FIB Bulletin - Bond of Reinforcement in Concrete

Punching of Flat Slabs with Large Amounts of Shear
Reinforcement
THSE NO 5409 (2012)

PRSENTE le 13 juillet 2012
LA FACULT DE L'ENVIRONNEMENT NATUREL, ARCHITECTURAL ET CONSTRUIT
LABORATOIRE DE CONSTRUCTION EN BTON
PROGRAMME DOCTORAL EN STRUCTURES
COLE POLYTECHNIQUE FDRALE DE LAUSANNE

POUR L'OBTENTION DU GRADE DE DOCTEUR S SCIENCES
PAR
Stefan Lips
accepte sur proposition du jury:

Prof. L. Laloui, prsident du jury
Prof. A. Muttoni, directeur de thse
Prof. K. Beyer, rapporteur
Prof. J. Hegger, rapporteur
Prof. A. M. Pinho Ramos, rapporteur
Suisse
2012
In loving memory of my Dad

Albert Lips
Prface
Les planchers-dalles sont une mthode de construction trs rpandu en plusieurs pays.
Cependant, dans certains cas leur comportement ltat limite ultime est encore insatisfaisant
cause de la fragilit trop importante en cas de rupture par poinonnement et plusieurs aspects
lis leur dimensionnement sont encore peu clairs. Pour ces raisons, depuis une douzaine
dannes, le poinonnement des dalles en bton arm reprsente un domaine important de
recherche au Laboratoire de Construction en Bton de lEPFL. Comme lont montr les travaux
prcdents, une armature transversale nest pas seulement utile pour augmenter la rsistance au
poinonnement, mais permet aussi damliorer sensiblement la capacit de dformation. Pour
ces raisons, dans la construction des planchers-dalles, lutilisation darmature contre le
poinonnement est de plus en plus rpandue et un modle physique permettant de dimensionner
larmature transversale a t dveloppe par M. Fernndez Ruiz et le soussign. Dans le cadre
de sa recherche, M. Lips a pu valider ce modle par une campagne dessais systmatique sur des
chantillons de dimensions relles et avec armatures proches de la ralit ainsi quamliorer le
modle dans sa prdiction de la capacit de dformation. Pour le mode de rupture caractris par
lcrasement de la premire bielle comprime qui dfinit la limite suprieure de la rsistance au
poinonnement, M. Lips a dvelopp un nouveau modle bas partiellement sur la thorie de la
fissure critique. Ce modle permet de dterminer de faon trs prcise la capacit de
dformation des dalles en cas de poinonnement et reprsente donc une amlioration importante
des connaissances dans ce domaine.
Flat slabs are a widespread construction method in several countries, in spite of the fact that, in
some cases, their ultimate limit state behavior is unsatisfactory because of their brittleness in the
case of failure by punching shear. Some points related to their design and dimensioning of flat
slabs remain unclear, which is why, over the past twelve years, the phenomenon of punching
shear failure of reinforced concrete slabs has been an important research field for the Structural
Concrete Laboratory of EPFL. As previous research works have shown, transverse shear
reinforcement does not only help increasing the punching shear strength, but also significantly
increasing the deformation capacity of flat slabs, for which the use of punching shear
reinforcement is becoming very common. The undersigned and Dr. Fernndez Ruiz have
developed a physical model for the design and dimensioning of transverse reinforcement to
prevent punching failure of flat slabs. In the framework of his research, Mr. Lips has validated
this model through a systematic test campaign on life-size specimens with realistic
reinforcement configurations. For the failure mode characterized by the crushing of the first
compression strut, which defines the upper limit of the punching shear strength, Mr. Lips
developed a new model partly based on the critical shear crack theory. This model allows
precisely determining the deformation capacity of flat slabs in case of punching and thus
constitutes a significant improvement of the knowledge in this field.
Lausanne, June 2012
Prof. Dr. Aurelio Muttoni
Acknowledgements
This research work was carried out at the Structural Concrete Laboratory (IBETON) at the
Ecole Polytechnique Fdrale de Lausanne (EPFL) under the supervision of Prof. A. Muttoni.
Thus, first of all, I would like to thank Prof. Muttoni for giving me the opportunity to do what I
like most and the liberty to do it as I like.
I would like to thank the jury members Prof. K. Beyer, head of the Earthquake Engineering and
Structural Dynamics Laboratory at the EPFL, Prof. J. Hegger, head of the Institute of Concrete
Structures at the Rheinisch-Westflische Technische Hochschule Aachen, Prof. A.M. Pinho
Ramos, head of the Structural Engineering Laboratory at the Universidade Nova de Lisboa, for
their engagement and for their valuable comments and suggestions. Additionally, I would like to
thank Prof. L. Laloui, head of the Laboratory of Soil Mechanics at the EPFL for his service as
president of the jury.
I wish to express my gratitude and sincere appreciation to the Swiss National Science
Foundation (Project # 121566) for financing this research work. Additionally, I would like to
thank the punching shear reinforcement manufacturer Fischer-Rista AG for allowing me to use
and to publish certain experimental data.
I would like to thank Dr. O. Burdet and Dr. M. Fernndez-Ruiz, Yvonne Bhl and all the
current and former PhD students of IBton with whom I shared a pleasant time with. Special
thanks go to my officemates Yaser Mirzaei and Jrgen Einpaul for the interesting, not always
work related, discussions and to Galina Argirova for the proof reading of my thesis as well as
for the moments of comfortable silence. Additionally, I would like to thank Fabio Brantschen
for the help with the French translation of the abstract.
I would like to address thanks to the lab technicians with special thanks to Grald Rouge and
Gilles Guignet without whom it would not have been possible to test 16 slabs in less than 4
months.
I would like to thank the members of the laboratories EESD and MCS for the enjoyable
memories. Special thanks to Talayeh Noshiravani and Hadi Kamyab for the great time inside as
well as outside of the EPFL.
I would like to thank all the people who in one way or another contributed and supported me
throughout my education although it is impossible to name here all by name. However, special
thanks go to Prof. A. Kenel who largely contributed to the fact that I started my PhD.
Finally, I would like to thank my family whose support allowed me to follow my dreams
whenever and especially wherever I wanted to.
iii
Abstract
Punching shear reinforcement is an efficient method to increase not only the strength but also
the deformation capacity of flat slabs supported by columns. Especially, the increase in
deformation capacity is desired so that the load can be distributed to other supports preventing a
total collapse of the structure in the case of the occurrence of a local failure. Thus, the research
presented herein addresses the punching strength as well as the deformation capacity of flat
slabs. Thereby, the focus is set on the analysis of the maximum increase in strength and rotation
capacity due to punching shear reinforcement. Therefore, the principal aim is the analysis of flat
slabs with large amounts of punching shear reinforcement. In addition to an experimental and
numerical investigation of flat slabs, another principal objective of the research project was the
development of an analytical model that enables accurate predictions of the punching strength
and the rotation capacity of flat slabs with large amounts of shear reinforcement. Thus, the
research presented herein can basically be divided into three main parts.
An experimental investigation of sixteen flat slab specimens with and without shear
reinforcement leads to new findings with respect to the punching strength and the loaddeformation response of flat slabs. The results of the tests serve for the validation of current
design codes and the Critical Shear Crack Theory. In addition to the specimens tested within
this research project, tests found in literature are used to investigate the influence of certain
parameters on the prediction of the punching strength.
A non-linear numerical model on the basis of the Finite Element Method enables the modeling
of the test specimens. This approach uses plane stress fields to calculate the moment-curvature
response of a discrete slab element. The thereby obtained flexural and torsional stiffness serve
as input parameters for a linear-elastic finite element analysis. This analysis enables the
modeling of the load-deformation response of the tested slab specimens leading to valuable
information regarding the state of deformation at different load levels.
The findings of the experimental and the numerical investigation support the development of an
analytical model. The theoretical background of this model is the Critical Shear Crack Theory,
which describes the punching strength as a function of the slab rotation. Thus, the developed
analytical model enables the calculation of the load-rotation response of flat slab specimens.
Moreover, the developed failure criteria enable the prediction of the punching strength as well
as the maximum rotation capacity. Finally, it is shown that the results obtained from the
developed model are in good agreement with results of tests performed within this research
project and of tests found in literature.
Keywords: punching shear, shear reinforcement, flat slabs, flexural response

v
Kurzfassung
Die Anordnung einer Durchstanzbewehrung ist eine wirksame Methode, um nicht nur den
Durchstanzwiderstand sondern auch das Verformungsvermgen von Flachdecken zu erhhen.
Vor allem das verbesserte Verformungsverhalten ist erstrebenswert, so dass die Lasten bei
einem lokalen Versagen umgelagert und ein sprdes Versagen des Bauwerks verhindert werden
kann. Deshalb befasst sich die hier prsentierte Forschungsarbeit mit dem Durchstanzwiderstand sowie dem Verformungsvermgen von Flachdecken. Dabei stand die Untersuchung
der maximalen Erhhung von Widerstand und Verformungsvermgen bei der Nutzung von
Durchstanzbewehrung im Vordergrund. Folglich ist das grundstzliche Thema der Arbeit die
Untersuchung von Flachdecken mit hohem Durchstanzbewehrungsgehalt. Neben einer
experimentellen und numerischen Untersuchung von Flachdecken soll ein Modell entwickelt
werden, welches den Durchstanzwiderstand sowie das Verformungsvermgen ermitteln kann.
Daher kann die Arbeit grundstzlich in drei Teile gegliedert werden.
Eine experimentelle Untersuchung von sechszehn Plattenausschnitten mit und ohne Durchstanzbewehrung fhrt zu neuen Erkenntnissen betreffend des Durchstanzwiderstandes und des
Verformungsvermgens. Die Versuchsergebnisse dienen zur Validierung von aktuellen
Bemessungsnormen und der Theorie des kritischen Schubrisses. Zustzlich zu den Versuchsresultaten dieser Forschungsarbeit werden Versuchsresultate aus der Literatur verwendet, um
den Einfluss verschiedener Parameter auf den ermittelten Durchstanzwiderstand zu untersuchen.
Ein numerisches Modell, basierend auf der Methode der Finiten Elemente, dient zur
Nachmodellierung der Versuchsplatten. Die Methode nutzt ebene Spannungsfelder um das
Verformungsverhalten des Querschnitts zu ermitteln. Die daraus gewonnenen Biege- und
Drillsteifigkeiten des Plattenquerschnitts dienen als Eingabeparameter fr eine linear-elastische
finite Elemente Berechnung. Diese Berechnung ermglicht das Nachbilden des
Verformungsverhaltens der Versuchsplatten, was zu wichtigen Erkenntnisse bezglich des
Verformungszustandes auf verschiedenen Laststufen fhrt.
Die Erkenntnisse der experimentellen und numerischen Untersuchung bilden die Grundlagen fr
die Entwicklung eines analytischen Modells. Als theoretische Basis dient das Modell des
kritischen Schubrisses, welches den Durchstanzwiderstand als Funktion der Plattenrotation
ermittelt. Dementsprechend erlaubt das analytische Modell die Berechnung des Rotationsverhaltens der Versuchsplatten. Des Weiteren ermglichen die entwickelten Bruchkriterien die
Bestimmung des Durchstanzwiderstandes und des maximalen Rotationsvermgens. Schliesslich
kann gezeigt werden, dass die Resultate des entwickelten Modells gut mit Versuchsresultaten
dieses Forschungsprojekts und Versuchsresultaten aus der Literatur bereinstimmen.
Stichworte : Durchstanzen, Durchstanzbewehrung, Flachdecken, Verformungsverhalten
vii
Rsum
Larmature de poinonnement est une mthode efficace pour augmenter non seulement la
rsistance mais aussi la capacit de dformation des planchers-dalles. En particulier,
lamlioration de la capacit de dformation est souhaitable afin quil soit possible de
redistribuer la charge en cas de rupture locale vitant ainsi un effondrement de toute la structure.
De ce fait, la prsente recherche traite de la rsistance au poinonnement et de la capacit de
dformation des planchers-dalles, avec pour point central leffet de larmature de
poinonnement sur laugmentation maximale de la rsistance au poinonnement et la capacit
de rotation. Lobjectif principal est donc lanalyse des planchers-dalles avec une quantit
importante darmatures de poinonnement. Paralllement lanalyse exprimentale et
numrique, un modle pouvant prdire la rsistance au poinonnement ainsi que la capacit de
la dformation dun plancher-dalle a t dvelopp. Le travail peut ds lors tre divis en trois
parties principales.
Linvestigation exprimentale des seize spcimens des planchers-dalles sans et avec armature de
poinonnement donne de nouvelles informations concernant la rsistance au poinonnement et
le comportement charge-dformation. Les rsultats de ces essais servent de validation des
normes de dimensionnement et de la thorie de la fissure critique. En plus des essais contenus
dans le cadre de ce projet de recherche, des essais de la littrature sont utiliss pour la recherche
de linfluence de certains paramtres quant la prdiction de la rsistance au poinonnement.
Un modle numrique bas sur la mthode des lments finis permet la modlisation des
spcimens essays. Cette mthode utilise des champs de contrainte en plan pour dterminer le
comportement en section. La rigidit obtenue par ce calcul est utilise comme valeur dentre au
calcul dlments finis linaires-lastiques et permet la modlisation pertinente du
comportement charge-dformation des spcimens tests.
Les rsultats des essais et du calcul numrique servent au dveloppement dun modle
analytique. Le modle est bas sur la thorie de la fissure critique, qui dfinit la rsistance au
poinonnement en fonction de la rotation de la dalle. Le modle analytique permet ainsi le
calcul du comportement charge-rotation dun plancher-dalle. De plus, les critres de rupture
dvelopps dans cette recherche permettent la prdiction de la rsistance au poinonnement et
de la capacit de la rotation. Les rsultats du modle propos donnent de bonnes corrlations
avec les rsultats des essais dans le cadre de cette recherche ainsi quavec les essais trouvs dans
la littrature.
Mots-cls: poinonnement, armature de poinonnement, planchers-dalles, comportement la

flexion
ix
Table of contents
Prface ........................................................................................................................................... i
Acknowledgements .................................................................................................................... iii
Abstract ........................................................................................................................................ v
Kurzfassung ...............................................................................................................................vii
Rsum......................................................................................................................................... ix
Table of contents ......................................................................................................................... xi
Abbreviations ...........................................................................................................................xvii
Notations .................................................................................................................................... xix
1.
2.
Introduction .......................................................................................................................... 1
1.1.
Research significance .................................................................................................. 3
1.2.
Objectives .................................................................................................................... 3
1.3.
Scope ........................................................................................................................... 4
1.4.
Organization ................................................................................................................ 5
1.5.
Personal contributions ................................................................................................. 6
Literature Review ................................................................................................................ 7

2.1.
Overview of previous research .................................................................................... 8
2.2.
ACI 318-11 (ACI 318 2011)...................................................................................... 11

2.2.1. Slabs without shear reinforcement ................................................................ 11
2.2.2. Slabs with shear reinforcement ..................................................................... 11
2.3.
Eurocode 2 (EC2 2004) ............................................................................................. 13

2.4.
German National Annex (NAD 2011) ....................................................................... 16

2.5.
SIA 262 (SIA 262 2003) ............................................................................................ 18

2.6.
Model Code (MC 2011)............................................................................................. 21
xi
2.6.1. Slabs without shear reinforcement ................................................................21

2.6.2. Slabs with shear reinforcement .....................................................................22
2.7.
Critical shear crack theory (CSCT)............................................................................25

2.7.1. Slabs without shear reinforcement ................................................................25
2.7.2. Slabs with shear reinforcement .....................................................................26
2.8.
Load-rotation response ..............................................................................................34

2.8.1. Analytical models .........................................................................................34
2.8.2. Numerical models .........................................................................................35
2.8.3. The Quadrilinear model (Muttoni 2008) .......................................................38
3.
Experimental Campaign....................................................................................................43
3.1.
Geometry and reinforcement .....................................................................................44
3.2.
Materials ....................................................................................................................46
3.3.
Test set-up ..................................................................................................................47
3.4.
Measurements ............................................................................................................48
3.5.
Results........................................................................................................................48
3.6.
Discussion of the results ............................................................................................52

3.6.1. Performance of the shear reinforcement .......................................................52
3.6.2. Column size ..................................................................................................52
3.6.3. Slab thickness................................................................................................53
3.6.4. Amount of shear reinforcement ....................................................................54
3.6.5. Shear deformations at column face ...............................................................54
3.6.6. Opening of the shear cracks ..........................................................................56
3.6.7. Strains in the studs ........................................................................................57
3.6.8. Deformations at the shear-critical region ......................................................62
4.
Validation of Code Provisions and the CSCT .................................................................63

4.1.
ACI 318-11 (ACI 318 2011) ......................................................................................64

4.1.1. Punching strength..........................................................................................64
4.1.2. Slab thickness................................................................................................64
4.1.3. Column size ..................................................................................................65
4.1.4. Shear reinforcement ratio ..............................................................................65
4.2.
Eurocode 2 (EC2 2004) .............................................................................................66

4.2.1. Strength prediction ........................................................................................66
4.2.2. Slab thickness................................................................................................66
xii
4.2.3. Column size .................................................................................................. 66

4.2.4. Shear reinforcement ratio.............................................................................. 67
4.3.
German National Annex to Eurocode 2 (NAD 2011) ............................................... 68

4.3.1. Strength prediction ........................................................................................ 68
4.3.2. Slab thickness ............................................................................................... 68
4.3.3. Column size .................................................................................................. 69
4.4.
SIA 262 (SIA 262 2003) ............................................................................................ 70

4.4.2. Slab thickness ............................................................................................... 70
4.4.3. Column size .................................................................................................. 71
4.5.
fib Model code (MC 2011) ........................................................................................ 72

4.5.2. Slab thickness ............................................................................................... 72
4.5.3. Column size .................................................................................................. 73
4.6.
Critical shear crack theory (CSCT) ........................................................................... 74

4.6.2. Slab thickness ............................................................................................... 75
4.6.3. Column size .................................................................................................. 75
4.6.5. Prediction of the slab response and failure criteria ....................................... 76
4.7.
Overview of the performance of the codes ................................................................ 79

4.7.2. Effective depth .............................................................................................. 84
4.7.3. Column size .................................................................................................. 85
4.7.4. Flexural reinforcement ratio ......................................................................... 87
4.7.5. Concrete compressive Strength..................................................................... 88
5.
Development of a Nonlinear Finite Element Approach.................................................. 89

5.1.
Flexural stiffness........................................................................................................ 91
5.1.1. Compatibility conditions............................................................................... 91
5.1.2. Material behavior .......................................................................................... 92
xiii
5.1.3. Equilibrium conditions..................................................................................96

5.2.
Shear stiffness ............................................................................................................99
5.3.
Analysis ...................................................................................................................100
5.4.
Comparison ..............................................................................................................101
5.4.1. Pure bending ...............................................................................................102
5.4.2. Pure torsion .................................................................................................104
5.4.3. Punching of slabs without shear reinforcement ..........................................106
5.4.4. Punching of slabs with shear reinforcement ...............................................109
6.
7.
Analysis of the Slab Response .........................................................................................113

6.1.
Analysis procedure ..................................................................................................114
6.2.
Global slab behavior ................................................................................................117
6.3.
Local slab behavior ..................................................................................................126
Development of an Analytical Model .............................................................................131

7.1.
General slab behavior ..............................................................................................132
7.2.
Load-rotation response ............................................................................................137

7.2.1. Global slab behavior ...................................................................................137
7.2.2. Local slab behavior .....................................................................................141
7.2.3. Equilibrium conditions................................................................................144
7.3.
Failure criteria ..........................................................................................................148
7.4.
Definition of model parameters ...............................................................................158

7.4.1. Limitation of the radial curvature r,lim at radius r2 .....................................158
7.4.2. Radius r0......................................................................................................160
7.4.3. Shear crack distance r2 ................................................................................161
7.4.4. Load distribution factor ............................................................................163
8.
Validation of the Analytical Model .................................................................................169

8.1.
Transformation of the specimen ..............................................................................170

8.1.1. Influence of the orthogonal reinforcement ..................................................171
8.1.2. Transformation of the column shape...........................................................176
8.1.3. Transformation of slab shape and loading conditions.................................178
8.1.4. Load application..........................................................................................181
8.2.
Other failure modes considered ...............................................................................186

8.2.1. Punching of slabs without shear reinforcement ..........................................186
8.2.2. Punching outside the shear-reinforced area ................................................187
xiv
8.3.
Validation with tests within this research ................................................................ 187

8.3.1. Strength and rotation predictions ................................................................ 188
8.3.2. Slab thickness ............................................................................................. 188
8.3.3. Column size ................................................................................................ 189
8.3.4. Shear reinforcement ratio............................................................................ 190
8.3.5. Prediction of the slab response and the failure criteria ............................... 191
8.4.
Comparison to tests from literature ......................................................................... 195

8.4.1. General ........................................................................................................ 195
8.4.2. Shear reinforcement ratio............................................................................ 196
8.4.3. Effective depth ............................................................................................ 197
8.4.4. Column size ................................................................................................ 198
8.4.5. Flexural reinforcement ratio ....................................................................... 199
8.4.6. Concrete compressive strength ................................................................... 200
8.4.7. Maximum aggregate size ............................................................................ 200
8.4.8. Predicted rotation at failure......................................................................... 201
9.
Conclusions and Future Research .................................................................................. 205

9.1.
Conclusions ............................................................................................................. 206
9.2.
Recommendations for future research ..................................................................... 210
Bibliography ............................................................................................................................. 211

Appendix A Test Database
Appendix B Example PL7
Appendix C Shear Reinforcement Ratio
xv
Abbreviations
ACI 318-11
American Concrete Institute Building Code (refer to reference ACI 318 2011)
Avg.
Average value
CSCT
Critical shear crack theory

(refer to references Muttoni 2008 and Fernndez Ruiz and Muttoni 2009)
COV
Coefficient of variation
EC 2004
Eurocode 2 (refer to reference EC2 2004)
MC 2010
fib Model Code 2010 (refer to reference MC 2011)
NAD 2011
German National Annex to Eurocode (refer to reference NAD 2011)
NLFEA
Nonlinear finite element analysis
SIA 2003
Swiss code 262 (refer to reference SIA 262 2003)
xvii
Notations
Asw
cross-sectional area of shear reinforcement
Asw1
cross-sectional area of shear reinforcement crossed by the outer shear crack
Ca
horizontal component of force in compression strut within the wedge element
Cb
horizontal component of force in compression strut in the outer slab segment
Cc
compression force due to bending
Cr2
total compression force action at radius r2
Ec
Youngs modulus of concrete
Ec0
Youngs modulus of uncracked concrete
Es
Youngs modulus of flexural reinforcing steel
Esw
Youngs modulus of shear reinforcing steel
EI0
flexural stiffness before cracking
EI1
tangential flexural stiffness after cracking (CSCT)
EIII
tangential flexural stiffness after cracking
Fc,a
inclined compression force in the strut within the wedge element
Fc,b
inclined compression force in the strut in the outer slab segment
Fc,c
compression force due to bending
Gc
shear modulus of concrete
Gc0
shear modulus of uncracked concrete
stiffness
Trc
tensile force acting at the column face
Tr2
tensile force acting at radius r2
punching shear load
VR
punching shear strength
Vflex
shear force associated with the flexural capacity of the slab specimen
VR,pred
predicted punching shear strength
VR,test
measured punching shear strength
VRc
concrete contribution to the punching strength
VRs
shear reinforcement contribution to the punching strength
VR,I
predicted punching strength for failure within the shear-reinforced area
VR,II
predicted failure load for failure of the compression strut
Vs
shear reinforcement contribution to the punching strength (CSCT)

xix
strain transformation matrix
parameter for quadratic function defining the distribution of the rotation
asx, asy
sectional area of longitudinal reinforcement per unit width
distance between load application points
b0
control perimeter (unless noted otherwise set at d/2 of the border of the support region
with circular corners)
b1
distance between load application point and the slab edge
bext
control perimeter outside the shear-reinforced area
bw
width of the compression strut
side length of the column

parameter for quadratic function defining the distribution of the rotation
effective depth (i.e. distance from extreme compression fiber to the centroid of the
longitudinal tensile reinforcement)
deff
measured effective depth (distance from extreme compression fiber to the centroid of
the longitudinal tensile reinforcement)
dg
maximum diameter of concrete aggregate
dg0
reference aggregate size (16 mm)
dn
nominal effective depth (distance from extreme compression fiber to the centroid of
the longitudinal tensile reinforcement)
dw
diameter of shear reinforcement
dv
shear resisting effective depth of the slab (MC)
dv,ext
distance between the flexural reinforcement and the bottom end of the vertical branch
of the shear reinforcement
fc
average compressive strength of concrete (measured on cylinders)
fct
average tensile strength of concrete
fy
yielding strength of flexural reinforcement
fyw
yielding strength of shear reinforcement
fyw,ef
effective stress in the shear reinforcement accounting for limited anchorage of the
shear reinforcement in thin slabs (EC, NAD)
slab thickness
ht
stud length, length of the vertical branch of the stirrup
change in slab thickness
factor accounting for size effect (EC)
k1
factor accounting for the reduction in strength due to transverse strains
k2
factor accounting for the stress distribution within the compression strut
xx
kr
factor accounting for slab rotation (SIA)
ksh
ratio of ultimate strength to the yielding strength of reinforcing steel
ksys
coefficient accounting for the performance of the shear reinforcement system (MC)
factor accounting for slab rotation (MC)
side length of the slab specimen

span between columns (SIA)
lai
smaller distance between one end of the shear reinforcement and the shear crack
las
larger distance between one end of the shear reinforcement and the shear crack
lbi
distance between the bottom end of the shear reinforcement and the shear crack
lbs
distance between the top end of the shear reinforcement and the shear crack
lcut
length of the cut for the numerical modeling of the column vicinity
lw
length of the vertical branch of the shear reinforcement
moment per unit width
m0
reference moment per unit width
m1
first principal moment per unit width
mc
sectional moment per unit width due to the stresses in the concrete
mcr
cracking moment per unit width
mr
radial moment per unit width
mR
nominal moment capacity per unit width
ms
sectional moment per unit width due to the stresses in the longitudinal reinforcement
(MC)
mt
tangential moment per unit width
mt,int
tangential moment per unit width within the shear-critical region
mt,ext
tangential moment per unit width outside the shear-critical region
mx, my
moment per units length in direction of the reinforcing bars
mxy
torsional moment per unit length
ncut
number of vertical branches of shear reinforcement crossing the outer shear crack
ns
number of vertical branches of shear reinforcement per radius
nr
number of vertical branches of shear reinforcement in the first perimeter
r0
radius from which the rotations are assumed to be constant

radius of the critical shear crack (CSCT, Quadrilinear model)
r1
radius of the resultant shear force crossing the crack

radius of the zone in which cracking is stabilized (CSCT, Quadrilinear model)
r2
radius of the critical shear crack
rc
radius of a circular column
xxi
rcr
radius of cracked zone
rq
radius of the load introduction at the perimeter
rp
radius of the zone in which cracking is stabilized
rs
radius of circular isolated slab element
ry
radius of yielded zone
s0
distance measured with respect to slab plane between the border of the support region
and the first vertical branch of shear reinforcement
s1
distance measured with respect to slab plane between two adjacent vertical branches of
shear reinforcement of same radius
sc
horizontal width of the compression strut
scr
average crack spacing
st
distance measured with respect to slab plane between two adjacent vertical branches of
the stirrups
vertical displacement
wi
crack opening at the ith vertical branch of shear reinforcement
wlim
limit crack width defining the anchorage condition of the shear reinforcement
x1
height of the compression zone at the column face
x2
height of the compression zone at the outer slab segment
xa
height of the compression zone due to the force in compression strut between the 1st
row of shear reinforcement and the column
xb
height of the compression zone due to the force in compression strut between the 2nd
row of shear reinforcement and the column
xc
height of the compression zone due to bending
xel
height of the compression zone calculated with linear-elastic material behavior
x, y
coordinates (unless noted otherwise corresponding to the horizontal slab plane)
coordinate perpendicular to the horizontal slab plane
change in slab thickness
vertical displacement due to shear deformations at the column face
length increment
change in stress in the shear reinforcement
angle of a slab segment
xxii
inclination of the compression strut between the 1st row of shear reinforcement and the
column
inclination of the compression strut between the 2nd row of shear reinforcement and
the column
cr
inclination of the outer shear crack
efficiency factor of the bending reinforcement for stiffness calculation
shear strains
partial safety factor for concrete (NAD)
partial safety factor for steel (NAD)
deformations
strains
1, 2
principal strains
strain at the peak stress of concrete
strain in the shear reinforcement
yielding strain of the shear reinforcement
direction of the principal stresses and/or strains
factor accounting for localization of rotation in the shear crack
load distribution factor

coefficient accounting for the performance of the shear reinforcement system (CSCT)
partial load distribution factor (between concrete and shear reinforcement)
partial load distribution factor (between perimeters of shear reinforcement)
1, 2
fitting parameter
flexural reinforcement ratio
shear reinforcement ratio (calculated according to Appendix C)
stresses in the flexural reinforcement
smax
maximum stress in the flexural reinforcement
stresses in the shear reinforcement
bond strength
curvature at stabilized crack phase (CSCT)
cr
curvature at cracking
curvature at stabilized crack phase
curvature in radial direction
r,lim
limit of curvature in radial direction
curvature in tangential direction
yielding curvature
xxiii
TS
decrease in curvature due to tension stiffening
slab rotation
constant slab rotation at the outer part of the slab specimen
R,pred
predicted rotation at failure
R,test
measured rotation at failure
xxiv
1. Introduction
The development of modern reinforced concrete enabled for the first time the use of slabs as a
structural element in civil engineering. In contrast to steel and timber structures, reinforced
concrete structures were no longer limited to columns and beams, thus allowing new
constructions methods. At first, reinforced concrete slabs were still supported on girders.
However, at the beginning of the 20th century the use of flat slabs prevailed. The advantages of
flat slabs compared to slabs on girders were early recognized and are still valid today. In 1914,
Eddy and Turner (Eddy and Turner 1914) wrote:
The superiority of flat slab floor supported directly on columns, over other forms
of construction when looked at from the standpoint of lower cost, better lighting,
greater neatness of appearance, and increased safety and rapidity of construction,
is so generally, or rather universally conceded as to render any reliable
information relative to the scientific computation of stresses in this type of
construction of great interest.
Nearly hundred years ago, Eddy and Turner highlighted the principal challenge of the design of
flat slabs at that time, which concerned the calculation of the stresses. Since no design
guidelines existed at first, the design of flat slabs was generally based on experimental data.
This led to the development of rather different approaches for the design of flat slabs, which can
be best displayed in the varying layouts used for the flexural reinforcement. For example in the
United States, Turner used a four-way flat slab system (Figure 1.1a; Figure 1.2a) whereas
Condron introduced a two way system (Figure 1.1b; Figure 1.2b) (Eddy and Turner 1914;
Condron 1913). At nearly the same time, Maillart independently developed a similar system in
Europe (Maillart 1926). In this time period, other flat slab systems were proposed that one
would today consider as rather special. For example, Smulski (Smulski 1918) developed a
circumferential flat slab system, which consisted of radially and tangentially arranged
reinforcement (Figure 1.1c).
Despite the different flexural reinforcement layouts, all previously mentioned approaches used
an enlarged column head. Although in each approach the shape slightly differed, the main
purpose of the enlargement was to enable the transfer of the load from the slab to the column
and thus prevent a punching failure. Additionally, the enlargement contributed to the flexural
capacity of the slab since it reduced the span between the columns. With time, the enlargement
of the column was replaced by a steel head within the slab.
Chapter 1
(a)
(b)
(c)
Figure 1.1: Reinforcement layouts for in early developed flat slab systems: (a) four-way flat slab
system, (b) two-way flat slab system, and (c) circumferential flat slab system (Taylor and
Thompson 1916)
The increasing number of research on punching led to a better understanding and a better
prediction of the punching strength. The consequence was that flat slabs could be designed
without any special reinforcement against punching. However, the problematic of such an
approach is that punching failure is a rather brittle failure mode and it can occur without any
warning signs. Throughout history, this led to several severe collapses with numerous
casualties. In order to prevent such accidents, integrity reinforcement was introduced to increase
the residual strength of a slab-column connection after the occurrence of a punching failure.
Figure 1.2:
Longitudinal reinforcement layout in the case of (a) a four-way slab system (concrete
removed after testing) (Talbot and Gonnerman 1918) and (b) a two-way slab system (before
casting) (Condron 1913)
However, in this context, the question rises if it is desired to prevent a collapse passively with
integrity reinforcement or should active failure prevention be desired instead. Analogous to the
first use of flat slabs, reinforcement can be placed in different ways but the objective should
always be to place it where it performs best. In order to determine the performance of flat slabs,
2
Introduction
the design should not only consider the force capacity but also of the deformation capacity. The
consideration of the deformation capacity easily leads to the conclusion that generally slabs
without shear reinforcement cannot provide sufficient deformation capacity. This is confirmed
by several building collapses, Thus, the question should not be whether punching shear
reinforcement should be used or not, but what amount of punching shear reinforcement is
necessary and where should it be placed so that it provides a satisfactory deformation capacity.
These questions can only be answered by the investigation of the load-deformation response of
flat slabs. However, the prediction of the displacements of flat slabs is not an easy task to
achieve. Generally, the non-linear response of reinforced concrete is challenging. Moreover, the
load concentration in the column vicinity requires a rather sophisticated model. Therefore, this
thesis intends to not focus solely on the punching strength but also on the response of flat slabs.
1.1. Research significance

Punching shear reinforcement is an efficient way to increase not only the strength but also the
deformation capacity of slab-column connections. However, the analysis of such a connection is
rather complex and includes several challenges. One challenge is the difference in performance
of different types of punching shear reinforcement. Each type leads to a rather different
performance, largely depending on the anchorage condition of the shear reinforcement system
and the distribution of the shear reinforcement. Moreover, the amount and the arrangement of
the shear reinforcement do not only influence the performance but also define the failure mode.
Consequently, the punching strength depends on various parameters that have to be investigated
individually. Currently, only scarce systematic research on this subject exists in literature for
full-scale specimens. Therefore, this research project focuses on the detailed investigation of
punching of full-scale slab specimens with large amounts of shear reinforcement.
1.2. Objectives
The objective of this research is to gain a better understanding of punching of flat slabs with
shear reinforcement. Thereby, the focus should be set on the analysis of the maximum increase
in strength and rotation capacity due to punching shear reinforcement. Therefore, the principal
aim is the analysis of flat slabs with large amounts of punching shear reinforcement. Within this
framework, several aspects should be investigated such as the load-deformation response of the
slab, the failure mechanism, and the load contribution of the shear reinforcement. Based on this
investigation, a simplified model should be developed that enables the prediction of the
punching strength and the rotation at failure for the investigated cases.
Chapter 1
1.3. Scope
The research presented herein basically focuses on the investigation of punching of flat slabs
with large amounts of shear reinforcement. This includes the investigation of slabs without and
with low amounts of shear reinforcement in order to analyze the influence of the shear
reinforcement. All the investigated cases refer to interior columns supported by square or
circular columns without any constraints at the boundary. Thus, the research concerns solely
symmetrically loaded slabs that were not subjected to membrane forces. The shear
reinforcement considered was limited to vertical, pre-installed shear reinforcement systems
(Figure 1.3).
(a)
(b)
(c)
(d)
(e)
(f)
(g)
Figure 1.3: Examples of shear reinforcement systems: (a) corrugated double headed shear studs, (b)
smooth double headed shear studs, (c) steel offcuts, (d) headed stirrups, (e) stirrups with lap
at the vertical branch, (f) stirrups or shear links, (g) continuous stirrups or cages of shear
links
The choice of the type, the amount, and the distribution of the shear reinforcement defined also
the failure modes that were considered within this research. Therefore, the framework of the
research was generally limited to the investigation of the failure within the shear reinforced area
(Figure 1.4a) and the failure due to crushing of the concrete strut near the column (Figure 1.4c).
However, failure outside the shear-reinforced area (Figure 1.4b) was not investigated in detail
but was considered in the validation of the code provisions, the critical shear crack theory, and
the model presented herein. On the other hand, failure modes such as delamination of the
concrete core (Figure 1.4d) or failure between the transverse reinforcement (Figure 1.4e), which
result directly from detailing that contradicts common design practice (e.g. large spacing,
insufficient anchorage), were neither analyzed nor used for the validation.
Introduction
(a)
(d)
(b)
(e)
(c)
(f)
Figure 1.4: Possible failure modes of slabs with shear reinforcement: (a) failure within shear-reinforced
area, (b) failure outside shear-reinforced area, (c) failure close to the column due to crushing
of concrete, (d) delamination of the concrete core, (e) failure between the transverse
reinforcement, and (f) flexural failure
1.4. Organization
The thesis covers three main parts, namely the experimental investigation including the
validation of code provisions and the critical shear crack theory, the numerical analysis of the
slab response, and the development and validation of an analytical model. After the
introduction, Chapter 2 presents an overview of the previously performed research on punching
of flat slabs with shear reinforcement. Additionally, it shows the code provisions and the
formulations of the critical shear crack theory used for the calculation within this research.
Afterwards, Chapter 3 presents an overview of the test campaign that was carried out within this
research project and presents selected results that are used and further discussed in the
subsequent chapters. In addition, the experimental part consists of a code validation presented in
Chapter 4 by which the test results from this research and test results found in literature are
compared to the predictions of current design codes and the critical shear crack theory.
Chapter 5 presents the development of a nonlinear finite element approach that is based on
plane stress fields. This approach is used for a detailed analysis of the global and local slab
behavior presented in Chapter 6.
Based on the experimental investigation and the numerical analysis, an analytical model was
developed. The mechanical basis and the derivation of the equations of this model are presented
in Chapter 7. In Chapter 8, this model is validated based on test results from this research and
test results found in literature.
Finally, the thesis closes with the conclusions of the three parts, followed by an outlook for
future research.
Chapter 1
1.5. Personal contributions

Within the research presented herein, following personal contributions were made:
Development and performance of an extensive experimental test campaign
In-depth analysis of measurement data obtained from the test campaign
Validation of current code provisions and the critical shear crack theory
Development and application of a constitutive model in order to model the slab

behavior subjected to a combination of flexural and torsional moments
Investigation of the slab response with respect to a global and a local part
Development and validation of an analytical model based on the critical shear crack
theory for the prediction of the punching strength and the rotation capacity of flat slabs
2. Literature Review
The start of the use of flat slabs supported by columns in the beginning of the 20th century led to
various research on the punching strength of flat slabs. At first, research covered mainly slabs
without punching shear reinforcement, followed by investigations on flat slabs with punching
shear reinforcement. Coming from the beam design, the first shear reinforcement used were
bent-up bars. Later on, new systems have been developed such as different stirrup systems and
shear studs. The change of the punching shear reinforcement system was always accompanied
by research on this subject resulting in new findings for different shear reinforcement systems.
Additionally, it can be noted that not only the shear reinforcement systems changed by time but
also the demand on the behavior. At first the increase of the punching strength occupied
researchers interest. However, later they diverted their focus on the deformation capacity and
safety of slab column connections. This influenced the research of flat slabs as well as the
further development of punching shear reinforcement systems.
This chapter gives a brief overview of the developments regarding punching of flat slabs with
shear reinforcement. Afterwards, it presents current code provisions and the critical shear crack
theory (CSCT) with respect to the formulations that were used for the calculations within this
research. Therefore, only formulations for symmetric slabs without shear reinforcement or with
vertical shear reinforcement for failure within the shear-reinforced area, failure outside the
shear-reinforced area, and failure of the concrete strut near the column are presented. Moment
transfer, asymmetric geometrical or loading conditions, prestressing, or inclined shear
reinforcement are not considered in the calculations. Thus, no formulations regarding these
subjects will be presented.
Since certain punching shear models such as the critical shear crack theory depend on the slab
deformation, this chapter additionally presents a short overview of different methods to predict
the response of a flat slab. For this discussion, the approaches will be separated into analytical
and numerical methods. Finally, a more detailed explanation of an analytical approach proposed
by Muttoni (Muttoni 2008) will be presented.
Chapter 2
2.1. Overview of previous research

This brief summary presents the main developments of research on the punching of flats slabs
with punching shear reinforcement. Considering the extensive amount of research on flat slabcolumn performed over the last decades, a complete review of all experimental investigations
and developed models would go beyond the scope of this work. Additionally, it can be noted
that overviews of models already exist such as the fib Bulletin 12 (FIB 2001). Therefore, this
brief summary concentrates on research that is seen as most crucial with respect to the work
within the herein presented research project.
A good starting point for such a summary is certainly the research from Kinnunen and
Nylander. Their contribution in 1960 (Kinnunen and Nylander 1960) was one of the first and
probably most important contribution with respect to the modeling of punching. The proposed
model led to the further development of other punching shear models. Although this approach
was developed for slabs without punching shear reinforcement, it served as basis for other
researchers who implemented punching shear reinforcement. In 1963, based on the model of
Kinnunen and Nylander (Kinnunen and Nylander 1960), Andersson (Andersson 1963)
developed an approach that considers shear reinforcement (Figure 2.1a). In the tests that he
performed for the model validation, he used bent-up bars and continuous stirrups as punching
shear reinforcement.
In 1974, the American Concrete Institute published the Special Publication 42 about shear in
reinforced concrete in which Part 4 was devoted to shear in slabs. Amongst other contributions,
Hawkins (Hawkins 1974) published a paper presenting an overview of tests performed with
different punching shear reinforcement systems such as steel heads, bent-up bars, and stirrups.
He concluded that shear reinforcement increases the punching strength even for small slabs and
that the detailing is crucial to increase the strength and to avoid undesired failure modes.
During the seventies, Ghali and Dilger from the University of Calgary, Canada, focus on
improving existing shear reinforcement systems, which were at this time generally bent-up bars
or different types of stirrups. They found that the anchorage conditions of the shear
reinforcement are crucial. At first, they used cut-off of standard I-shaped steel beams (Langohr
et al. 1976). Afterwards, they collaborated with Andr (Andr 1979; Andr et al. 1979; Andr
1981) who introduced stud rails as new punching shear reinforcement in Germany. This new
system led to an extensive research throughout the eighties and nineties (Seible et al. 1980;
Dilger and Ghali 1981; Van der Voet et al. 1982; Mokhtar et al. 1985; Ghali 1989; Elgabry and
Ghali 1990; Ghali and Hammill 1992; Megally and Ghali 1994; Hammill and Ghali 1994;
Birkle and Dilger 2008). This research was accompanied by the development of the shear
friction model that was first developed for shear in beams (Loov 1998; Tozser 1998) and later
applied for slab-column connections (Dechka 2001; Birkle 2004).
Literature Review
During the eighties and nineties, Regan started his research on punching with and without shear
reinforcement at the Polytechnic of Central London (Regan 1983; Regan 1985; Regan 1986). In
the research team of Regan, Shehata developed a model for slabs without shear reinforcement
that was based on the approach of Kinnunen and Nylander (Shehata 1985; Shehata and Regan
1989; Shehata 1990). Later, Gomes extended Shehatas model by implementing the contribution
of the shear reinforcement (Gomes and Regan 1999a; Gomes and Regan 1999b) (Figure 2.1b).
Further research has been conducted by Regan and Samadian (Regan and Samadian 2001) and
Oliveira et al. (Oliveira et al. 2000) who continued their work in Brazil leading to several recent
publications about punching tests with shear reinforcement (Trautwein et al. 2011; Carvalho et
al. 2011).
In the United Kingdom in the beginning of the nineties, Chana and Desai performed an
extensive experimental campaign of punching shear tests with shear reinforcement (Chana and
Desai 1992; Chana 1993). Thereby, they tested slabs with conventional shear links and slabs
with a special shear reinforcement system consisting of links welded together to a cage (known
as shearhoop system). The main objective of this investigation was to show the improved
performance of the prefabricated system compared to the conventional shear links and its code
applicability.
Starting in the nineties, Broms presented a further development of the model of Kinnunen and
Nylander (Broms 1990a) and introduced a combination of stirrups and bent-up bars as punching
shear reinforcement (Broms 1990b) (Figure 2.1c). He showed that this system allows an
increase in the deformation capacity compared to slabs with only stirrups. The increase in
deformation capacity of slab-column connections and the further development of this model
have been his main research interest over the years leading to various publications (Broms
2000a; Broms 2000b; Broms 2006; Broms 2007a; Broms 2007b). In 2005, he summarized a
main part of his earlier work in his dissertation treating design methods for punching of flat
slabs and footings with and without shear reinforcement (Broms 2005).
Also in the last two decades, the research group of Hegger at the Rheinisch-Westflischen
Technischen Hochschule Aachen in Germany performed extensive experimental research on
punching of flat slabs and foundations and thoroughly investigated the structural behavior of
slabs with and without punching shear reinforcement (Hegger et al. 2006; Hegger et al. 2007;
Hegger et al. 2009; Hegger et al. 2010; Siburg and Hegger 2011). With respect to punching of
slabs with punching shear reinforcement, the dissertations written by Beutel (Beutel 2003) and
Husler (Husler 2009) contributed largely to the understanding of the flat slab behavior.
As already mentioned, more recent work has been performed in Brazil regarding punching shear
reinforcement in combination with prestressing (Carvalho et al. 2011) and the performance of
punching shear reinforcement that does not embrace the flexural reinforcement (Trautwein et al.
2011). Other recent experimental research has been performed at Imperial College London in
the United Kingdom by Vollum et al. (Vollum et al. 2010) in which the arrangement of the
punching shear reinforcement was investigated.
9
Chapter 2
In this summary, the research of Muttoni performed at the Ecole Polytechique Fdrale de
Lausanne is excluded since the latest contribution regarding the Critical Shear Crack Theory
(CSCT) with respect to punching with punching shear reinforcement will be presented in detail
subsequently in Subchapter 2.7. However, before the CSCT is described, the current code
provisions used within this research will be presented.
(a)
Andersson (Andersson 1963):

The slab part between the column face and the
edge of the slab is assumed to rotate rigidly.
This outer slab part is assumed to be carried
by a compression zone that is supported by
the column.
Failure occurs if a defined tangential
compressive strain at a defined distance away
of the column is reached.
(b)
Gomes and Regan (Gomes and Regan 1999a):

It is assumed that the slab can be divided into
three parts: an outer, a wedge, and a column
part. However, despite the separation, it is
assumed that the entire slab except the column
part rotates as a rigid body.
Failure is assumed to occur either if the shear
stress on any surface below the shear crack
reaches the sliding resistance or if the
maximum principal stress reaches the indirect
tensile strength at a section with a defined
radius.
(c)
Broms (Broms 2005):

Failure is assumed to occur either if the
tangential compressive strain reaches a
defined value or if the compression stress in
the fictitious internal column capital reaches a
critical value. In both cases, the failure
criterion depends on the state of stress in the
flexural reinforcement.
Figure 2.1: Selected models that are based on the model of Kinunnen and Nylander (Kinnunen and
Nylander 1960): (a) model proposed by Andersson (Andersson 1963), (b) model proposed
by Gomes and Regan (Gomes and Regan 1999a), and (c) model proposed by Broms (Broms
2005)
10
Literature Review
2.2. ACI 318-11 (ACI 318 2011)

2.2.1. Slabs without shear reinforcement
The code provision of ACI 318-11 regarding punching of slabs without shear reinforcement is
rather simple and straightforward. The area at the control perimeter is multiplied by an
admissible shear stress. Thus, the punching strength is defined as:
1
3
(2.1)
where b0 is a control perimeter, d is the effective depth of the slab, fc is the compressive strength
of concrete in MPa.
The control perimeter is defined in clause 11.11.1.2, which would suggest that the perimeter is,
alike other codes, circular at the corners. However, clause 11.11.1.3 allows using straight sides
at the corner in the case of square or rectangular columns. Since, in practice, it seems more
reasonable to use the largest control perimeter allowed, the critical perimeter is used with
straight lines for the comparison of the tests with the ACI 318-11 code. Therefore, the critical
perimeter used within this research corresponds to the drawing shown in Figure 2.2b.
(a)
Figure 2.2:
(b)
Control perimeter according to ACI 318-11 clause (a) 11.11.1.2 and (b) 11.11.1.3
2.2.2. Slabs with shear reinforcement

According to ACI 318-11 the punching strength for failure within the shear-reinforced area can
be calculated by adding the concrete and the shear reinforcement contributions, whereby the
concrete contribution generally corresponds to half of the punching strength of slabs without
shear reinforcement. However, in the case of double headed studs, ACI 318-11 (11.11.5.1)
proposes the concrete contribution as 3/4 of the punching strength of slabs without shear
reinforcement. Thus, the punching strength in slabs with stirrups is defined as:
1
6
(2.2)
11
Chapter 2
and in slabs with studs defined as:

1
4
(2.3)
where b0 is a control perimeter set at d/2 of the border of the support region, d is the effective
depth of the slab, fc is the compressive strength of concrete in MPa, Asw is the cross-sectional
area of one perimeter of shear reinforcement around the column, sw is the distance between
perimeters of shear reinforcement, and fyw is the yield strength of the shear reinforcement.
The provision for punching outside the shear-reinforced area is similar to the provision for the
punching strength of slabs without shear reinforcement. However, for this failure mode, the
control perimeter is set at a distance of d/2 from the last line of shear reinforcement and the
admissible shear stress at this perimeter is half of the one allowed in the case of punching
without shear reinforcement. Thus, the punching strength for punching outside the shearreinforced area is defined as:
1
6
(2.4)
where bout is a control perimeter set at a distance of d/2 from the last line of shear reinforcement,
d is the effective depth of the slab, and fc is the compressive strength of concrete in MPa.
Figure 2.3: Control perimeter for punching shear verification outside the shear- reinforced area
according to ACI 318-11
The maximum punching strength is defined as the multiple of the punching strength of slabs
without shear reinforcement. Generally, ACI 318-11 proposes this factor to be 1.5. However, a
factor of 2 may be used in the case of headed shear studs (ACI 318-11 11.11.5.1). It has to be
noted that in the case of an increase of the maximum punching strength the detailing rules
change. In fact, if a factor of 2 is used, the spacing between the studs is limited to 0.5d.
However, certain investigated test specimens within this research have a distance of 0.75d
12
Literature Review
between the studs and thus do not fulfill this detailing rule. Nevertheless, for the calculations
performed within this research, this restriction of the spacing is not considered. Thus, for all the
investigated specimens within this research, the maximum punching strength of specimens with
stirrups is defined as:
1
(2.5)
(2.6)
and of specimens with studs as:

2
where b0 is a control perimeter set at d/2 of the border of the support region, d is the effective
depth of the slab, and fc is the compressive strength of concrete in MPa.
2.3. Eurocode 2 (EC2 2004)

The Eurocode provision for punching without shear reinforcement is based on an empirical
formulation for the prediction of the shear strength of beams. The adjustment is mainly made by
fitting the control perimeter so that the formulation agrees well with test results. Unlike
ACI 318-11, the provision of EC2 2004 accounts for the flexural reinforcement ratio and size
effects. Thus, the punching strength is defined as:
0.18
100
(2.7)
where b0,EC is a control perimeter set at 2d of the border of the support region with circular
corners, d is the effective depth of the slab, fc is the compressive strength of concrete in MPa,
is the flexural reinforcement ratio limited to the maximum of 2%, k is a factor accounting for the
size effect that is defined as:
1
200
2.0
(2.8)
13
Chapter 2
Figure 2.4:
Control perimeter according to EC2 2004
Equation 2.7 leads to a resistance of zero if the reinforcement ratio goes to zero. Thus, it implies
a low shear strength for small reinforcement ratios. Therefore, a minimum punching shear
strength was introduced accounting only for the tensile strength of the concrete and a size effect
factor k. The minimum punching shear stress is defined as:
0.035
(2.9)

Similar to ACI 318-11, EC2 2004 proposes the summation of the concrete and the shear
reinforcement contributions, whereby the concrete contribution corresponds to 75% of the
punching strength of slabs without shear reinforcement. This reduction is made to account for
the activation of the shear reinforcement and that the concrete strength reduces due to the
vertical movement of the punching cone when the shear reinforcement is yielding (EC2
Commentary 2008). Thus, the punching strength can be calculated as:
0.75
1.5
(2.10)
where Asw is the area of one perimeter of shear reinforcement around the column, sw is the radial
spacing of perimeters of shear reinforcement, d is the effective depth, and fyw,ef is the effective
stress in the shear reinforcement accounting for limited anchorage of the shear reinforcement in
thin slabs and fyw,ef is defined as:
,
1.15 250
0.25
(2.11)
where d is the effective depth in mm and fyw is the yielding strength of the shear reinforcement
in MPa.
14
Literature Review
The punching strength for failure outside the shear-reinforced area is similarly defined as for
punching strength of slabs without shear reinforcement. The only difference to the formulation
for punching of slabs without shear reinforcement is the length of the control perimeter, which
is in this case is taken at the outer perimeter leading to the expression:
0.18
100
(2.12)
where bout is a control perimeter set at a distance of 1.5d from the outermost perimeter of shear
reinforcement. All other parameters correspond to the formulation of punching of slabs without
according to EC2 2004
For the maximum punching strength, EC2 2004 uses a similar approach as for the calculation of
the strength of the compression strut in a reinforced concrete beam. Therefore, the strength is
directly related to the concrete compressive strength, the column perimeter, and the effective
depth. Thus, the maximum punching strength is defined as:
0.3 1
250
(2.13)
where b0,in is a control perimeter set at the border of the support region, d is the effective depth
of the slab, and fc is the compressive strength of concrete.
15
Chapter 2
2.4. German National Annex (NAD 2011)

NAD 2011 is an amendment of EC2 2004 for Germany. Therefore, only the differences between
the two code provisions will be discussed. For the provision for slabs without shear
reinforcement the only difference is the assumption for factor CR,c, which depends on the ratio
of the control perimeter to the effective depth in order to improve the performance of the
provision of EC2 2004 for smaller column sizes. Thus, the punching strength is defined as:
100
(2.14)
where b0,EC is a control perimeter set at 2d of the border of the support region with circular
corners, d is the effective depth of the slab, fc is the compressive strength of concrete, is the
flexural reinforcement ratio (see Equation 2.16 for additional conditions), k, which is defined
according to Equation 2.8, is a factor accounting for the size effect, and factor CR,c, which is
defined as:
,
4:
4:
,
0.18
0.18 0.1
(2.15)
0.6
where b0,in is a control perimeter set at the border of the support region and d the effective depth.
As in EC2 2004, the flexural reinforcement ratio is limited to 2%. Additionally, NAD 2011
limits the flexural reinforcement ratio by:
0.5
0.5 0.85
1.15
1.5
2.0%
(2.16)
where c and s are partial safety factors, fc is the compressive strength of concrete, and fy the
yielding strength of the reinforcement.

With respect to the failure mode within the shear-reinforced area no additional changes have
been made. However, the reduction of VRc due to the adjusted definition of factor CR,c (Equation
2.15) has to be considered. Thus, the punching strength is defined as:
0.75
1.5
(2.17)
where Asw is the area of one perimeter of shear reinforcement around the column, sw is the radial
spacing of perimeters of shear reinforcement, d is the effective depth, and fyw,ef is the effective
stress in the shear reinforcement defined by Equation 2.11.
16
Literature Review
The provision for failure outside the shear-reinforced area corresponds to the EC2 2004
provision except for the fact that an orthogonal layout of the shear reinforcement is not allowed.
However, in this thesis, this limitation is neglected for the comparisons of the code predictions
to the experimental data presented in Chapter 4. Additionally, for the calculations within this
research, a reduced value of the effective depth dout is used as it is proposed by Hegger et al.
(Hegger et al. 2010). This reduction is due to the fact that at the outer perimeter the shear force
is not transferred to the bottom surface of the slab such as in the case of a column but to the
bottom end of the outermost shear reinforcement. Consequently, the punching strength
predicted by NAD 2011 using the proposed (Hegger et al. 2010) reduced value of the effective
depth dout is somewhat smaller than the one predicted by EC2 2004. Thus, the punching strength
is defined as:
,
100
(2.18)
reinforcement and dout is the distance between the flexural reinforcement and the bottom end of
the vertical branch of the shear reinforcement as proposed by Hegger et al. (Hegger et al. 2010).
All other parameters correspond to the formulation of punching without shear reinforcement.
It has to be noted that NAD 2011 does not allow spacing larger of 2d between vertical branches
of shear reinforcement in the outermost perimeter. However, this rule was not considered in the
comparison of the code provision to the test results in Chapter 4. In the cases in which the
spacing was larger than 2d, the external perimeter bout was reduced according to the provision of
EC2 2004 (Figure 2.5).
(a)
(b)
according to NAD 2011 and to the proposal of Hegger et al. (Hegger et al. 2010)
A clear difference exists for the code provisions for failure of the concrete strut (maximum
punching strength) between NAD 2011 and EC2 2004. While EC2 2004 uses a beam analogy to
estimate the punching strength for failure of the concrete strut, NAD 2011 uses a multiplication
17
Chapter 2
of the punching strength of slabs without shear reinforcement. Thus, the maximum punching
strength is given by:
1.4
(2.19)
The multiplication factor of 1.4 can be changed for different shear reinforcement systems if an
accreditation for the system is obtained. Thus, Chapter 4 shows additionally the results of the
calculations that were performed with a factor of 1.9 for double headed studs.
2.5. SIA 262 (SIA 262 2003)

The punching provision of SIA262 2003 for slabs without shear reinforcement is based on the
CSCT. Therefore, the punching strength depends on the slab rotation. This rotation is estimated
by the design shear load and the flexural strength of the slab. In design practice, it is sufficient
to verify that the punching shear strength VRd, calculated with the design load Vd, is larger than
the design load Vd. However, it has to be noted that the thereby calculated VRd does not
correspond to the actual punching strength of the slab. The actual punching strength is obtained
at the point where the punching strength VR corresponds to the applied load V. Thus, in order to
obtain the actual punching strength, the formulation has to be solved for V so that it equals VR.
Generally, the punching strength is defined as:
0.3
(2.20)
where b0 is a control perimeter set at d/2 of the border of the support region with circular
corners, d is the effective depth of the slab, fc is the compressive strength of concrete, and kr is
defined as:
1
0.45
0.9
1
2.2
(2.21)
with
/
0.15
(2.22)
where l is the span between the columns, mR is the flexural strength, and m0 is the moment due
to the applied load.
For inner columns, m0 can be assumed as:
8
where V is the applied shear force.
18
(2.23)
Literature Review
If the slab is designed with either a concrete with maximum aggregate sizes dg smaller than
32 mm or reinforcing steel with a yielding strength larger than 500 MPa, ry and d in Equation
2.21 must be respectively adjusted using the following two equations:
,
min
48
; 1 min
16
500
;1
(2.24)
and
min
Figure 2.7:
48
16
;1
(2.25)
Control perimeter according to SIA262 2003

With respect to the failure within the shear reinforced area, SIA262 2003 is mainly based on a
strut and tie model used in design of beams by which the inclination angle is fixed at 45. This
means that it neglects any contribution of the concrete. Thus, the punching strength is defined
as:
(2.26)
where Asw is the area of shear reinforcement intersected by the potential failure surface (conical
surface with angle 45) and fyw is the yield strength of the shear reinforcement.
19
Chapter 2
Similar to the approach of EC2 2004 and NAD 2011, for the failure at the outer perimeter,
SIA262 2003 uses the same approach as for slabs without shear reinforcement. The only
difference to the provision for slabs without shear reinforcement is that the control perimeter is
taken at a distance of 0.5dv from the outermost shear reinforcement perimeter. For the punching
strength outside the shear-reinforced area, SIA262 2003 does not use the effective depth d to
calculate the external perimeter but the distance from the flexural reinforcement to the bottom
end of the shear reinforcement to account for the difference in shear transfer, This approach is
also proposed by Hegger et al. (Hegger et al. 2010) for EC2 2004 and NAD 2011. Thus, the
punching strength is defined as:
0.3
(2.27)
reinforcement and dout is the distance between the flexural reinforcement and the bottom end of
the vertical branch of the shear reinforcement. All other parameters correspond to the
formulation of punching of slabs without shear reinforcement.
(a)
(b)
according to SIA262 2003
The provision of the failure of the concrete strut uses the assumption that the maximum
punching strength is related to the punching strength of slabs without shear reinforcement. This
is similar to the provision of NAD 2011. However, as already discussed for SIA262 2003
provisions for slabs without shear reinforcement, the predicted punching strength VR is a
function of the applied shear force V that is included in the factor kr. Therefore, the maximum
punching strength is not the proportionally increased punching strength estimated by the
formulation for slabs without shear reinforcement. Thus, the maximum punching strength is
defined as:
2
20
(2.28)
Literature Review
2.6. Model Code (MC 2011)

The Model Code (MC 2010) is like SIA262 2003 based on the CSCT. The punching strength
depends on the slab rotation, which results from the applied load and the stiffness of the slab
defined by the flexural strength. Since the punching strength depends on the applied load, the
equation has to be solved so that VRc = V, as it was described previously for SIA262 2003 in
Section 2.5. Another specialty of the MC 2010 is that different levels of approximation exist.
Level I approximation enables a fast pre-dimensioning, Level II approximation is recommended
for the typical design of new structure, Level III approximation is recommended either for
special design cases or for the analysis of existing structures, and Level IV approximation is
recommended for special design cases or for a more detailed assessment of existing structures
(Tassinari 2011). In this research, Level II and Level III approximation is used for the prediction
of the tested specimen presented herein and Level II approximation is used for the comparison
to tests from literature. For slabs without shear reinforcement, the punching strength is defined
as:
(2.29)
where b0 is a control perimeter set at d/2 of the border of the support region with circular
corners, dv is the shear-resisting effective depth of the slab, fc is the compressive strength of
concrete in MPa, and k is defined as:
1.5
1
0.9
0.6
(2.30)
where d is the effective depth in mm, is the rotation of the slab, and kdg is a factor accounting
for the influence of aggregate size defined as:
32
0.75
16
(2.31)
where dg is the maximum aggregate size in mm.
For a Level II calculation, the rotation of the slab can be estimated by:
1.5
(2.32)
where rs distance to the point where the radial bending moment is zero, d is the effective depth,
Es is the Youngs modulus of the flexural reinforcement, mR is the flexural strength, and ms the
average moment per unit length in the support strip due to the applied load.
21
Chapter 2
For inner columns, ms can be assumed as:

(2.33)
where V is the applied shear force.
In the case of a level III, the factor 1.5 can be decreased to 1.2 due to the more accurate
prediction of the average bending moment ms. Thus, the rotation for a Level III calculation can
be estimated by:
1.2
(2.34)
where rs distance to the point where the radial bending moment is zero, d is the effective depth,
Es is the Youngs modulus of the flexural reinforcement, mR is the flexural strength, and ms is
the is the average moment per unit length in the support strip determined by a linear-elastic
finite element analysis.
Figure 2.9:
Control perimeter according to MC 2010

While most punching provisions of MC 2010 are similar to SIA262 2003, the provision
regarding failure within the shear-reinforced area is different. While SIA262 2003 completely
neglects the contribution of concrete, MC 2010 takes the summation of the shear forces
transferred by the concrete and the shear reinforcement. Both values, the shear contribution of
the concrete and that of the shear reinforcement depend on the rotation of the slab accounting
for the activation of the shear reinforcement and the reduction in the concrete contribution with
increasing rotation. More information about the mechanical model behind this approach can be
found in the next subchapter explaining the CSCT in detail.
22
Literature Review
The contribution of the shear reinforcement can be calculated as the sum of the multiplication of
the cross-sectional area of the shear reinforcement within an area between a distance of 0.35dv
and dv from the column face (Figure 2.10) and the stresses in the shear reinforcement.
Therefore, the contribution of the shear reinforcement is defined as:
(2.35)
where Asw is the cross-sectional area of all the shear reinforcement intersected by the potential
failure surface (conical surface with angle 45) within a distance of 0.35dv to dv from the column
face and sw are the stresses in the shear reinforcement defined by the rotation of the slab and the
bond conditions of the shear reinforcement. The stresses in the shear reinforcement are given
by:
6
(2.36)
where Esw is the Youngs modulus of the shear reinforcement, is the rotation of the slab, fct is
the tensile strength of the concrete defining the maximum bond stress, fyw is the yield strength of
the shear reinforcement, d is the effective depth, and w is the diameter of the vertical branch of
the shear reinforcement.
(a)
(b)
Figure 2.10: Shear reinforcement considered by MC 2010
The concrete contribution can be calculated according to the provisions for punching of slabs
without shear reinforcement (Equation 2.29). Finally, the punching strength for failure within
the shear-reinforced area can be obtained by the summation of the contributions of the concrete
and the shear reinforcement.
(2.37)
Similar to other codes, MC 2010 uses the same formulation as for punching without shear
reinforcement for the calculation of the punching strength for failure outside the shearreinforced area. However, for failure outside the shear reinforced area, the control perimeter is
23
Chapter 2
set at the distance of 0.5dv,out from the outermost shear reinforcement perimeter. Thus, the
punching strength is defined as:
(2.38)
where k is defined according to Equation 2.30, fc is the concrete compressive strength, bout is a
control perimeter set at a distance of 0.5dv,out from the outermost perimeter of shear
reinforcement, and dv,out is the distance between the flexural reinforcement and the bottom end
of the vertical branch of the shear reinforcement.
(a)
(b)
according to MC 2010
With respect to the failure of the concrete strut, MC 2010 uses a similar approach as other codes
by increasing the punching strength of slabs without punching shear reinforcement by a factor.
It has to be noted that the predicted punching strength VR is a function of the applied shear force
V, which is included in the calculation of the rotation. Therefore, the maximum punching
strength is not the proportionally increased punching strength estimated by the formulation for
slabs without shear reinforcement. The maximum punching strength is defined as:
(2.39)
where factor ksys is proposed as 2.4 for slabs with stirrups and 2.8 for slabs with double headed
studs. All other parameters correspond to the formulation of punching of slabs without shear
reinforcement.
24
Literature Review
2.7. Critical shear crack theory (CSCT)

The principal hypothesis of the critical shear crack theory is that the punching strength depends
on the opening and the roughness of a critical shear crack (Muttoni 2008; Fernndez Ruiz and
Muttoni 2009; Muttoni and Fernndez Ruiz 2010). Therefore, the main parameters of the CSCT
are the rotation of the slab, the effective depth, and the aggregate size. Since the latter two
parameters are slab properties and are constant for each slab, the punching strength can be
defined as a function of the slab rotation. This function can be illustrated as a failure criterion
curve. The intersection point of the failure criterion and the flexural response of the slab (loadrotation relationship) define the punching strength of the slab (Figure 2.12).
failure criterion
load V
load-rotation
relationship
flexural
strength
VR
R
slab rotation
Figure 2.12: Calculation of the punching strength for slabs without shear reinforcement according to the
CSCT
As mentioned previously, the failure criterion accounts for the opening of a critical shear crack
and the roughness of the crack defining maximal admissible shear stresses at a control
perimeter. By considering these parameters, a semi-empirical failure criterion was developed.
The punching strength as a function of the slab rotation is defined as:
3/4
15
(2.40)
where is the slab rotation, d is the effective depth, dg0 is a reference aggregate size equal to
16 mm, dg is the maximum aggregate size, fc is the compressive strength of concrete in MPa,
and b0 is the control perimeter set at a distance of 0.5d from the support region with circular
corners.
25
Chapter 2
Figure 2.13: Control perimeter used for the CSCT
The load-rotation relationship can be assumed to be independent from the failure criterion and it
can be determined by any suitable method such as an analytical model or a non-linear finite
element analysis. Different methods will be discussed in Subchapter 2.8.

Similar to the case without shear reinforcement, the CSCT defines different failure criteria as a
function of the rotation of the slab for each failure mode if shear reinforcement is present
(Fernndez Ruiz and Muttoni 2009). Figure 2.14 shows an overview of the principal failure
modes namely (a) failure within the shear-reinforced area, (b) crushing of the concrete strut, (c)
failure outside the shear-reinforced area, and (d) flexural failure.
(a)
(b)
failure criterion punching
outside shear-reinforced area

VR
load-rotation relationship
within shear-reinforced
area (Vc+Vs)
load V
load V
failure criterion
crushing of
the concrete
failure criterion
crushing of the concrete
VR
shear reinforcement
contribution (Vs)
area (Vc+Vs)
concrete contribution (Vc)

R
R
slab rotation
slab rotation
(c)
(d)
failure criterion
punching outside
shear-reinforced area
VR

area (Vc+Vs)
load V
load V
failure criterion
failure criterion

area (Vc+Vs)
VR,flex
R
slab rotation
slab rotation
Figure 2.14: Calculation of the punching strength according to the CSCT for (a) failure within the shearreinforced area, (b) crushing of concrete, (c) failure outside the shear-reinforced area, and
(d) flexural failure
26
Literature Review
The failure within the shear reinforced area is assumed to occur when a critical shear crack
opens and strains localize. It is assumed that the concrete carries a portion of the shear force
depending on the crack opening and the roughness of the critical crack. The other portion is
carried by the shear reinforcement, which also depends on the opening of the critical crack.
Therefore, the strength can be defined as the sum of the two components:
(2.41)
Vc
concrete
contribution
Vs
shear reinforcement
contribution
Figure 2.15: Contribution of the concrete and the shear reinforcement
The concrete contribution to the punching strength can be estimated using the same hypotheses
as for slabs without shear reinforcement, according to Equation 2.40. Using the same
hypotheses, one can calculate the contribution of the shear reinforcement as the sum of the
stresses in the shear reinforcement w times the cross sectional area of the shear reinforcement
Asw:
(2.42)
It is assumed that the strains and thus the stresses depend on the opening of a critical shear crack
w, which is assumed to open proportionally to the product of the rotation and the effective
depth d. Thus, the crack width can be expressed as:
(2.43)
or as:
(2.44)
where is a constant value that is proposed to be 0.5 (Fernndez Ruiz and Muttoni 2009).
Based on the distribution of the shear reinforcement and the location and the inclination of the
critical shear crack, one can calculate the crack opening at the position of the shear
reinforcement (Figure 2.16). Within this research, the tip of the critical shear crack is assumed
to be at the column face with an inclination angle of 45 (/4) as it is proposed by Fernndez
Ruiz and Muttoni (Fernndez Ruiz and Muttoni 2009).
27
Chapter 2
Therefore, the crack width at the position of the shear reinforcement can be expressed as:
0.5
cos
(2.45)
where is the rotation of the slab and hi is the vertical distance of the crack tip to the point
where the crack crosses the shear reinforcement.
Figure 2.16: (a) Opening of the crack at an angle of 45, (b-d) stress distribution in the stud depending on
the crack opening and the bond conditions, and (e) stress distribution if no bond is
considered
In the case of smooth shear reinforcement by which no bond between reinforcement and
concrete occurs, the stresses and the strains are constant over the whole length of the shear
reinforcement (Figure 2.16e). Therefore, the strains can be integrated along the whole length of
the stud in order to obtain the crack width wi. By using this relationship and assuming a linearelastic material model, one can obtain the stresses in the shear reinforcement by:
(2.46)
where Es is the Youngs modulus of and lw is the length of the shear reinforcement.
In the case of ribbed shear reinforcement by which bond along the vertical branch occurs, the
stresses in the shear reinforcement depend on the stress at the crack and on the bond conditions.
The assumption of a perfectly-plastic bond law leads to a linear distribution of the stresses along
the length of the shear reinforcement. The difference in stress in the shear reinforcement per unit
length can be calculated by the bond force per unit length divided by the cross sectional area of
the shear reinforcement. Thus, the difference in stress can be expressed as:
(2.47)
where b is the bond strength, dw is the diameter of the shear reinforcement, and x is a unit
length.
28
Literature Review
Assuming a linear-elastic material model for the reinforcing steel the difference in strain can be
obtained by:
(2.48)
where Es is the Youngs modulus, b is the bond strength, dw is the diameter of the shear
reinforcement, and x is a unit length.
The crack width can be stated as the total deformation within the shear reinforcement:
(2.49)
where a and b are the anchorage points of the shear reinforcement and w is the strain in the
Since the length of the shear reinforcement is rather short, full anchorage is not necessarily
achieved by bond along the vertical branch but at each end of the shear reinforcement.
Therefore, the stress distribution also depends on the distance between the crack and the
anchorage provided at each end of the shear reinforcement leading to three different cases of
stress distributions that can occur as shown in Figure 2.16(b-d).
In the first case, the anchorage is on both sides performed by bond (Figure 2.16b). This case
occurs if the crack width wi is smaller than the limit:
4
(2.50)
reinforcement, and lai is the shorter distance between the crack and one end of the shear
reinforcement (lai = min(lbi; lbs)).
In this case, the stresses can be obtained from:

4
(2.51)
reinforcement, wi is the crack opening, and fyw is the yielding strength of the shear
reinforcement.
29
Chapter 2
The second case occurs if the distance between the crack and one end of the shear reinforcement
is too small to provide sufficient anchorage solely by bond. The distance at the other end
however provides sufficient anchorage length (Figure 2.16c). This case occurs if the crack width
wi is between the limits:
4
(2.52)
reinforcement, lai is the shorter distance between the crack and one end of the shear
reinforcement (lai = min(lbi; lbs)), and las is the longer distance between the crack and one end of
the shear reinforcement (las = max(lbi; lbs)).

2
(2.53)
reinforcement, wi is the crack opening, lai is the shorter distance between the crack and one end
of the shear reinforcement, and las is the longer distance between the crack and one end of the
shear reinforcement, and fyw is the yielding strength of the shear reinforcement.
The third case occurs if neither one of the distances between the crack and the ends of the shear
reinforcement is large enough in order to provide sufficient anchorage solely by bond. This case
occurs if the crack width wi is larger than the limit:
2
(2.54)
the shear reinforcement (las = max(lbi; lbs)).
(2.55)
30
Literature Review
Using these formulations, one can calculate the stresses at each vertical branch of the shear
reinforcement. However, since this approach is very sensitive to the actual position of the shear
reinforcement and even more to the distinction if a vertical branch is inside or just outside the
critical crack, the calculation method is more applicable with a smeared shear reinforcement
approach (refer to Chapter 7.4.3). Therefore, the calculations performed within this research use
a shear reinforcement ratio calculated at a control perimeter set at a distance of 0.5d from the
border of the support:
(2.56)
where nr is the number of radii of shear reinforcement, dw is the shear reinforcement diameter, s0
is the distance between the first row of shear reinforcement and the column face, s1 is the
distance between two adjacent reinforcements at the same radius, c is the side length of the
column, and d is the effective depth of the slab.
For the strains of the smeared shear reinforcement, an average value calculated at the half
distance of the critical crack can be assumed. Using this assumption, the crack opening at middistance can be estimated by:
0.5
cos
(2.57)
where is the rotation of the slab and d is the effective depth.
In the case of this smeared shear reinforcement approach only two cases of stress distribution in
the shear reinforcement have to be distinguished depending on the crack opening and the bond
conditions. The first case covers small crack openings which allow the shear reinforcement to
be activated solely by bond. Therefore, the bond length is smaller than the distance of the crack
to the end of the shear reinforcement (assumed to be lw/2). This limit can be calculated by:
4
(2.58)
where b is the bond strength, dw is the diameter of the shear reinforcement, Es is the Youngs
modulus, and lw is length of the shear reinforcement.
31
Chapter 2
For crack opening smaller than the limit (w < wlim), the stresses in the shear reinforcement can
be calculated by:
4
(2.59)
modulus, and fyw is the yielding strength of the shear reinforcement.
For crack opening equal or larger as the limit (w wlim), the stresses in the shear reinforcement
can be calculated by:
(2.60)
modulus, lw is length of the shear reinforcement, and fyw is the yielding strength of the shear
reinforcement.
Figure 2.17: (a) Opening of the crack assuming smeared shear reinforcement concentrated at middistance, (b) stress distribution in the stud if w < wlim, and (c) stress distribution in the stud if
w wlim
For failure at the outer perimeter, the CSCT uses a formulation that is similar to the formulation
for slabs without shear reinforcement. In fact, the only difference to the formulation for slabs
without shear reinforcement is that the control perimeter changes. Instead of being set a distance
of d/2 from the column face, the control perimeter is set at a distance of d/2 from the outermost
perimeter of shear reinforcement.
3/4
15
(2.61)
16 mm, dg is the aggregate size, fc is the compressive strength of concrete, dout is the distance
between the flexural reinforcement and the bottom end of the vertical branch of the shear
32
Literature Review
reinforcement, and bout is the control perimeter set at a distance of 0.5d from the outermost
perimeter of shear reinforcement.
Figure 2.18: Control perimeter for punching shear verification outside the shear- reinforced area used in
the CSCT
For the failure of crushing of the concrete strut, the CSCT uses a simplified formulation. It is
assumed that the compressive strength in the strut depends on the transverse strains. Since the
transverse strain depends on the rotation of the slab, a failure criterion as a function of the
rotation is proposed that multiplies the punching strength of slabs without shear reinforcement
by a factor , which accounts for different shear reinforcement systems. Factor of 3.0 is
proposed by Fernndez-Ruiz and Muttoni (Fernndez Ruiz and Muttoni 2009) for studs or
systems with perfect anchorage conditions (steel offcuts, headed reinforcement). For other
systems, such as stirrups or where the reinforcement is developed by bond, a value of = 2.5 is
adopted. Additionally, the CSCT considers the aggregate size dg, which accounts for the width
of the crushing zone where strains localize (Muttoni and Fernndez Ruiz 2010). Thus, the
punching strength is given by:
3/4
15
(2.62)
16 mm, fc is the compressive strength of concrete, and b0 is the control perimeter set at a
distance of 0.5d from the support region with circular corners.
If the load-rotation curve does not intersect with any of the previously mentioned failure
criterion, it is assumed that no punching failure occurs and a flexural failure can be expected.
This can be the case for slabs with low amount of flexural reinforcement.
33
Chapter 2
2.8. Load-rotation response

As discussed previously, several models limit the punching strength depending on the state of
deformation and thus depending on the slab response. Therefore, models such as the CSCT
require the prediction of the load-rotation curve in order to obtain the punching strength. This
prediction of the load-rotation curve of the slab is in the case of the CSCT independent of the
failure criterion and can be obtained by any suitable model. Generally, the models can be
separated into two groups: analytical and numerical models. The analytical models are typically
limited to axisymmetric slab element whereby the numerical models provide the use of arbitrary
shapes. The following subsections present a brief overview of the analytical and numerical
models used for the prediction of the load-rotation behavior of slabs related to punching.
Subsequently, a more detailed explanation of the Quadrilinear model proposed by Muttoni
(Muttoni 2008) is given.
2.8.1. Analytical models

In the case of a linear-elastic slab analysis, an exact solution of the slab response can be found
analytically for simple cases such as that of an axisymmetric slab. In fact, an axisymmetric slab
with an axisymmetric loading leads to a rather simple differential equation (Equation 2.63) that
can be solved by stepwise integration. However, an analytical solution can only be easily found
for linear-elastic slab behavior for which the flexural rigidity D is constant.
1
(2.63)
where the flexural rigidity D is
12 1
(2.64)
For nonlinear behavior such as it is the case of reinforced concrete, the flexural rigidity is a
function of the state of stress defined by the internal forces acting at a distance r away from the
support. However, the differential equation can still be solved by using a numerical method as it
will be explained in the next subchapter or by using simplified method in which a certain
kinematic of the slab is assumed. For this, Kinnunen and Nylander (Kinnunen and Nylander
1960) assumed a kinematic by which a slab portion outside a shear crack is regarded as rigid in
radial direction. Therefore, it is assumed that this slab portion has a constant rotation. Using a
bilinear slab behavior and the equilibrium condition at the outer slab segment, the load-rotation
curve of the slab can be calculated. Muttoni (Muttoni 2008) proposed a more refined approach
that assumes a quadrilinear slab behavior. This model (subsequently referred to it as
34
Literature Review
Quadrilinear) is considered to be the most suitable analytical model in combination with the
CSCT and it will thus be explained in more detail in Subchapter 2.8.3.
2.8.2. Numerical models

Another method to calculate the load-rotation curve is the numerical approach. Either a
numerical method can be applied to simply solve numerically the differential equation of an
axisymmetric slab considering the nonlinear slab behavior or a more sophisticated numerical
method can be used such as the finite element method to model the slab as a whole. Guandalini
(Guandalini 2005) uses a numerical approach in order to numerically solve the axisymmetric
case. For this, he calculates the forces acting on a small slab element as a function of the state of
strain in the slab. By using the equilibrium conditions within a discrete element (Figure 2.19)
and by applying the boundary conditions, he obtains the response of the axisymmetric slab.
qAi
viri
mtir
(mri+m)(ri+r)
mriri
(vi+v)(ri+r)
ri
r
mtir
Figure 2.19: Solving of axisymmetric cases by numerical integration
However, for more general geometries and loadings, an axisymmetric solution might not be
suitable. In these cases an approach based on the finite element method is more promising.
Starting in the last century, various research has been performed regarding the application of the
finite element method for reinforced concrete. Due to the enormous amount of different
developments in this area, it is neither possible nor the objective of this research to present the
development of elements used for the modeling of reinforced concrete in its entirety. Therefore,
the following sections present a selection of finite elements analyses and developments.
35
Chapter 2
Generally, the elements chosen for the analysis of flat slabs are either plate (or shell) elements
or 3D solid elements. Whereby the latter is only used for the modeling of a limited portion of
the slab (e.g. punching test specimens) since the modeling of a flat slab requires a large amount
of elements leading to an extensive need of computational resources and calculation time. The
advantage of the use of 3D solid elements is that the concrete and the reinforcement can be
modeled separately so that the modal accounts of the orthogonal reinforcement. Additionally,
the punching shear reinforcement can be implemented in the same model allowing the analysis
of the load transfer path in the column vicinity. With respect to punching of flat slabs, Beutel
(Beutel 2003) and Husler (Husler 2009) investigated several punching tests with and without
shear reinforcement by using models with 3D solid elements. The analysis led to acceptable
results regarding the response of the investigated test specimens with and without shear
reinforcement. Nevertheless, it has to be mentioned that a calculation using 3D elements is
sophisticated so that the input and model parameters have to be chosen carefully and the details
such as the anchorage of the shear reinforcement (Beutel 2003) have to be appropriately
modeled.
A simpler approach is the use of 2D elements such as plate or shell elements. However, the
main challenge by using plate or shell elements is the implementation of the nonlinear response
of a reinforced concrete section. Several methods exist that can account for the nonlinearity in
the 2D finite element analysis. One possibility is the use of elements that consist of different
layers which are separately integrated (Figure 2.20) and later assembled for the whole element.
Thus, every layer has its own stiffness matrix which depends on the state of deformation.
Formulations for this method with respect to plate and shell elements have been developed,
amongst others, by Hand et al. (Hand et al. 1973) and by Vecchio and Polak (Vecchio 1989;
Polak 1992; Polak and Vecchio 1993). The main difficulty lies in the definition of the torsional
stiffness. For this, Hand et al. (Hand et al. 1973) introduced a shear retention factor that
accounts for dowel action and aggregate interlock so that the element can provide shear
stiffness. Polak (Polak 1992) defined the torsional stiffness as a function of the concrete
stiffness in the principal directions.
Steel
stresses
Figure 2.20: Layered element
36
Concrete
stresses
Literature Review
Another possibility to introduce the material nonlinearity into the finite element method is by
using a modified stiffness approach. This method was introduced by Jofriet and McNeice
(Jofriet and McNeice 1971) and by Bell and Elms (Bell and Elms 1972). In this method, the
constitutive relationship of a reinforced concrete element is calculated in advance leading to the
secant stiffness of the element depending on the state of deformations. Afterwards, the secant
stiffness is used for a linear finite element calculation leading to a new state of deformation for
each element. This routine will be repeated until a certain tolerance is met. The main advantages
of this method are firstly the calculation speed since the element response has to be calculated
only once and secondly the robustness of the calculation. Due to the fact that the constitutive
relationship is calculated in advance, the response of the element is well defined so that for each
state of deformation, a defined stiffness exists.
Vaz-Rodrigues (Vaz Rodrigues 2007) and Tassinari (Tassinari 2011) used this method with
respect to shear test and punching test calculations, respectively. The main problem however
was the introduction of the torsional stiffness after cracking. Similar to the shear retention factor
introduced by Hand et al. (Hand et al. 1973), Vaz-Rodriques (Vaz Rodrigues 2007) introduced
the torsion retention factor to account for the torsional stiffness of the slab (Figure 2.21). For
regular punching tests Vaz Rodrigues (Vaz Rodrigues 2007) and Tassinari (Tassinari 2011)
empirically determined the torsion retention factor to be t = 1/8. Since this value is completely
empirical and thus it is only valid for the geometries and loadings according to the tests on
which it is based on, a more sophisticated formulation is desired. Chapter 5 presents a
constitutive model that allows modeling the bending and the torsional response of a reinforced
slab element.
m
mR
EI1
mi
mcr
EI0
0
0
Dx,i or Dy,i
1
i
Figure 2.21: Definition of the stiffness matrix using the torsion retention factor t (Tassinari 2011)
37
Chapter 2
2.8.3. The Quadrilinear model (Muttoni 2008)

The Quadrilinear model assumes a similar slab behavior as Kinnunen and Nylander (Kinnunen
and Nylander 1960). It is proposed to divide the slab into two parts: an inner and an outer part.
The inner core over the column has a constant curvature and moment. The outer slab segment is
assumed to undergo rigid body deformation in radial direction. Thus, the slab deforms
accordingly to a conical shape which leads to a constant slab rotation . Figure 2.22 illustrates
the kinematic of the axisymmetric slab and the forces acting on the outer slab element.
(a)
(b)
V
2
r0
-mt
-mr
rq rs
V
2
r0
rc
Figure 2.22: Assumed slab kinematic: (a) outer slab segment with acting forces and (b) deformed shape
of the slab at the column vicinity
By looking at the outer slab segment, one can define the equilibrium of moments at the section
at a radius of r0:
(a)
(2.65)
(b)
-y
r=-/r0
r
neglected
-cr
-m
-mR
-mR
-mcr
mr
t=-/r0
-y
TS
EI1
-mcr
-1 -cr
EI0
1
-mR
-mcr
mt
r0
ry
r1
rcr
-cr
-1
-y
rs
Figure 2.23: (a) Distribution of the radial and tangential moments and curvatures and (b) assumed
quadrilinear moment-curvature relationship (Muttoni 2008)
38
Literature Review
At the inner slab element the radial and tangential curvatures are assumed to be constant. Thus,
for r r0, the curvatures become:
(2.66)
At the outer slab segment (r > r0) the tangential curvature becomes:
(2.67)
In the axisymmetric slab element, this curvature distribution leads to the radial and tangential
moments. Assuming a quadrilinear moment-curvature relationship as shown in Figure 2.23b,
the moments are defined by the curvatures. The parameters for this relationship are defined as
follows:
Assuming that the influence of the reinforcement can be neglected for the phase before
cracking, the cracking moment can be estimated as:
(2.68)
12
(2.69)
The stiffness before cracking can be expressed as:
leading to the cracking curvature of:

2
(2.70)
After cracking, the stiffness of the reinforced slab specimen decreases. If a linear-elastic
behavior of the concrete and the reinforcing steel is assumed, the stiffness after cracking can be
estimated as:
(2.71)
where c is the height of the compression zone defined as:
(2.72)
where Es is the Youngs modulus of the reinforcing steel, Ec is the Youngs modulus of the
concrete, is the flexural reinforcement ratio, d is the effective depth, and is an efficiency
39
Chapter 2
factor that accounts for the orthogonal reinforcement layout. Within this research, a value of
= 0.75 is used for the calculations of the punching strength. This value will be discussed in
detail in Chapter 8 regarding the validation of the analytical model developed within this thesis.
If one assumes a rigid-plastic concrete behavior, the flexural strength can be determined as:
(2.73)
where is the flexural reinforcement ratio, d is the effective depth, fy is the yielding strength of
the reinforcement, and fc is the compressive strength of the concrete.
The consideration of a constant contribution of the tension stiffening that can be estimated as:
1
6
(2.74)
This equation leads to the curvature at the beginning of the stabilized cracked phase of:
(2.75)
and to the curvature at yielding of:
(2.76)
The different phases of the moment-curvature relationship divide the slab into four regions that
have to be integrated (Figure 2.23): The region at which the reinforcement is yielding, which
can be calculated by:
,
(2.77)
the region where the cracking is stabilized, which can be calculated by:
,
(2.78)
and the region where the concrete is cracked, which can be calculated by:
40
(2.79)
Literature Review
Using these formulations, one can obtain the shear force as a function of the slab rotation at the
outer slab segment by solving Equation 2.65:
2
ln
ln
ln
ln
(2.80)
where rq is the radius of the load application point, rc is the radius of the column, mr is the
moment in radial direction acting at radius r0, r0 is the radius of the critical shear crack, mR is the
flexural strength, ry is the radius of the yielded zone, r1 is the radius of the zone in which
cracking is stabilized, rcr is the radius of the cracked zone, rs is the radius of the slab, EI0 is the
flexural stiffness before cracking, EI1 is the flexural stiffness after cracking, TS is the decrease
in curvature due to tension stiffening, is the slab rotation at the outer slab segment, and
operator x is x for x 0 and 0 for x < 0.
41
3. Experimental Campaign
The experimental campaign within this research project consisted of sixteen full-scale slab
specimens. All slabs had the same plan dimensions (3.00 m x 3.00 m) and the same flexural
reinforcement ratio (1.5%). The investigated parameters were the column size (defined as
column size to effective depth ratio c/dn), the slab thickness (h), the amount of shear
reinforcement (w), and the type of shear reinforcement system. An overview of the test
parameters is illustrated in Figure 3.1. Various continuous measurements have been taken such
as the applied load, the slab rotation, the vertical displacements, the surface deformations, the
change in slab thickness, and the strains in the shear reinforcement. This chapter that is based on
the paper (Lips et al. 2012b) presents a summary of the tests performed within the research
project. A more detailed description of the test set-up and the measurement devices as well as
all the measured data can be found in the test reports (Lips et al. 2010; Lips et al. 2012a).
(a)
(b)
w,n [%]
1.5
Series PL
Series PF
1 PL6
PL7/9
PL10
PF2/4/5
PF1
0.5
0
0
PL1
PL6
PL9
PL7
PL8
PF4
PF1/2/3
PL8
PF3
PL12
PL12
PL11
PL11
PV1
1
Series PL
Series PF
2
c/dn [-]
PL3
3
PL1/3 PV1
100
200
300
h [mm]
PL10
PF5
PL4 PL5
400
500
Figure 3.1: Overview of the test program: variation of the c/dn ratio and (b) variation of the slab
thickness (nominal values)
43
Chapter 3
3.1. Geometry and reinforcement

All specimens had plan dimensions of 3.0 x 3.0 m and a constant flexural reinforcement ratio of
1.5%. The reinforcement layout was orthogonal and parallel to the slab edges. In a first series,
the column size c was varied (130 mm; 260 mm; 520 mm), whereas the slab thickness h was
kept constant (250 mm) leading to a constant effective depth d and thus to a variation of the
column size to effective depth ratio c/dn. In a second series, the slab thickness h was varied
(250 mm; 320 mm; 400 mm) whereas the column size to effective depth ratio c/dn was kept
constant. In a third series, the amount of shear reinforcement was varied (0.93%; 0.47%;
0.23%). In this series the thickness (h = 250 mm) and the column size (c = 260 mm) were kept
constant. The main parameters of the slab specimens are summarized in Table 3.1.
S. 3
Series 2
Series 1
Table 3.1: Main parameters of the test specimens
Specimen1
PL1
PL6
PF1
PV1
PL7
PF2
PL3
PL8
PF3
h
[mm]
250
250
250
250
250
250
250
250
250
c
[mm]
130 x 130
130 x 130
130 x 130
260 x 260
260 x 260
260 x 260
520 x 520
520 x 520
520 x 520
deff
[mm]
193
198
209
210
197
208
197
200
209
c/deff
[-]
0.67
0.66
0.62
1.24
1.32
1.25
2.64
2.60
2.49
fc,test
[MPa]
36.2
36.6
31.1
34.0
35.9
30.4
36.5
36.0
37.1
eff
[%]
1.63
1.59
1.50
1.50
1.59
1.51
1.59
1.57
1.50
PL4
320
340 x 340
267
1.27
30.5
1.58
PL9
320
340 x 340
266
1.28
32.1
1.59
PF4
320
340 x 340
274
1.24
32.5
1.54
PL5
PL10
PF5
PL11
PL12
400
400
400
250
250
440 x 440
440 x 440
440 x 440
260 x 260
260 x 260
353
343
354
201
201
1.25
1.28
1.24
1.29
1.29
31.9
33.0
33.4
34.2
34.6
1.50
1.55
1.50
1.56
1.56
fy
[MPa]
583
583
583
709
583
583
583
583
583
531 20
580 26
531 20
580 26
531 20
580 26
580
580
580
554
554
w,eff
[%]
1.01
0.79
0.93
0.79
0.85
0.79
fyw
[MPa]
519
536
519
536
519
536
Studs
Stirrups
Studs
Stirrups
Studs
Stirrups
0.93
516
Studs
0.79
550
Stirrups
0.82
0.79
0.23
0.47
563
550
592
592
Studs
Stirrups
Studs
Studs
System
PL: (Lips et al. 2012a); PF: (Lips et al. 2010); PV: (Fernndez Ruiz et al. 2010)
In addition to the geometric variation, series one and two were performed with two different
shear reinforcement systems: corrugated double-headed shear studs and cages of continuous
stirrups. Figure 3.2 illustrates the two shear reinforcement systems. The shear studs were
radially arranged whereby the number of studs in a row was determined so that no punching at
the outer perimeter could occur. The diameter and the number of radii were chosen so that the
44
Experimental Campaign
shear reinforcement ratio was approximately equal (w,eff between 0.79 and 1.01%) except for
specimen PL11 and PL12, for which a quarter and a half of the previous shear reinforcement
ratios were used (w equal to 0.23% and 0.46%, respectively). Table 3.2 shows the diameter and
number of studs used for each specimen. The cages of continuous stirrups consisted of bent bars
(10 mm) that were welded together with straight bars (6 mm) to form a cage. The spacing
between each vertical branch of the stirrups was constant 100 mm, leading to a constant shear
reinforcement ratio of w = 0.79%. One cage with dimensions 1200 x 1200 mm was placed at
the center of the slab surrounded by eight smaller cages with dimensions 600 x 600 mm to
prevent a failure at the outer perimeter of the shear-reinforced area. Table 3.3 shows the
parameters of the continuous stirrups for each specimen.
(a)
(b)
Figure 3.2: Shear reinforcement used and their distribution: (a) corrugated studs and (b) cages of
continuous stirrups (dimensions in mm)
45
Chapter 3
Table 3.2: Parameters of the corrugated studs
Specimen
PL6
PL7
PL8
PL9
PL10
PL11
PL12
s0
[mm]
80
80
80
100
130
80
80
w
[mm]
14
14
14
18
22
10
10
s1
[mm]
160
160
160
200
260
160
160
nr
[-]
12
16
24
16
16
8
16
ns
[-]
6
7
7
6
5
7
7
w
[%]
1.01
0.93
0.85
0.93
0.82
0.23
0.47
hw
[mm]
215
215
215
285
365
215
215
Table 3.3: Parameters of the continuous stirrups

Specimen
PF1-3
PF4
PF5
dt
[mm]
10
10
10
st
[mm]
100
100
100
t
[%]
0.79
0.79
0.79
ht
[mm]
200
270
345
3.2. Materials
For all specimens, normal strength concrete with a maximum aggregate size of 16 mm was
used. The compressive strength was determined on cylinders with a height of 320 mm and a
diameter of 160 mm at 14 days, 28 days, and the day of testing. For the flexural reinforcement,
hot-rolled steel with an average yielding strength ranging between 531 and 583 MPa was used.
For the punching shear reinforcement, the studs consisted of hot-rolled steel with an average
yielding strength ranging between 516 to 591 MPa, whereas the stirrups consisted of coldformed steel with an average yielding strength ranging between 536 and 550 MPa. Details can
be found in Table 3.1.
46
3.3. Test set-up

Figure 3.3a shows the test set-up for the specimen with a thickness of 250 mm. This test set-up
was already used for previous punching tests. However, for the specimen with a thickness of
320 mm and 400 mm, it had to be strengthened (Figure 3.3b). In both cases the load was applied
by four hydraulic jacks underneath the strong floor. Four tension bars running through the floor
were connected to four spreader beams that distributed the load to eight tension bars. These bars
applied the load at the top surface of the slab. The support consisted of two massive steel beams
that distributed the load on the strong floor.
(a)
(b)
Figure 3.3: Test set-up for slabs with thickness of (a) h=250 mm and (b) h=320 mm and h=400 mm
(dimensions in mm)
47
Chapter 3
3.4. Measurements
Various continuous measurements were recorded during the tests. Load cells measured the
applied load at the hydraulic jacks as well as the reaction forces at the support. Five
inclinometers circularly arranged on the top surface of the slab at a distance of 1.38 m from the
center measured the rotation of the slab (Figure 3.4b). Several Linear Variable Differential
Transducers (LVDTs) have been placed on the top and the bottom side of the slab to measure
the vertical displacements. Additionally, two LVDTs were used to measure the change of
thickness of the slab by using a small rod, which was fixed at the bottom surface, put through a
hole in the slab (8 mm) and connected to the LVDT fixed on the top surface of the slab (Figure
3.14). Omega-shaped extensometers with a measuring length of 50 and 100 mm measured the
surface deformation of the slab. They were placed axially and diagonally at the top and bottom
surface of the slab. In addition, 12 strain gauges measured the strains in the punching shear
reinforcement. After the test, the slab specimens were cut in half along the weaker axis to
analyze the punching cone in detail.
(a)
(b)
Figure 3.4: Plan view of the location of (a) the load application and (b) the rotation measurements
(dimensions in mm)
3.5. Results
All specimens failed in punching, except specimen PL8, which underwent large deformation
without failure. This test was stopped before flexural failure occurred. Table 3.4 presents the
experimentally obtained punching strength and the measured rotation along the weaker axis at
failure of all specimens. Additionally, Table 3.4 lists the normalized punching strength and the
normalized rotation at failure for each slab. The normalization of the strength accounts for
column size, depth of the member, and concrete compressive strength and the normalization of
the strength accounts for the critical shear crack width and the roughness of the crack. The
normalized results are graphically shown in Figure 3.5.
48
Table 3.4: Overview of the test results
Specimen
VR,test
R,test
PL1
PL6
PF1
PV11
PL7
PF2
PL3
PL8
PF3
PL4
PL9
PF4
PL5
PL10
PF5
PL11
PL12
[kN]
682
1363
1043
974
1773
1567
1324
2256
2242
1625
3132
2667
2491
5193
4717
1176
1633
[]
6.0
18.6
9.5
7.6
32.0
16.7
13.2
46.8
6.5
26.2
14.0
4.7
18.0
13.4
11.9
22.0
[MPa]
0.52
1.00
0.76
0.47
0.91
0.81
0.41
0.69
0.64
0.50
0.95
0.77
0.44
0.93
0.80
0.60
0.83
[-]
0.04
0.12
0.06
0.05
0.20
0.11
0.08
0.31
0.05
0.22
0.12
0.05
0.19
0.15
0.07
0.14
(Fernndez Ruiz et al. 2010)
MPa
1.5
no shear reinf.
studs
stirrups
0.5
0
0
Figure 3.5:
0.1
0.2
0.3
0.4
0.5
Normalized strength as a function of the normalized rotation for the test specimens
Figure 3.6 illustrates the crack pattern at the section along the axis after punching failure,
obtained by a saw cut after testing. In addition, pictures of the cut sections are shown in Figure
3.7. The inclination of the failure surface was rather steep for members with large amounts of
shear reinforcement and particularly for specimens with studs. For lower amounts of shear
reinforcement and for specimens with stirrup cages, the angle of the failure surface was
somewhat flatter and the critical shear crack crossed two or three rows of shear reinforcement.
Specimens with studs as shear reinforcement had basically two failure areas: spalling of the
concrete cover at the bottom side of the column and crushing of concrete at the top surface near
the anchorage of the first row of studs. In the case of stirrups, these two failure areas could still
be recognized but were less dominant than for specimens with studs.
49
Chapter 3
Figure 3.6:
50
Cracking pattern along the weak axis after punching failure
PL1
PL6
PF1
PL7
PF2
PL3
PL8
PF3
PL4
PL9
PF4
PL5
PL10
PF5
PL11
PL12
Figure 3.7: Saw-cuts along the weak axis after punching failure
51
Chapter 3
3.6. Discussion of the results

3.6.1. Performance of the shear reinforcement
Figure 3.8 shows the load-rotation curves of specimens with shear reinforcement and the
reference specimens without shear reinforcement. It can be observed that the strength and the
rotation capacity significantly increases if shear reinforcement is provided. This significant
increase occurred by all slab specimens as it is shown in Figure 3.8a for specimens with a small
c/d ratio, in Figure 3.8b for the standard specimens, and in Figure 3.8c for specimens with large
thickness. Additionally, it can be observed that the performance of a slab specimen depends
somewhat on the shear reinforcement system. The slabs with studs showed a higher strength and
a larger rotation capacity than slabs with stirrups. This difference can be seen in all the
investigated cases such as in the case of small columns (Figure 3.8a), in the case of the standard
specimen (Figure 3.8b), as well as in the case of increased thickness (Figure 3.8c).
(a)
(b)
(c)
1.2
MPa
PL6 (studs)
0.9
PL10 (studs)
PL7 (studs)
PF5 (stirrups)
PF2 (stirrups)
PF1 (stirrups)
0.6
PL1 (reference)
PV1 (reference)
PL5 (reference)
0.3
0
0
0.1
0.2
0.3
0.4
0.5 0
0.1
0.2
0.3
0.4
0.5 0
0.1
0.2
0.3
0.4
0.5
Figure 3.8: Load-rotation curves for slab specimens with and without shear reinforcement:
(a) c = 130 mm / h = 250 mm; (b) c = 260 mm / h = 250 mm; (c) c = 440 mm / h = 400 mm
3.6.2. Column size

Figure 3.9 shows the load-rotation curves for specimens supported by various column sizes. In
the case of slabs without shear reinforcement the test results follow a clear tendency. The
strength and the rotation capacity increase with larger column sizes. However, due to the larger
perimeter b0, the normalized strength is smaller in the case of larger columns (Figure 3.9a). A
similar tendency can be observed for the test specimens with studs (Figure 3.9c). However,
whereas in the case of slabs without shear reinforcement the normalized strength increases more
by decreasing column sizes, in the case of slabs with studs the increase for smaller columns is
smaller. This is even more significant in the case of stirrups, for which the normalized strength
even decreases for small columns (PF1).
52
(a)
(b)
(c)
1.2
MPa
PL6 (c=130 mm)
0.9
PL7 (c=260 mm)
PF1 (c=130 mm)

PF2 (c=260 mm)
PL8 (c=520 mm)
PF3 (c=520 mm)
0.6
PL1 (c=130 mm)
PV1 (c=260 mm)
PL3 (c=520 mm)
0.3
0
0
Figure 3.9:
0.1
0.2
0.3
0.4
0.5 0
0.1
0.2
0.3
0.4
0.5 0
0.1
0.2
0.3
0.4
0.5
Load-rotation curves for slab specimens with various column sizes: (a) without shear
reinforcement; (b) with stirrups; (c) with studs
3.6.3. Slab thickness

Figure 3.10 shows the load-rotation curve for specimens with various thicknesses. It can be seen
that since the three normalized load-rotation curves nearly coincide, the strength develops
approximately proportional to the normalization parameter b0 d and the rotation capacity
inversely proportional to the normalization parameter d. This indicates a similar influence of
size for specimens without shear reinforcement, for specimens with stirrups, and for specimens
with studs.
(a)
(b)
(c)
1.2
MPa
PL9 (h=320 mm)
0.9
PF2 (h=250 mm)

PF5 (h=400 mm)
PL7 (h=250 mm)

PL10 (h=400 mm)
PF4 (h=320 mm)
0.6
PL4 (h=320 mm)
PV1 (h=250 mm)
PL5 (h=400 mm)
0.3
0
0
0.1
0.2
0.3
0.4
0.5 0
0.1
0.2
0.3
0.4
0.5 0
0.1
0.2
0.3
0.4
0.5
Figure 3.10: Load-rotation curves for specimens with various thicknesses for (a) specimens without shear
reinforcement, (b) specimens with stirrups, and (c) specimens with studs
53
Chapter 3
3.6.4. Amount of shear reinforcement

Figure 3.11 shows the load-rotation curves for specimens with different amount of shear
reinforcement. It can be seen that the load-rotation curves of all the specimens coincide. Thus,
the flexural stiffness is not influenced by the shear reinforcement. However, the amount of shear
reinforcement has an influence on the strength and rotation capacity. Even a rather small
amount of shear reinforcement increases the strength and the rotation capacity (PL11). By
doubling the amount of shear reinforcement the punching strength and the rotation capacity can
be further increased. However, if even more shear reinforcement is added, the punching strength
does not significantly increase. This can be explained by the change in failure mode between
specimens PL12 and PL7. While PL12 had a failure within the shear-reinforced area, PL7 had a
failure due to crushing of the concrete strut close to the column. Consequently, further increase
of the amount of shear reinforcement would also lead to crushing of the concrete strut so that it
will not result in a significant higher punching strength or in a significant larger rotation
capacity. In addition, it seems that in the case of PL7 the flexural reinforcement is yielding since
the curve shows the beginning of a plateau. Therefore, the maximum punching strength is
reached and cannot be further increased without increasing the flexural capacity.
MPa
1.2
PL7 (w=0.93%)
0.9
PL12 (w=0.47%)
0.6
PL11 (w=0.23%)
PV1 (w=0.00%)
0.3
0
0
0.1
0.2
0.3
0.4
0.5
Figure 3.11: Load-rotation curve for specimens with various amount of shear reinforcement
3.6.5. Shear deformations at column face

The shear deformations at the column face cannot be measured directly. However, they can be
estimated by using certain assumptions. Firstly, it has to be assumed that all the shear
deformations occur at the column face, thus no shear deformation occurs between point B and C
shown in Figure 3.12. Additionally, the rotation is assumed to be constant between point B and
C. Both assumptions are supported by the measured displacements between point B and C.
Another assumption is that the rotation occurs within a small region at the column face so that
the extrapolated displacements at the column face calculated by the rotations are approximately
zero.
54
AB
x0
x1
Figure 3.12: Calculation of the shear deformations at the column face
Using theses assumption, the shear deformations at the column face can be estimated from the
measured displacement at the bottom side of the slab (Figure 3.12). By an extrapolation of these
displacements, one can obtain the shear deformations at the column face:
(3.1)
Figure 3.13 shows the normalized load-deformation curves for shear deformations near the
column face. By comparing the slabs with shear reinforcement (PL7 and PF2) to the reference
slab (PV1), it can be noted that large shear deformations occur only if shear reinforcement is
present. Similar to the measured load-rotation behavior, the slab with studs (PL7) leads to larger
shear deformations than the slab with stirrups (PF2). With respect to the column size, Figure
3.13b shows that the smaller the column is the larger the shear deformations are. In comparison
to specimen PL6 (c = 130 mm; c/d = 0.66), which had rather large shear deformations, specimen
PL8 (c = 520 mm; c/d = 2.60), by which no punching failure occurred, had significantly smaller
shear deformations. A difference in shear deformation occurs also for slab specimens with
different thicknesses. However, the differences are rather small. Nevertheless, it can be noted
that the stiffer the slab is the smaller the normalized shear deformations are. This is again
similar to the observed load-rotation behavior.
(a)
(c)
(b)
1.2
PL9 (h=320 mm)
MPa
PL6 (c=130 mm)
PL7 (h=250 mm)
PL7 (studs)
0.9
PL7 (c=260 mm)

PL10 (h=400 mm)
PF2 (stirrups)
PL8 (c=520 mm)
0.6
PV1 (reference)
0.3
0
0
0.01
0.02
0.03
0.04
0.05 0
0.01
0.02
0.03
0.04
0.05 0
0.01
0.02
0.03
0.04
0.05
Figure 3.13: Normalized shear deformations at the column face as a function of the normalized shear
force of selected specimens: (a) varying shear reinforcement system, (b) varying column
size, and (c) varying slab thickness
55
Chapter 3
3.6.6. Opening of the shear cracks

The measurement of the change in slab thickness close to the column can be interpreted as an
indirect measurement of the opening of the shear cracks. Figure 3.14a shows the location of the
thickness measurements whereby the exact location depends on the slab thickness. Generally, it
can be said that distance x is approximately d/2 and y between 0 and 35 mm depending on the
reinforcement layout. The exact distances for each specimen can be found in the test reports
(Lips et al. 2012a) and (Lips et al. 2010). The change in slab thickness has been measured by
using a small rod that was fixed on the bottom side, crossed the slab vertically through a hole in
the slab, and was connected to a LVDT on the top side of the slab.
(a)
(b)
Figure 3.14: Measurement of change in slab thickness: (a) location in plan view and (b) detail of the
measurement installation (dimensions in mm)
Figure 3.15 shows the curve of the change in slab thickness, or the vertical shear crack opening,
as a function of the normalized applied load. Generally, it can be noted that specimens without
shear reinforcement (PL1-PL5) experience as soon as the first shear crack opens significant
changes in thickness. In contrast, the shear reinforcement present in specimens PL6-PL12
controls the crack opening of the first shear crack, leading to a higher strength and larger crack
openings. With respect to the column size, Figure 3.15a shows that in the case of specimens
without shear reinforcement the crack opening is larger for larger column sizes. In the case of
specimens with shear reinforcement the crack opening is nearly the same for all column sizes.
With respect to the slab thickness (Figure 3.15b), it can be generally noted that the thicker the
slab is the larger the crack openings are. However, the difference between the specimens PL4 /
PL5 and PL9 / PL10 is rather small. With respect to the amount of shear reinforcement (Figure
3.15c), it can be seen that an increase in crack opening occurs between specimen PL11 and
PL12. A further increase in shear reinforcement does not seem to influence the crack opening as
it can be seen by comparing specimen PL12 and PL7. Additionally, it can be noted that the
specimen with stirrups (PF2) experiences less change in thickness than the specimen with studs
(PL7).
56
(a)
(b)
(c)
1.2
MPa
PL6 (c=130 mm)
PL9 (h=320 mm)
PL10
PL7 (h=250 mm) (h=400 mm)
PL7 (c=260 mm)
0.8
PL7 (w=0.93%)
PL12 (w=0.47%)
PF2 (w=0.79%; stirrups)
PL8 (c=520 mm)
0.6
PL11 (w=0.23%)
PL4 (h=320 mm)
PL5 (h=400 mm)
PL1 (c=130 mm)
0.4
PL3 (c=520 mm)
0.2
0
0
h [mm]
15 0
10
h [mm]
10
15 0
h [mm]
10
15
Figure 3.15: Normalized shear deformations at the column face as a function of the normalized shear
force of selected specimens: (a) varying column size, (b) varying slab thickness, and (c)
varying the amount of shear reinforcement
3.6.7. Strains in the studs

Strain gauges measured the strain in the shear reinforcement. They were placed on the first three
vertical branches at the top and bottom side of the shear reinforcement in axial and diagonal
direction (Figure 3.16a and Figure 3.18a). In order to show the different behavior of the shear
reinforcement for failure due to crushing of the concrete strut and for failure within the shearreinforced area, two different specimens were further analyzed. Specimen PL7, which had a
large amount of shear reinforcement (w = 0.93%), exhibit crushing of the concrete strut and
specimen PL12, which had only half of the amount of shear reinforcement of the first specimen
(w = 0.46%), showed a failure within the shear-reinforced area. The illustration of the strains in
the studs for each specimen at different load levels and the load-deformation curves reveal the
difference in behavior of the two investigated specimens.
(a)
(b)
JD-1/7
JD-5/11
JD-3/9
y
60% VR
80% VR
100% VR
JD-2/8
JD-4/10
JD-6/12
Figure 3.16: Strain measurements on the shear studs of slab specimen PL7: (a) instrumentation and crack
pattern and (b) measured strain for load levels of 60%, 80%, and 100% of failure load
Figure 3.16 shows the average of the axial and diagonal strain measurements in the studs for
several load levels and Figure 3.17 shows the load-strain curves for the studs in axial and
diagonal direction for slab specimen PL7. It can be seen that yielding only occurs at the top end
of the first and second stud at load levels close to the failure load. Additionally, it can be noted
57
Chapter 3
that the strains in the second row of studs are generally higher than in the first one. This leads to
the conclusion that load must have been transferred from the second row of studs directly to the
column.
(a)
(b)
2000
1800
1600
1400
V [kN]
1200
1000
800
600
400
200
0
0
[]
10
JD1
JD2
JD3
JD4
JD5
JD6
15 0
JD7
JD8
JD9
(JD10)
JD11
JD12
5
[]
10
15
Figure 3.17: Strain measurements as a function of the applied load of studs (a) in axial direction and (b)
in diagonal direction for slab specimen PL7
In comparison to specimen PL7, specimen PL12, which exhibited a failure within the shear
reinforced area, behaved slightly different. The second row of studs started yielding on the top
end of the stud whereas the strain measurements in the first row of studs on top and on the
bottom showed lower strains.
(a)
(b)
JD-1/7
JD-5/11
JD-3/9
y
60% VR
80% VR
100% VR
JD-2/8
JD-4/10
JD-6/12
Figure 3.18: Strain measurements on the shear studs of slab specimen PL12: (a) instrumentation and
crack pattern and (b) measured strain for load levels of 60%, 80%, and 100% of failure load
With respect to the strain development in the studs, Figure 3.19 shows that specimen PL12
performs similarly to specimen PL7. Generally, the strains in the studs of specimen PL12 are
larger in the second row of studs. In fact, the strains measured in that row started to increase
significantly at the load level at which the reference specimen punched (PV1: VR = 974 kN).
Afterwards, the strains in the studs further increased and the second row of studs started
yielding just before reaching the failure load (JD-3).
58
(a)
(b)
2000
1800
1600
1400
V [kN]
1200
1000
800
400
200
0
0
JD7
JD8
JD9
JD10
JD11
JD12
JD1
JD2
JD3
(JD4)
JD5
JD6
600
[]
15 0
10
[]
10
15
Figure 3.19: Strain measurements as a function of the applied load of studs in (a) axial direction and (b)
in diagonal direction for slab specimen PL12
In order to analyze the contribution of each row of studs and the concrete, the force inside the
studs was estimated and compared to the total shear force. Assuming that the force in the studs
is proportional to the strains, thus no influence of bond at the measured location are assumed,
the force in one stud can be estimated by multiplying the strain by the Youngs modulus and the
cross sectional area of the stud. Additionally, assuming that the slab behaves in a perfectly
symmetric manner, the total shear force can be calculated by multiplying the force in one stud
by the number of studs in one perimeter. Therefore, the contribution of the shear reinforcement
can be estimated by:
(3.2)
Based on Figure 3.16 and Figure 3.18, it can be assumed that only the force in the first two rows
of studs is directly transferred to the column. Forces from studs further away will be first
transferred to another stud closer to the column. Therefore, the total shear force contribution of
the shear reinforcement can be described as the sum of the shear force in the first two rows of
studs, whereby it has to be noted that for the sum the force calculated at the bottom end of the
studs of the first row and the force calculated at the top surface were used. This can be
explained by the fact that it is assumed that a portion of the load of the second stud will be
introduced to the first stud through bond stresses. In addition to the shear forces in the studs, it
is assumed that a portion of the load will be transferred by the concrete. Therefore, it can be
stated that:
, ,
, ,
(3.3)
Using Equation 3.3, the contribution of the concrete can be estimated as a function of the
rotation. Despite the rather coarse assumptions, a general behavior can be presented. Figure 3.20
shows the normalized load-rotation curve with the estimated contribution of the studs and the
59
Chapter 3
concrete. Generally, it can be noted that for low rotations the estimated contribution of the
concrete is dominant up to a certain point at which the concrete contribution starts to decrease.
In the case of specimen PL7 (Figure 3.20a) the point at which the concrete contribution starts to
decrease is approximately at the rotation at which the reference specimens without shear
reinforcement (PV1) punched. Additionally, it can be noted that after a first decay, the
contribution of concrete stays constant before it descends again with increasing rotation until
failure occurs.
A different behavior can be observed in the case of specimen PL8, which did not fail in
punching. The changes in the concrete contribution are less dominant than for specimen PL7.
However, again it can be noted that the rotation at which the contribution of the concrete is
maximal is close to the rotation at failure of the reference specimen (PL3). After the peak, the
concrete contribution decreases slowly until the point at which the test was stopped. Specimen
PL10 shows a similar behavior to specimen PL7, although the estimation shows a certain
stabilization at the level of the maximum of the estimated concrete contribution. This maximum
is again near the failure of the reference specimen.
(a)
(b)
1
MPa
0.9
PL7
PL8
0.8
V
0.7
0.6
Vw,2,top
0.5
Vw,2,top
Vw,1,bot
0.4
0.3
Vw,1,bot
0.2
Vc
0.1
(c)
Vc
0
1
0.9
(d)
PL10
PL12
MPa
0.8
0.7
0.6
Vw,2,top
0.5
Vw,1,bot
0.4
Vw,2,top
0.3
Vw,1,bot
0.2
Vc
0.1
0
0
Vc
0.1
0.2
0.3 0
0.1
0.2
0.3
Figure 3.20: Contribution of the first to rows of studs and the concrete to the normalized strength as a
function of the normalized slab rotation for slab specimen (a) PL7, (b) PL8, (c) PL10, and
(d) PL12
60
The results obtained for specimen PL12 also follows the same tendency as specimen PL7.
Again, it can be noted that the peak of the concrete contribution occurs at the rotation at which
the reference specimen reaches its punching strength. Similar to specimen PL7, specimen PL12
shows stabilization of the concrete contribution after a certain decay of concrete contribution. At
a certain rotation, the contribution of the concrete starts to decrease until punching failure
occurs.
Interesting to note is that in the case of PL12, which had a failure within the shear reinforced
area, the contribution of concrete is still at a rather high level compared to the contribution of
the shear reinforcement, unlike in the case of specimen PL7 that had a failure of the concrete
strut. This leads to the conclusion that the contribution of the concrete should not be neglected
for the failure within the shear-reinforced area. Additionally, it can be noted that the concrete
contribution is not constant and thus is influenced by the slab rotation. These observations are in
agreement with the hypotheses of the CSCT, which considers these effects in its mechanical
model.
Using the same assumption as for Equation 3.2, one can calculate the force in the studs based on
the strain measurements on the top side of the first and second row of studs. Figure 3.21 shows
the shear force at the top side of the first and second stud with respect to the total shear force as
a function of the normalized slab rotation for slab specimen PL7 and PL9. It can be seen that
after a first increase the shear force ratio is nearly constant before it increases shortly before
failure. Another interesting fact is that the sum of the calculated shear force in each stud is
larger than the total shear force. Thus, a part of the force is transferred from the second row of
studs to the first row of studs.
(a)
100
90
(b)
PL7
PL9
Vi/Vtot [%]
80
70
Vw,1,top
Vw,2,top
60
Vw,1,top
Vw,2,top
50
40
30
20
10
0
0
0.1
0.2
0.3 0
0.1
0.2
0.3
Figure 3.21: Shear force at the top side of the first and second stud with respect to the total shear force as
a function of the normalized slab rotation for slab specimen (a) PL7 and (b) PL9
61
Chapter 3
3.6.8. Deformations at the shear-critical region

On the top and bottom surface of the slab specimens, omega-shaped extensometers were placed
to measure the surface deformations. The location of each measurements device and all the
results of the measurements can be found in the test reports (Lips et al. 2010; Lips et al. 2012a).
Selected results of deformation measurements on the bottom side of the slab specimens PL1,
PL6, and PF1 show the general behavior observed in all specimens with respect to the radial and
tangential direction. The shown measurements were taken by the measurement devices placed
according to Figure 3.22.
Figure 3.22: Location of the surface deformation measurements (dimensions in mm)
Figure 3.23 shows the normalized load-deformation curve of the surface deformation at the
bottom surface of specimens without shear reinforcement (PL1), with studs (PL6), and with
stirrups (PF1). All specimens experienced similar radial deformation at the beginning with
stabilized or even decreasing strains at the load level where the reference specimens (PL1)
punched. Afterwards, the strains of the other two specimens remained approximately constant
until a load level close to punching. Just before punching failure occurred, the radial strains
increased significantly. With respect to the tangential deformation, it can be noted that the
specimens with and without shear reinforcement show the same behavior. The three curves
nearly coincide until each specimen reaches its punching strength. Interesting to note is that PL6
reached strains above 4, which could lead to softening and further on to spalling of the
concrete.
(a)
(b)
1.2
tangential deformations
radial deformations
MPa
PL6 (studs)
PL6 (studs)
0.8
PF1 (stirrups)
PF1 (stirrups)
0.6
PL1 (none)
PL1 (none)
0.4
0.2
0
-2.5
-2.0
-1.5
-1.0
r []
-0.5
-5.0
-4.0
-3.0
-2.0
-1.0
t []
Figure 3.23: Normalized load-deformation curve of the bottom surface deformations (a) in radial
direction and (b) in tangential direction
62
4. Validation of Code Provisions and the CSCT

The results from the experimental campaign presented in the previous chapter were compared to
current code provisions (ACI 318 2011; EC2 2004; NAD 2011; SIA 262 2003; MC 2011) and
to the CSCT (Muttoni 2008; Fernndez Ruiz and Muttoni 2009) to investigate the performance
of these models. The used formulations of the code provisions and the CSCT are presented in
Chapter 2. The code provisions were investigated with respect to the predicted punching
strength and the influence of several different parameters such as the slab thickness, the column
size, and the shear reinforcement ratio. For this, the punching strength was calculated
accordingly to the provisions of the previously mentioned codes as a function of the investigated
parameter. These calculations were performed using the following assumptions: the concrete
compression strength was chosen as 33.5 MPa (average of the tests), the yielding strength of the
flexural reinforcement as 575 MPa (average of the tests), the yielding strength of the shear
reinforcement as 550 MPa (average of the tests), the shear reinforcement as 0.9% if not varied,
the effective depth as 210 mm if not varied, and the ratio of the column size to the effective
depth as 1.24 if not varied. Additionally, the predicted punching strength was compared with the
experimentally obtained strength of punching tests from literature in order to investigate the
general performance of the investigated models.
63
Chapter 4
4.1. ACI 318-11 (ACI 318 2011)

4.1.1. Punching strength
In comparison to the results from the tests performed within this research project, it can be
noted that ACI 318-11 predicts somewhat conservative punching strength. Although it is a
rather simple approach, the coefficient of variation is relatively small leading to a 5%-fractile of
1.10. Figure 4.1 shows the ratio of the punching strength of the tests to the one predicted by
ACI 318-11 as a function of the shear reinforcement ratio. All points are above 1.0 showing that
ACI 318-11 leads to safe estimates of the punching strength for the performed tests.
2
ACI 318-11
Avg. = 1.28
COV = 8.4%
5% =1.10
1.8
VR,test / VR,pred []
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
0
0.5
1
w [%]
h=250 mm
h=320 mm
h=400 mm
1.5
2
Figure 4.1: Ratio of the experimentally obtained punching strength to the punching strength predicted
according to ACI 318-11 (ACI 318 2011)

Figure 4.2a shows the normalized predicted punching strength as a function of the effective
depth d. Due to the specified parameter, mainly due to the large shear reinforcement ratio, only
crushing of the concrete strut (maximum punching strength) is governing. ACI 318-11 does not
account for influences of the effective depth on the normalized punching strength. Therefore,
the calculation leads to a horizontal line for slabs without shear reinforcement and slabs with
shear reinforcement since a proportional increase is used. The distinction between studs and
stirrups using an increasing factor of 2 instead of 1.5 seems to be reasonable.
64
Validation of Code Provisions and the CSCT
4.1.3. Column size

Figure 4.2b shows the normalized predicted punching strength as a function of the ratio of the
column size to the effective depth c/d. Again the predicted failure mode is crushing of the
concrete strut. ACI 318-11 predicts smaller normalized punching strength for larger c/d ratios,
which is in good agreement with the test results although this tendency does not seem to apply
to specimens with small column sizes and stirrups (PF1: c/d = 0.62, stirrups). However, the
safety margin is large enough so that the predicted strength is still lower than the experimentally
obtained punching strength.
4.1.4. Shear reinforcement ratio

Figure 4.2c shows the normalized predicted punching strength as a function of the shear
reinforcement ratio. The failure modes are well addressed although in the case of specimen
PL12 (w = 0.47%) the experimentally obtained failure mode seems to be failure within the
shear-reinforced area, whereas the predicted failure mode is crushing of the concrete strut.
Again, the larger increase of the strength in the case of double headed studs compared to
stirrups complies with the test results.
(b)
(a)
1.5
ACI 318-11
Test (studs)
Test (stirrups)
Test (none)
1.25
MPa
(c)
1
0.75
0.5
studs
studs
studs
stirrups
stirrups
stirrups
none
none
0.25
0
0
Figure 4.2:
100
200
300
d [mm]
400
500
0.5
1.5 2 2.5
c/d [-]
3.5
4 0
0.25
0.5
0.75
w [%]
1.25
Punching strength according to ACI 318-11 (ACI 318 2011) as a function of (a) the
effective depth, (b) the column size to effective depth ratio, and (c) the shear reinforcement
ratio
65
Chapter 4
4.2. Eurocode 2 (EC2 2004)

4.2.1. Strength prediction
In comparison to the tests performed within this research project, it can be noted that EC2 2004
may lead to unsafe design. Due to the rather low average and a large coefficient of variation, the
5%-fractile is only 0.76. Several factors contribute to this rather low value. Firstly, EC2 2004
does not distinguish between the different shear reinforcement systems since the provision
applies implicitly only for shear links. This fact leads to scattered results. Secondly, as it is
explained later, the provision regarding the maximum punching strength is not well addressed,
which can lead to an overestimation of the punching strength.
2
EC 2 : 2004
Avg. = 1.00
COV = 14.6%
5% =0.76
1.8
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
0
0.5
1
w [%]
h=250 mm
h=320 mm
h=400 mm
1.5
2
according to EC2 2004

depth d. While the strength is well predicted for specimens without shear reinforcement and for
specimens with studs, the provision overestimates the strength for specimens with stirrups. This
is even more pertinent considering the fact that EC2 2004 implicitly applies only for shear links
and not for studs. It clearly shows that a distinction between double headed studs and stirrups
could increase the accuracy of the strength predictions. Another interesting fact is that
EC2 2004 accounts for size effects in the case of slabs without shear reinforcement (decrease of
normalized strength with increase of the effective depth) but not in the case of slabs with shear
reinforcement that fail due to crushing of the concrete strut.
4.2.3. Column size

column size to the effective depth c/d. In the case of specimens without shear reinforcement, the
strength is generally well predicted. However, it overestimates the strength for small column
66
sizes, which could lead to unsafe design. In the case of specimens with shear reinforcement, it
can be clearly seen that the beam analogy used in the EC2 2004 provision does not work. It
underestimates the strength for small column sizes and the strength increases rapidly with
increasing column size so that it overestimates the strength for c/d ratio of 1.5. With further
increase of the column size, the predicted strength would increase even more, if it was not
limited by other predicted failure modes such as failure within the shear-reinforced area or as in
this case by the flexural strength. Consequently, this could lead to unsafe design, especially in
the case of large column sizes and slabs with large flexural and shear reinforcement ratios.

reinforcement ratio. For the failure within the shear-reinforced area, the strength is well
predicted. However, again in the case of failure of the concrete strut and the specimen with
stirrups, the strength is overestimated.
(b)
(a)
1.5
EC 2 : 2004
Test (studs)
Test (stirrups)
Test (none)
MPa
1.25
1
(c)
failure within
shear-reinforced area
failure of
concrete
crushing
studs or
stirrups
studs or
stirrups
0.75
0.5
studs or
stirrups
lexural
failure
none
none
0.25
0
0
100
200
300
d [mm]
Figure 4.4:
400
500
0.5
1.5 2 2.5
c/d [-]
3.5
4 0
0.25
0.5
0.75
w [%]
1.25
Punching strength according to EC2 2004 as a function of (a) the effective depth, (b) the
column size to effective depth ratio, and (c) the shear reinforcement ratio
67
Chapter 4
4.3. German National Annex to Eurocode 2 (NAD 2011)

The approach adapted by NAD 2011 for the calculation of the maximum punching strength
leads to a higher average of the ratio of the experimentally to the predicted strength than
EC2 2004 does. However, the coefficient of variation is around 15% leading to a 5%-fractile
below 1.0. One reason for the coefficient of variation of around 15% is again that no distinction
between different shear reinforcement systems was made since the code provision applies
implicitly only for shear links. Another reason for this variation is that the average of specimens
without shear reinforcement is 1.01 and the average of the specimens with shear reinforcement
is 1.25. Therefore, the scatter results from the different safety margins of the approaches used
for slabs with and without shear reinforcement. The statistical values will show better results if
the specimens with and the specimens without shear reinforcement are analyzed separately. This
would lead to a coefficient of variation of only 4% for specimens without shear reinforcement
and 12% for specimens with shear reinforcement, whereby the 5%-fractile increases to 0.94 and
1.01, respectively. A further improvement can be obtained, by using different factors relating
the punching strength of slabs without shear reinforcement to the maximum punching strength
for double headed studs and for stirrups as it is used in practice. Figure 4.5 shows the
comparison to the test results for a constant factor of 1.4 (Figure 4.5a) and by using a factor of
1.4 for stirrups and 1.9 for double headed studs (Figure 4.5b).
(a)
(b)
2
1.6
1.4
1.2
1
0.8
0.6
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
0
NAD : 2011
Avg. = 1.07
COV = 9.4%
5% =0.90
1.8
NAD : 2011
Avg. = 1.14
COV = 14.6%
5% =0.89
1.8
0.5
1
w [%]
h=250 mm
h=320 mm
h=400 mm
1.5
2
0.4
0.2
0
0
0.5
1
w [%]
h=250 mm
h=320 mm
h=400 mm
1.5
2
(a) according to NAD 2011 and (b) according to NAD 2011 with adjusted factor for double
headed studs (stirrups: 1.4 / double headed studs: 1.9)

depth d. Since the formulation for specimens without shear reinforcement is the same as in
EC2 2004, the prediction of the strength corresponds well to the experimentally obtained
68
strength as it was already shown previously. Due to the adjusted approach for the failure of the
concrete strut, the predictions are always on the safe side for specimens with shear
reinforcement. However, especially when studs are used and the prediction of the strength is
calculated with the factor used in the code (factor: 1.4), the predictions are rather conservative
as it can be seen by comparing the test results of the studs to the predicted strength curve for
stirrups. Again, it can be noted that using a larger factor for double headed studs (factor: 1.9)
improves the predictions as it can be seen by comparing the test results of the studs to the
predicted strength curve for studs.
4.3.3. Column size

column size to the effective depth c/d. In comparison to the approach of EC2 2004, it can be
seen that the adapted provision in NAD 2011 for the maximum punching strength leads to better
predictions of the punching strength. The predictions are always on the safe side. Moreover, due
to the adjusted C factor for small c/d ratios, the provision accounts for the lower normalized
strength in the case of small column sizes.

Figure 4.6b shows the normalized predicted punching strength as a function of the shear
reinforcement ratio. For low amounts of shear reinforcement the predictions are the same as for
EC2 2004 since the same formulations were used. However, the change in failure mode is
predicted at lower shear reinforcement ratios as other codes (ACI 318 2011; EC2 2004) do. This
may lead to a different prediction of the failure mode and thus to more conservative predictions.
(b)
(a)
1.5
NAD : 2011
Test (studs)
Test (stirrups)
Test (none)
1.25
MPa
(c)
1.9
(studs)
1.9
(studs)
0.75
1.9
(studs)
1.4
(stirrups)
1.0
(none)
1.4
(stirrups)
1.0
(none)
0.5
0.25
0
0
Figure 4.6:
100
200
300
d [mm]
400
500
0.5
1.5 2 2.5
c/d [-]
3.5
1.4
(stirrups)
4 0
0.25
0.5
0.75
w [%]
1.25
Punching strength according to NAD 2011(including the predictions calculated with the
adjusted factor for studs) as a function of (a) the effective depth, (b) the column size to
effective depth ratio, and (c) the shear reinforcement ratio
69
Chapter 4
4.4. SIA 262 (SIA 262 2003)

In comparison to the results from the tests performed within this research project, it can be
noted that SIA262 2003 predicts a somewhat conservative punching strength. Additionally,
SIA262 2003 leads to scattered results, mainly due to the fact that it does not distinguish
between different shear reinforcement systems and that it uses a relatively low increasing factor
of 2.0. Furthermore, SIA262 2003 neglects any concrete contribution for the provision
regarding failure within the shear-reinforced area leading to enormously conservative values for
slabs with low amounts of shear reinforcement. These two conditions lead to a rather large
scatter and thus to a rather large coefficient of variation of 15% resulting in a 5%-fractile around
1.0 despite the large average value of 1.31.
2
SIA 262 : 2003
Avg. = 1.31
COV = 14.7%
5% =0.99
1.8
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
0
0.5
1
w [%]
h=250 mm
h=320 mm
h=400 mm
1.5
2
according to SIA262 2003

depth d. While the provision for slabs without shear reinforcement leads to a good agreement
with the test results, the provision of the maximum punching strength leads to rather
conservative results. Since no distinction between different shear reinforcement systems is
made, the difference between the predicted and measured values is, especially in the case of
slabs with studs, relatively large.
70
4.4.3. Column size

column size to the effective depth c/d. Again, the provision for slabs without shear
reinforcement leads to a good agreement with the test results. As noted previously, the provision
of the maximum punching strength leads generally to conservative predictions. However, in the
case of stirrups and small column sizes, the low factor relating the punching strength of slabs
without shear reinforcement to the maximum punching strength of 2.0 is justified and can thus
be seen as a lower limit.

reinforcement ratio. For the failure within the shear-reinforced area, SIA262 2003 neglects the
concrete contribution leading to low estimates of strength for slabs with low amounts of shear
reinforcement. In this case, the punching strength calculated with the formulation for slabs
without shear reinforcement needs to be considered. Furthermore, since the maximum punching
strength is rather low when using a factor of 2.0 relating the punching strength of slabs without
shear reinforcement to the maximum punching strength, the prediction of failure within the
shear-reinforced area is restricted to a small range of shear reinforcement ratios (w = 0.40%0.65%). This clearly leads to an underestimation of the punching strength.
(b)
(a)
1.5
SIA 262 : 2003

Test (studs)
Test (stirrups)
Test (none)
1.25
MPa
(c)
1
0.75
studs or
stirrups
studs or
stirrups
studs or
stirrups
none
0.5
none
0.25
0
0
Figure 4.8:
100
200
300
d [mm]
400
500
0.5
1.5 2 2.5
c/d [-]
3.5
4 0
0.25
0.5
0.75
w [%]
1.25
Punching strength according to SIA262 2003 as a function of (a) the effective depth, (b) the
column size to effective depth ratio, and (c) the shear reinforcement ratio
71
Chapter 4
4.5. fib Model code (MC 2011)

Regarding the results from the tests performed within this research project, it can be noted that
MC 2010 predicts safe estimates of the punching strength. The ratios between the
experimentally obtained and the predicted values are somewhat scattered. The scatter results
mainly from the rather low estimates of the punching strength for failure within the shearreinforced area. This failure mode is also predicted in the cases, for which a failure due to the
crushing of the concrete strut was observed (e.g. PL6). This effect is observed for a Level II and
a Level III calculation. Generally, it can be noted that a Level II and a Level III calculation lead
to nearly the same results for all the test specimens.
(a)
(b)
2
1.8
1.6
1.4
1.2
1
0.8
0.6
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
0
MC 2010
Level III
Avg. = 1.29
COV = 12.2%
5% =1.03
1.8
MC 2010
Level II
Avg. = 1.28
COV = 12.2%
5% =1.02
0.5
1
w [%]
h=250 mm
h=320 mm
h=400 mm
1.5
2
0.4
0.2
0
0
0.5
1
w [%]
h=250 mm
h=320 mm
h=400 mm
1.5
2
according to MC 2010 (a) Level II and (b) Level III

Figure 4.10a and Figure 4.11a show the normalized predicted punching strength as a function of
the effective depth d for a MC 2010 Level II and Level III calculation, respectively. It can be
noted that both calculation methods show good agreement with the test results for specimens
without shear reinforcement. In the case of specimens with shear reinforcement, both
calculation methods show good agreement for effective depth smaller than 300 mm. However,
for effective depths larger than 300 mm the predicted failure modes changes from crushing of
the concrete strut to failure within the shear-reinforced area leading clearly to an
underestimation of the punching strength. This is mainly due to the formulation regarding the
activation of the shear reinforcement. For thicker and thus stiffer slabs, the stresses, which
depend on the slab rotation, are relatively small in the shear reinforcement according to the
provision since it assumes that the shear reinforcement is not fully activated. This assumption
permits only the consideration of small stresses in the shear reinforcements and thus leads to a
minor increase of strength due to the shear reinforcement (VR,s).
72
4.5.3. Column size

Figure 4.10b and Figure 4.11b show the normalized predicted punching strength as a function of
the ratio of the column size to the effective depth c/d for a MC 2010 Level II and Level III
calculation, respectively. Generally, the Level II and the Level III calculations show good
agreement with the experimentally obtained punching strength for specimens with and without
shear reinforcement. Similar to thick slabs, the predicted failure mode changes in the case of
slabs with smaller c/d ratios. This change limits the normalized strength for smaller columns,
which correspond to the test results for specimens with stirrups. For slabs with studs it may lead
to an underestimation of the punching strength.

Figure 4.10c and Figure 4.11c show the normalized predicted punching strength as a function of
the shear reinforcement ratio for a MC 2010 Level II and Level III calculation, respectively. It
can be seen that the provision for failure within the shear-reinforced area leads to conservative
estimates of the punching strength. This is mainly due to the fact that only shear reinforcement
within a band between 0.35d to d around the column is considered. In the performed tests only
one perimeter of studs is within this zone and is thus considered in the calculation. A calculation
with a shear reinforcement ratio (smeared shear reinforcement) would clearly increase the
predicted strength and therefore would correspond better to the test results.
(b)
(a)
1.5
MC 2010 L-II
Test (studs)
Test (stirrups)
Test (none)
1.25
MPa
(c)
1
0.75
ksys=2.8
ksys=2.8
ksys=2.4
ksys=2.4
0.5
ksys=2.8
ksys=2.4
ksys=1.0
ksys=1.0
0.25
0
0
100
200
300
d [mm]
400
500
0.5
1.5 2 2.5
c/d [-]
3.5
4 0
0.25
0.5
0.75
w [%]
1.25
Figure 4.10: Normalized punching strength according to MC 2010 Level II as a function of (a) the
ratio
73
Chapter 4
(b)
(a)
1.5
MC 2010 L-III
Test (studs)
Test (stirrups)
Test (none)
1.25
MPa
(c)
1
0.75
ksys=2.8
ksys=2.4
ksys=2.8
ksys=2.4
ksys=2.8
ksys=2.4
0.5
ksys=1.0
ksys=1.0
0.25
0
0
100
200
300
400
500
d [mm]
0.5
1.5 2 2.5
c/d [-]
3.5
4 0
0.25
0.5
0.75
1.25
w [%]
Figure 4.11: Normalized punching strength according to MC 2010 Level III as a function of (a) the
ratio
4.6. Critical shear crack theory (CSCT)

In comparison to the tests performed within this research project, it can be noted that the CSCT
predicted strength agree very well with the experimentally obtained strength (Figure 4.12a). The
average is close to 1.0 and the coefficient of variation is small (6.8%). Only the strength of the
specimen with stirrups and a small column size (PF1) is somewhat overestimated. However,
with respect to the rotation, it can be seen that the CSCT generally underestimates the rotations
(Figure 4.12b). Additionally, it can be noted that the ratio between the predicted and the
experimentally obtained rotation lead to scattered results. In order to address this observation,
the predicted and the measured load-rotation curves will be further analyzed subsequently.
(a)
(b)
2
3
CSCT : V
Avg. = 0.99
COV = 6.8%
1.8
1.4
1.2
1
0.8
0.6
0.4
0.2
0
0
2.5
R,test / R,pred []
1.6
0.5
1
w [%]
h=250 mm
h=320 mm
h=400 mm
1.5
2
CSCT :
Avg. = 1.57
COV = 27.2%
2
1.5
1
0.5
0
0
0.5
1
w [%]
h=250 mm
h=320 mm
h=400 mm
1.5
2
Figure 4.12: Ratio of (a) the experimentally obtained punching strength to the punching strength
predicted according to the CSCT and (b) the experimentally obtained rotation at failure to
the rotation at failure predicted according to the CSCT
74

Figure 4.13a shows the predicted punching strength as a function of the effective depth d. It can
be seen that the CSCT predicts the strength very well for specimens with and without shear
reinforcement. Moreover, the difference in strength between specimens with studs and with
stirrups is well addressed. It is interesting to note that for specimens without shear
reinforcement the predicted normalized strength slightly increases with decreasing thickness,
which corresponds to the test results. On the other hand, in the case of specimens with shear
reinforcement, the predicted normalized strength slightly increases with decreasing thickness for
large column sizes until a certain thickness, at which the predicted normalized strength
decreases with decreasing thickness. This can be explained by the fact that in the case of smaller
slab thicknesses, and thus lower stiffness, the model predicts yielding of the flexural
reinforcement leading to a smaller normalized strength.
4.6.3. Column size

Figure 4.13b shows the predicted punching strength as a function of the ratio of the column size
to the effective depth c/d. Generally, the CSCT leads to good agreement with the test results. It
accounts for the increase in normalized strength with decreasing column size for slabs with and
without shear reinforcement. However, for extremely small column sizes the normalized
strength may decrease in certain cases. For example, for specimen PF1 with stirrups and a
column size of c = 130 mm, the experimentally obtained normalized strength is considerably
lower than the predicted one.

reinforcement ratio. Again, the normalized strength predicted by the CSCT shows good
agreement with the test results. The increase in strength with increasing shear reinforcement
ratio is well predicted as it corresponds very well with the experimental results. Moreover, the
change in failure mode is accurately predicted. Again, the distinction between the different
shear reinforcement systems is properly addressed.
75
Chapter 4
(b)
(a)
1.5
CSCT
Test (studs)
Test (stirrups)
Test (none)
1.25
MPa
(c)
1
studs
stirrups
0.75
studs
stirrups
studs
stirrups
0.5
none
none
0.25
0
0
100
200
300
d [mm]
400
500
0.5
1.5 2 2.5
c/d [-]
3.5
4 0
0.25
0.5
0.75
w [%]
1.25
Figure 4.13: Normalized punching strength according to the CSCT as a function of (a) the effective
depth, (b) the column size to effective depth ratio, and (c) the shear reinforcement ratio
4.6.5. Prediction of the slab response and failure criteria

As discussed previously, while the punching strength is well predicted with a low coefficient of
variation (Figure 4.12a), the predicted rotations are not only generally lower than the
experimentally obtained rotation but lead also to scattered results compared to the
experimentally obtained rotations (Figure 4.12b). This indicates an overestimation of the
stiffness of the slab specimens by the CSCT. In order to address this issue, the measured loadrotation curves are compared to the predicted load-rotation curves and the failure criteria.
Figure 4.14 illustrates the normalized load-rotation curves for specimens without shear
reinforcement. Although the predicted load-rotation curve behaves slightly stiffer than the
measured load-rotation curve, they generally correspond well. It can be noted that the difference
of the predicted and measured curves are larger for specimens for which higher strength is
obtained, thus for specimens with larger column sizes (PL3) or larger thicknesses (PL4 and
PL5). However, even for larger strength, the agreement to the test results is still acceptable. This
corresponds also to Figure 4.12b which shows that for specimens without shear reinforcement
the ratios between the predicted to the experimentally obtained rotation are within a small range.
In fact, if only the specimens without shear reinforcement are considered, one can obtain an
average value of 1.27 and a coefficient of variation of only 7%. This leads to the conclusion that
this issue needs only be addressed in the case of large shear forces as it is the case of slab
specimens with shear reinforcement.
76
(a)
(b)
1
(c)
PL1
PV1
PL3
MPa
0.9
0.8
Quadrilinear
0.7
Quadrilinear
Quadrilinear
0.6
0.5
measured
measured
measured
0.4
0.3
Vc
Vc
0.2
Vc
0.1
0
(d)
(e)
PL4
0.05
0.1
0.15
0.2
PL5
MPa
0.9
0.8
Quadrilinear
Quadrilinear
0.7
0.6
0.5
measured
measured
0.4
0.3
Vc
0.2
Vc
0.1
0
0
0.05
0.1
0.15
0.2 0
0.05
0.1
0.15
0.2
Figure 4.14: Measured normalized strength, calculated normalized strength (Quadrilinear), and failure
criterion (Vc) as a function of the normalized rotation for specimens without shear
reinforcement (a-e)
Figure 4.15 illustrates the load-rotation curves for specimens with shear reinforcement. The
trend mentioned previously is clearly visible. The slab specimens with shear reinforcement
show not only a less stiff behavior than predicted but also reach their flexural strength much
below the theoretical flexural capacity. This reduced stiffness leads to the underestimated
rotations and to the scattered results as shown in Figure 4.12b. Another aspect of the difference
in the load-rotation behavior of the test specimens compared to the predictions is that the
intersection point between the slab behavior and the failure criterion is far from being close to
the actual measured failure. However, due to the fit of the value that was chosen as 3 for studs
and 2.5 for stirrups, the strength can be predicted accurately. Nevertheless, it has to be reminded
that neither the predicted load-rotation curve nor the failure criterion represents the actual slab
behavior accurately.
77
Chapter 4
(a)
(b)
1.5
(c)
PL6
PL7
PL8
MPa
Quadrilinear
Quadrilinear
1
Vc+Vs
measured
0.5
Vc+Vs
measured
Vc
Vc
Vc
Vc
Vc+Vs
Quadrilinear
measured
Vc
Vc
(d)
(e)
1.5
(f)
PL9
PL10
Quadrilinear
MPa
Quadrilinear
Quadrilinear
Vc+Vs
PL11
Vc+Vs
measured
measured
measured
0.5
Vc
Vc
Vc
Vc
Vc
Vc+Vs
Vc
(g)
(h)
1.5
(i)
PL12
PF1
PF2
Quadrilinear
MPa
Quadrilinear
Quadrilinear
1
Vc+Vs
measured
measured
measured
Vc+Vs
Vc
0.5
Vc+Vs
Vc
Vc
Vc
Vc
Vc
(j)
(k)
1.5
(l)
PF3
PF4
PF5
Quadrilinear
MPa
Quadrilinear
1
Quadrilinear
Vc
Vc
0
0
0.1
0.2
measured
measured
measured
0.5
Vc+Vs
Vc+Vs
Vc+Vs
0.3 0
0.1
0.2
Vc
Vc
Vc
Vc
0.3 0
0.1
0.2
0.3
Figure 4.15: Measured normalized strength, calculated normalized strength (Quadrilinear), concrete
contribution for failure criterion (Vc), failure criterion for failure within the shear-reinforced
area (Vc+Vs), and failure criterion for crushing of the concrete strut (Vc) as a function of the
normalized rotation for specimens with studs (a-g) and stirrups (h-l)
78
4.7. Overview of the performance of the codes

In order to give an overview of the general performance of the code provisions and the CSCT,
the results obtained by them were compared to the results from 132 tests found in literature
whereby the selection consists of 28 tests of specimens without shear reinforcement and 104
tests of specimens with shear reinforcement. The selection of test specimens without shear
reinforcement consists only of specimens that were used as reference specimens in series
investigating slabs with shear reinforcement. More details of the test specimens used in this
investigation and the results of the calculations can be found in Appendix A.
For the comparison of the different approaches, the ratio of the experimentally obtained
punching strength to the predicted strength was calculated with each approach. From these
results, statistical values such as the average and the coefficient of variation (COV) were
determined for different sets of specimens. Additionally, the 5%-fractile for an assumed normal
distribution with corresponding average and standard deviation was determined for the code
provisions. Since the CSCT is not a code provision and thus does not have any safety margin in
its formulation, the 5%-fractile has no meaning and is therefore not shown. After the statistical
analysis, the ratios of the experimentally obtained punching strength to the predicted strength
will be illustrated with respect to different parameters namely the shear reinforcement ratio, the
effective depth, the column size to effective depth ratio, the flexural reinforcement ratio, and the
concrete compressive strength. These comparisons reveal the advantages and the drawbacks of
each code provision and the CSCT.
If all specimens are considered, all the codes except MC 2010 and the mechanical model of the
CSCT lead to similar results as it can be seen in Table 4.1. Only EC2 2004 has due to the low
average a rather low 5%-fractile value of only 0.82. In contrast, MC 2010 is the only code that
has a 5%-fractile close to 1.0 mainly due to the low coefficient of variation of 11.6%, which is
by far the lowest of all investigated code provisions. Since MC 2010 is based on the CSCT, the
CSCT shows a similar low coefficient of variation of 10.8%. Additionally, the CSCT agrees
well with the test results leading to an average value close to 1.0. Although this statistical
analysis gives a good general overview of the performance of the different approaches, a
differentiation of specimens with and without shear reinforcement will lead to a more detailed
overview of the performance of each code and the CSCT.
Table 4.1:
# specimens
Average
COV
5% fractile
1
Statistical analysis of all investigated tests from literature
ACI2011
132
1.37
19.7%
0.93
EC2004
132
1.09
14.8%
0.82
NAD2011
132
1.22 / 1.181
15.2% / 16.1%1
0.92 / 0.871
SIA2003
132
1.37
20.3%
0.91
MC2010
132
1.24
11.6%
1.00
CSCT
132
0.98
10.8%
-
with consideration of the increased factor for specimens with studs (1.9 instead of 1.4)
79
Chapter 4
Table 4.2 shows the statistical analysis for the investigated tests divided into specimens with
and without shear reinforcement. In the case of specimens without shear reinforcement, it can be
noted that all the models lead to similar scatter except ACI 318-11 that shows much more
scatter. This is mainly due to the fact that it does not consider the influence of the flexural
reinforcement ratio. Although this simplified design approach leads to scatter compared to test
results, it was reasonably justified by Alexander and Hawkins (Alexander and Hawkins 2005)
who noted that the code should not predict the punching strength accurately but define a lower
limit. In comparison to ACI 318-11, EC2 2004 and NAD 2011 have a much smaller coefficient
of variation. However, due to the average close to 1.0, the 5%-fractile for EC2 2004 and
NAD 2011 is smaller than the one for ACI 318-11. It has to be noted that EC2 2004 and
NAD 2011 have nearly the same formulation for specimens without shear reinforcement except
for coefficient C, which is constant in EC2 2004 whereas it depends on the column size in
NAD 2011. SIA262 2003 and MC 2010 have nearly the same average and coefficient of
variation since they both are based on the CSCT.
Table 4.2:
Specimens
considered
Specimens without
shear reinforcement
Specimens with
shear reinforcement
Statistical analysis of tests from literature for specimens with and without shear
reinforcement
# specimens
Average
COV
5% fractile
# specimens
Average
COV
5% fractile
ACI2011
EC2004
NAD2011
SIA2003
MC2010
CSCT
28
1.25
16.7%
0.91
104
1.40
19.7%
0.95
28
1.03
12.0%
0.82
104
1.10
15.1%
0.83
28
1.03
11.7%
0.84
104
1.27 / 1.221
13.1% / 15.2%1
1.00 / 0.911
28
1.12
7.9%
0.98
104
1.44
19.1%
0.99
28
1.17
7.8%
1.02
104
1.25
12.0%
1.01
28
0.95
8.4%
104
0.99
11.1%
-
With respect to the specimens with shear reinforcement, it can be noted that while MC 2010 and
CSCT show approximately a small coefficient of variation, the other codes show a rather large
scatter of the results. However, due to a sufficient high average, ACI 318-11, SIA262 2003, and
NAD 2011 lead to a 5%-fractile close to 1.0. In contrast, EC2 2004 has a low average leading
together with a rather high coefficient of variation to a 5%-fractile well below 1.0. The
difference between EC2 2004 and NAD 2011 is significant although they use nearly the same
formulations. The main modification of NAD 2011 compared to EC2 2004 is the approach for
the calculation of the punching strength for the failure of the concrete strut. Interesting to note is
that the calculation according to NAD 2011 with consideration of a larger factor for the increase
in the punching strength for specimens with studs (factor = 1.9), the statistical values seem to be
worse compared to the calculations according to NAD 2011 with the standard factor
(factor = 1.4). However, this can be explained by the fact that the change in the factor lead only
to an improvement of the prediction for specimens with studs that have a prediction of failure
80
due to crushing of the concrete strut whereas the prediction of slabs with other shear
reinforcement systems and failure modes do not change. Therefore, the average of a certain set
of tests is reduced which leads to a larger scatter of the results when all specimens are
considered. If the different failure modes are considered, this scatter in the results is less
apparent. In order to address the predictions for the different failure modes in more detail, the
statistical values were calculated for each model with respect to the predicted failure mode.
Table 4.3 presents the average, the coefficient of variation, and the 5%-fractile for each
approach depending on the failure mode predicted by the corresponding model. Therefore, the
number of specimens with a certain predicted failure mode is different for each approach. It can
be noted that generally one failure mode dominates. This dominant failure mode has the largest
average and in most cases the largest coefficient of variation. For example, ACI 318-11 has for
failure modes crushing of concrete strut and failure within the shear-reinforced area an average
of around 1.3 and a coefficient of variation of around 13%. For failure outside the shearreinforced area however, the average is nearly 1.5 and the coefficient of variation at 20%. This
raises the question if the ACI 318-11 provisions address this failure mode correctly.
In the case of SIA262 2003 the dominant failure mode is failure within the shear-reinforced
area. While the averages of the other failure modes are around 1.25, the average for failure
within the shear-reinforced area is above 1.5. This can be explained by the fact that
SIA262 2003 considers solely the contribution of the shear reinforcement in the formulation of
this failure mode leading to conservative estimates. Consequently, this failure mode becomes
dominant with a rather large average value.
In contrast to ACI 318-11 and SIA262 2003, EC2 2004 shows the opposite tendency. The
critical failure mode is in the case of crushing of the concrete strut although it has the lowest
average. However, while the coefficient of variation is 11% and 14% for the other failure
modes, it is 18% for failure of the concrete strut, leading to a 5%-fractile of only 0.78.
Therefore, it can be concluded that this code provision may overestimate the punching strength
for this mode of failure. In contrast, the statistical values for this failure mode are significantly
improved using NAD 2011 resulting in a 5%-fractile above 1.0 due to the modified formulation
for crushing of the concrete strut. Moreover, the consideration of a reduced effective depth for
the calculation of the punching strength for failure outside the shear-reinforced area leads to a
larger average resulting in a 5%-fractile above 1.0.
MC 2010 and CSCT show similar results. The slightly more dominant failure mode is for both
approaches failure within the shear-reinforced area. The average and the coefficient of variation
are slightly larger than the ones for the other two failure modes. However, the coefficient of
variation is still small leading together with a sufficient large average to a 5%-fractile above 1.0
for all failure modes. It can be noted that MC 2010 has in comparison to the other code
provisions not only the lowest coefficient of variation but is also the most balanced approach
with respect to the different failure modes.
81
Chapter 4
Table 4.3:
Predicted failure
mode
crushing of the
concrete strut
within shearreinforced area
outside shearreinforced area
Statistical analysis of tests from literature according to the predicted failure mode
# specimens
Average
COV
5% fractile
# specimens
Average
COV
5% fractile
# specimens
Average
COV
5% fractile
ACI2011
EC2004
NAD2011
SIA2003
MC2010
CSCT
32
1.29
13.6%
1.00
8
1.27
12.6%
1.01
63
1.48
20.3%
0.99
22
1.11
17.9%
0.78
17
1.11
11.0%
0.91
61
1.11
14.5%
0.85
37 / 341
1.25 / 1.111
11.4% / 12.8%1
1.02 / 0.871
7 / 81
1.14 / 1.131
11.2% / 10.8%1
0.93 / 0.931
60 / 621
1.30 / 1.291
13.7% / 13.8%1
1.01 / 1.001
28
1.28
9.6%
1.08
65
1.55
17.3%
1.11
9
1.22
8.8%
1.04
14
1.21
7.5%
1.06
43
1.32
11.7%
1.07
45
1.23
10.1%
1.02
22
0.92
7.7%
46
1.03
12.1%
36
1.00
8.8%
-
All the conclusions drawn on the basis of the statistical values can be additionally supported by
the detailed analysis of the ratio of the experimentally obtained punching strength to the
predicted strength as a function of several parameters such as the shear reinforcement ratio, the
effective depth, the column size to effective depth ratio, the flexural reinforcement ratio, and the
concrete compressive strength.
82

Figure 4.16 shows the ratio of the experimentally obtained punching strength of the tests from
literature to the predicted strength as a function of the shear reinforcement ratio. ACI 318-11
shows scattered but safe results for all shear reinforcement ratios. Additionally, it can be noted
that for larger amounts of shear reinforcement only crushing of the concrete and failure outside
the shear-reinforced area is predicted. In contrast, EC2 2004 predicts even for large amount of
shear reinforcement a failure within the shear-reinforced area. This effect results from the
overestimation of the crushing failure load as already discussed previously in this chapter.
Consequently, EC2 2004 has a tendency that the larger the shear reinforcement ratio is the lower
the ratio of the experimentally obtained strength to the predicted strength is. NAD 2011 avoids
these drawbacks of EC2 2004 and thus leads to safe estimates even for large amounts of shear
reinforcement. SIA262 2003 shows a clear underestimation of the strength in the case of low
amount of shear reinforcement. Due to the fact that SIA262 2003 neglects any concrete
contribution if shear reinforcement is present and thus the punching strength of slabs without
shear reinforcement is governing for low amounts of shear reinforcement (Figure 4.8c), the
punching strength is largely underestimated for these cases. MC 2010 as well as the CSCT
shows good agreement to the test results for all shear reinforcement ratios. No tendency occurs
in both cases.
(a)
(b)
3
2.5
(c)
no shear reinf.
crushing
within
outside
lexural failure
ACI 318-11
EC 2 : 2004
NAD 2011
2
1.5
1
0.5
0
(d)
(e)
(f)
3
SIA 262 : 2003
CSCT
MC 2010 Level II
2.5
2
1.5
1
0.5
0
0
0.5
1
w [%]
1.5
0.5
1
w [%]
1.5
0.5
1
w [%]
1.5
Figure 4.16: Ratio of the experimentally obtained punching strength of tests from literature to the
predicted strength as a function of the shear reinforcement ratio
83
Chapter 4
4.7.2. Effective depth

literature to the predicted strength as a function of the effective depth. ACI 318-11 shows
scattered but safe results without any particular tendency. EC2 2004 shows somewhat scattered
results without any clear tendency. However, it seems that EC2 2004 slightly overestimates the
strength in the case of thicker slabs but due to the scarce set of tests with effective depths larger
than 300 mm, no explicit conclusion can be drawn. Again, NAD 2011 improves the provisions
of EC2 2004 and thus leads to safer estimates of the strength for specimen with shear
reinforcement. Similar to ACI 318-11, SIA262 2003 shows scattered results, mainly due to the
negligence of the concrete contribution, for all thicknesses without showing any tendency. Also
no tendency occurs for MC 2010 and the CSCT, which both show good agreement for all
effective depths. Interesting to note is that unlike the other codes, MC 2010 as well as the CSCT
predicts failure within the shear-reinforced area for large effective depths. This can be explained
by the fact that the stresses in the shear reinforcement depend in both models on the slab
rotation. For small rotations such as it is the case for thick slabs, it is assumed that the shear
reinforcement is not fully activated, thus the yielding strength is not reached, leading to a
predicted failure within the shear-reinforced area.
(a)
(b)
3
(c)
EC 2 : 2004
ACI 318-11
NAD 2011
2.5
2
1.5
1
0.5
no shear reinf.
crushing
within
outside
lexural failure
(d)
(e)
(f)
3
SIA 262 : 2003
CSCT
MC 2010 Level II
2.5
2
1.5
1
0.5
0
0
100
200
d [mm]
300
400
100
200
d [mm]
300
400
100
200
d [mm]
300
predicted strength as a function of the effective depth
84
400
4.7.3. Column size

literature to the predicted strength as a function of the column size to the effective depth ratio.
As for the previously investigated parameters, ACI 318-11 does not show any tendency with
respect to the c/d ratio. EC2 2004 shows no tendency as one would expect according to Figure
4.4b. In fact, the estimates are not worse for large c/d ratios than for smaller ones. However, it is
interesting to note that unlike the other codes, EC2 2004 predicts only once a failure of the
concrete strut for specimens with moderate to large c/d ratios (c/d > 1.5). This observation
corresponds well to Figure 4.4b that shows a large increase in strength for larger c/d ratios for
failure of the concrete strut. Again, NAD 2011 overcomes this deficiency by using a different
formulation for failure of the concrete strut leading to larger ratios for larger column sizes.
Neither SIA262 2003, MC 2010, nor the CSCT show a clear trend with respect to the c/d ratio.
However, an interesting aspect is that MC 2010 predicts no failure of the concrete strut for small
c/d ratios unlike the predictions of the other codes. This is due to the fact that in the case of
small columns the punching strength for failure of the concrete strut might be overestimated as
shown in Figure 4.10b and Figure 4.11b. Although this strength might be overestimated,
MC 2010 provides still safe estimates since the considered cross sectional area of shear
reinforcement decreases significantly with smaller c/d ratios so that failure within the shearreinforced area is predicted (refer to Figure 4.10b). Similar to MC 2010, the CSCT predicts
larger normalized strength for smaller column sizes as it can be seen by comparing Figure 4.10b
and Figure 4.13b. However, in contrast to MC 2010, the CSCT may also predict failure of the
concrete strut for smaller column sizes as it is the case for the calculations shown in Figure
4.13b. This difference can mainly be explained by the fact that the CSCT allows more cross
sectional area of shear reinforcement for the calculation (compare subchapter 2.6.2 and 2.7.2)
leading to the prediction of larger punching strength for failure within the shear-reinforced area.
Consequently, the punching strength might be overestimated in the case of small column sizes
as it can be seen in Figure 4.18f.
85
Chapter 4
(a)
(b)
3
(c)
EC 2 : 2004
ACI 318-11
NAD 2011
2.5
2
1.5
1
0.5
0
no shear reinf.
crushing
within
outside
lexural failure
(d)
(e)
(f)
3
SIA 262 : 2003
CSCT
MC 2010 Level II
2.5
2
1.5
1
0.5
0
0
2
c/d [-]
2
c/d [-]
2
c/d [-]
predicted strength as a function of the column size to effective depth (c/d) ratio
86
4.7.4. Flexural reinforcement ratio

literature to the predicted strength as a function of the flexural reinforcement ratio. It can be
noted that all the codes predict a punching failure even for low flexural reinforcement ratios,
which correspond to the experimental observations. Generally, none of the models leads to an
obvious trend with respect to the flexural reinforcement ratio. Nevertheless, ACI 318-11 seems
to lead to lower ratios of experimentally obtained and predicted strength for lower flexural
reinforcement ratios than in the case of higher flexural reinforcement ratios.
(a)
(b)
3
(c)
EC 2 : 2004
ACI 318-11
NAD 2011
2.5
2
1.5
1
no shear reinf.
crushing
within
outside
lexural failure
0.5
0
(d)
(e)
(f)
3
SIA 262 : 2003
CSCT
MC 2010 Level II
2.5
2
1.5
1
0.5
0
0
2
flex [%]
2
flex [%]
2
flex [%]
predicted strength as a function of the flexural reinforcement ratio
87
Chapter 4
4.7.5. Concrete compressive Strength

literature to the predicted strength as a function of the concrete compressive strength. Since
most tests were performed with normal strength concrete with a compressive strength between
20 MPa and 40 MPa, the comparison lead to a cloud of points and no clear conclusions can be
drawn. However, it can be noted that no approach shows obvious deficiencies for tests with
concrete with compressive strength larger than 40MPa. Nevertheless, more test data is desired
in order to analyze the trend of each approach with respect to the compressive strength.
(a)
(b)
3
ACI 318-11
2.5
(c)
no shear reinf.
crushing
within
outside
lexural failure
EC 2 : 2004
NAD 2011
2
1.5
1
0.5
0
(d)
(e)
(f)
3
SIA 262 : 2003
CSCT
MC 2010 Level II
2.5
2
1.5
1
0.5
0
0
20
40
60
fc [MPa]
80
20
40
60
fc [MPa]
80
20
40
60
fc [MPa]
predicted strength as a function of the concrete compressive strength
88
80
5. Development of a Nonlinear Finite Element Approach

In order to study the load-deformation behavior of the slab specimens, a finite element approach
was developed. Since the nonlinearity in the behavior of cracked reinforced concrete has to be
considered in the calculation of the deformation, the calculation requires a sophisticated model.
Generally, two possible finite element methods seemed to be suitable: a two dimensional
approach with plate or shell elements or a three dimensional analysis with solid elements. Since
the complexity of a three dimensional model increases the number of (mostly unknown) input
parameters as well as generates problems in detailing the interaction of concrete and reinforcing
steel, the two dimensional approach was chosen. In order to implement the nonlinear behavior, a
modified stiffness approach was chosen by which the secant stiffness is calculated using a
plane stress field method.
Figure 5.1 shows the calculation procedure of the application in a flow chart. The approach is
based on a linear finite element analysis with modified stiffness for each element. For this, the
secant stiffness is calculated by the moment-curvature response of a single reinforced concrete
element (Figure 5.2) whereby the response is determined by the integration of the forces of
several plane stress field layers. Using this secant stiffness a linear finite element calculation
will be performed. The thereby calculated curvatures lead to a new set of stiffness parameters
for each element. This calculation is repeated until the difference of the reaction force and the
maximum rotation between two calculation steps is smaller than a certain tolerance. Afterwards,
the reaction force and the maximum rotation will be stored and the imposed displacement will
be increased. One of the advantages of this calculation is that all input parameters are well
defined and physically sound. Additionally, since the moment-curvature calculation is
performed only once, the calculation is less time consuming than other methods. However, this
procedure is only possible by using certain assumptions that are further explained in subchapter
5.3 in which the analysis is explained.
89
Chapter 5
Before the analysis is described, the calculation of the moment-curvature curves is presented
including the explanation of the input parameters and the used material properties. Finally, this
chapter presents a comparison of the nonlinear finite element approach to experimentally
obtained results from tests found in literature and from tests within this research project.
Calculation of momentcurvature relationships

(mx- x- y- xy; mxy- x- y- xy)
Linear-elastic FEM analysis

using E0 and G0
Calculation of the curvature

based on the nodal rotations
Determination of the stiffness

based on the curvature and the
moment-curvature
relationships
No
Adjustment of the secant

stiffness for the FEM Analysis
Calculation of the reaction

force V and the maximal
rotation x
Next load
(displacement)
step
Is the difference between

step i and i-1 small ?
Yes
Is the maximum
displacement reached?
Yes
Exit
Figure 5.1: Flowchart of the NLFEA
90
No
Development of a Nonlinear Finite Element Approach
5.1. Flexural stiffness

5.1.1. Compatibility conditions
Since it is assumed for the calculation of the flexural and torsional stiffness that normals to the
median plane of a slab remain straight and orthogonal to the median surface during
deformations, the strain at each level can be described as a linear function of the strains at midheight of the cross section (0,x, 0,y, 0,xy) and the curvatures (x, y, xy ):
,
,
(5.1)
2
Based on this, a strain profile can be determined from the given strains at mid-height of the
section and the curvatures.
(a)
(c)
(b)
Figure 5.2: Stress and strain distribution
From Mohrs circle of strains, one can calculate the principal strains (1, 2) and the direction of
the principal strains ().
z
2
2
z
(5.2)
where z is the distance to the mid-height of the cross section.

91
Chapter 5
1
arctan
2
(5.3)
By assuming that the directions of the principle strains coincide with the directions of the
principle stresses, one can calculate the stresses in the element based on the constitutive model.
5.1.2. Material behavior

Concrete
The concrete behavior is modeled on the basis of the approach of (Vecchio and Collins 1986).
The stress-strain relationship is described by a parabolic function, whereby the maximum
strength is determined by the transverse tensile strains (Equation 5.4). However, the parabolic
function was more generalized (Equation 5.5), so that adjustments could be easily made. This
modification allows implementing confinement or a residual strength.
0.8
(5.4)
0.34
where fc is the compressive strength, 1 the transverse tensile stain, and p the strain at the peak
compression stress.
0
where
(5.5)
and
92
The parabolic function is defined by the maximum compressive strength fc, which is reduced in
presence of transverse tensile stains 1, by the strain at the peak p, and by the maximum strain
r, at which a certain residual stress is obtained r. For the calculations herein, a peak strain of
p = -0.002, a maximum strain of r = -0.006, and a residual stress of r = 0 MPa were set (Figure
5.3).
(b)
(a)
1
0.8
c/fc,max [-]
fc,max / fc [-]
Model used
Vecchio/Collins
Thorenfeldt et al.
0.6
0.8
0.6
0.4
0.4
0.2
0.2
0
0
0.5
0
0
c,1 []
(c)
(d)
fc,max
1
0.8
c [MPa]
c / fct [-]
4
-c,2 []
0.6
0.4
0.2
r
0
0
ct
0.5
c,2 []
0
0
q
-c,2 []
Figure 5.3: Concrete stresses as a function of the principal (a) lateral tensile strains and (b) compressive
strains (Vecchio and Collins 1986; Thorenfeldt et al. 1987); (c) tension model and (d)
parameters for the implemented compressive stress-strain function
93
Chapter 5
Steel
The stress-strain behavior of the reinforcing steel is modeled as a bilinear function. The
behavior in tension and compression is assumed to be the same. The stresses increase linearly
with increase of the strain until the yielding strength fy is reached. After the yielding strength is
reached the stresses increase linearly until the ultimate strain u and the ultimate strength defined
as kshfy are reached. It has to be noted that this bilinear function was chosen in order to provide
numerical stability when the reinforcing bars are yielding. It is not supposed to model the strain
hardening accurately.
0
,
,
,
(5.6)
with
1
(b)
(a)
-su
su
-ksh
s []
s/fy [-]
s/fy [-]
ksh
-sy
sy
-1
s []
Figure 5.4: Stress-strain relationship for the reinforcing steel: (a) general view and (b) detailed view of
pre-yield behavior
Tension stiffening
The tension stiffening effect is the increase of the rigidity due to the tensile stresses in the
concrete between cracks transferred by bond of reinforcing bars at stabilized cracking of the
section (Ft > Fct). Various research has been performed in the past on this subject. However, in
the case of slabs subjected to bending, most formulations rely on rather coarse simplifications or
empirical adjustments. The difficulty lies mostly in defining the crack spacing, the bond
stresses, and the area of concrete contributing to the tension stiffening. For numerical analyses,
mostly two approaches are used in order to consider tension stiffening. The first one uses a
modified pseudo-behavior of concrete under tension (Vecchio and Collins 1986; Vecchio 1989;
94
Polak 1992; Polak and Vecchio 1993) and the second one uses a pseudo-behavior of the
reinforcing steel under tension (Sigrist 1995; Marti et al. 1998; Kaufmann and Marti 1998;
Alvarez et al. 2000). Within this research, the second approach was used.
fsy
ct
Figure 5.5:
s,TS
fct
y
s
ct
c,1
Considering tension stiffening by using (a) a modified concrete model (Vecchio and Collins
1986) or (b) a modified steel behavior (Sigrist 1995)
Assuming a rigid perfectly-plastic bond law, the stresses in the reinforcement bar reduce
linearly along the length. This assumption is valid as long as the reinforcing steel does not yield.
However, since the behavior of steel after yielding, steel hardening, is not modeled the tension
stiffening in the post-yield phase is neglected. On the basis of these assumptions the reduced
strain due to tension stiffening can be expressed as (Sigrist 1995):
1
(5.7)
where smax is the stress in the rebar at the crack, b is the bond strength, scr is the crack spacing,
is the diameter of the rebar, and Es is the Youngs modulus of the reinforcing steel.
Figure 5.6: Stresses and strains in a tension member considering tension stiffening
95
Chapter 5
The crack spacing is assumed to be the spacing of the flexural reinforcing bars. This rather
simple hypothesis can be supported by several considerations. Firstly, although several
researchers developed different approaches to express the crack spacing analytically, the
calculation methods are in most cases based on empirical simplifications and the variation of the
results from different research is rather high. Consequently, it seems more reasonable to use one
coarse but physically sound simplification than to use a more sophisticated formulation, which
uses several assumptions. Additionally, this basic assumption shows a good agreement with the
experimentally observed average crack spacing.
5.1.3. Equilibrium conditions

By using the stress-strain relationship for concrete and steel the internal forces can be
calculated. Because the concrete stresses (principal directions) and the steel stresses (direction
of the global coordinate system) are not equivalent, the stresses have to be transformed to the
same direction. Since in the NLFEA the stiffness will be defined in the direction of the global
coordinate system, the stresses need to be transformed into these directions. Generally, the
transformation of stresses (x, y, xy) can be performed by using the inverse of the transformed
strain transformation matrix [T] in order to obtain the stresses in another direction (x, y, xy).
The strain transformation matrix is defined as (Cook et al. 2002):
(5.8)
where for two dimensional problems, the transformation matrix can be expressed as:
cos
sin
2 sin
cos
sin
cos
2 sin
cos
sin
sin
cos
cos
cos
sin
(5.9)
Similarly, the strain transformation matrix can be used to obtain the stress transformation
relations (Cook et al. 2002):
-T
(5.10)
where [T]-T has the form

T
cos
sin
-T
sin
96
cos
sin
cos
sin
cos
2 sin
2 sin
cos
cos
cos
sin
(5.11)
It can be noted that the same relationship can be stated for curvatures:
(5.12)
and for moments:
-T
(5.13)
Since the calculation of the concrete stresses is performed in the principal direction, one has to
transform them into the direction of the reinforcement to form the equilibrium condition. Thus,
if the existing stresses are the principal stresses and is the principal direction (Equation 5.3),
Equation 5.10 becomes:
-T
(5.14)
By integrating the concrete stresses, one can obtain the concrete forces in the global directions.
Thus,
,
-T
(5.15)
The steel forces can be obtained by the summation of the forces in each rebar. Thus,
,
(5.16)
,
The first set of equilibrium conditions can be described as the sum of the internal and external
membrane forces in each direction. Thus,
0
(5.17)
Similarly, the moments due to the concrete stresses can be defined as:
97
Chapter 5
-T
(5.18)
and the moments due to the steel stresses:

,
(5.19)
,
leading to the second set of equilibrium conditions, which equals the sum of the internal and
external moments in each direction to zero. Thus,
0
(5.20)
Hence, for each state of deformation (0,x, 0,y, 0,xy, x, y, xy) the external forces (nx, ny, nxy, mx,
my, mxy) can be calculated. Assuming that the membrane forces (nx, ny, nxy) are equal to zero, one
can calculate the moment-curvature relationship for each state of deformation. Figure 5.7
illustrates the moment mx and mxy as a function of the curvature x and xy for a lateral curvature
y equal to zero. The calculated surfaces in Figure 5.7 clearly show the change in the slab
behavior depending on the different state of deformation. For example, the flexural stiffness
decreases if the torsional curvature increases. Even more pertinent is the change of the torsional
stiffness due to the flexural curvature.
(a)
(b)
120
200
100
100
mx [kN]
80
0
mxy [kN] 60
100
40
200
20
0
0.03
300
0.03
0.1
0.02
-1
xy [m ]
0.1
0.02
0.05
0.05
0.01
0.05
0
0.1
0
x [m-1]
xy [m-1]
0.01
0.05
0
0.1
Figure 5.7: Moment as a function of the curvatures x and xy for (a) mx and (b) mxy
98
x [m-1]
The dependence of the stiffness on the different curvatures can also be seen in Figure 5.8 that
shows the moment mx as a function of the curvature x for different levels of xy and the moment
mxy as a function of the curvature xy for different levels of x. Again, it can be seen that while a
difference in the flexural stiffness due to a torsional curvature only marginally occurs, the
difference in the torsional stiffness due to a flexural curvature is significant. Therefore, these
effects clearly need to be considered for an accurate analysis of a slab in which torsional
moments occur.
(a)
(b)
300
150
y=0.00 m-1
y=0.00 m-1
200
100
mxy [kN]
mx [kN]
100
0
100
50
xy=0.00 m-1
xy=0.01 m-1
xy=0.02 m-1
200
300
0.1
Figure 5.8:
0.05
0
x [m-1 ]
0.05
x=0.00 m-1
x=-0.03 m-1
x=-0.06 m-1
0
0
0.1
0.01
0.02
0.03
xy [m -1]
0.04
0.05
Interaction of the curvatures x and xy: (a) bending moment-curvature diagram in xdirection and (b) torsional moment-curvature diagram
Based on the calculated moment-curvature response of the element, the secant stiffness matrix K
can be determined, which is afterwards used for the linear-elastic finite element analysis. The
secant stiffness matrix becomes:
0
0
0
0
0
(5.21)
Although this diagonal matrix does not seem to account for the interaction between the different
curvatures, each stiffness (Kx, Ky, and Kxy) is determined as a function of all the curvatures (x,
y, and xy). Thus, the interaction between the different curvatures is accounted in the calculation
of each stiffness.
5.2. Shear stiffness

Flat slabs are generally thin structures. Thus, the shear deformations are relatively small and are
usually neglected in FEM calculations of slabs (e.g. by using Kirchhoff elements). In this case
the shear deformations are considered (based on the Mindlin theory) but with an uncracked
shear stiffness (Gxz = Gyz = Gc0). This assumption is valid for regions where the shear
deformations are small and thus for regions where no shear reinforcement is used. In regions
99
Chapter 5
with shear reinforcement, in the vicinity of the columns, the shear deformations can be
significant as it was presented in Chapter 3. As it will be discussed in Chapter 7, these large
shear deformations result from the compression strut close to the column and thus it is assumed
that they lead to a rigid body shift of the slab element outside the supported area. Therefore, it is
assumed that the load-rotation response of the slab is only slightly influenced by these shear
deformations.
5.3. Analysis
The analysis procedure consists of certain simplifications in order to reduce the demand of
resources and to increase the calculation speed. The first assumption concerns the calculation of
the flexural stiffness. The flexural stiffness is regarded as independent from the actions
perpendicular to the investigated direction. For example, it is assumed that the curvature y,
which acts perpendicular to the moment mx, does not influence the stiffness in x-direction. Thus,
only the curvatures x and xy are considered for the calculation of the stiffness EIx. This
assumption can be supported by the moment-curvature (mx-x) curves for different levels of
lateral y and torsional curvatures xy. Figure 5.9 illustrates the moment-curvature curves for
different levels of lateral and torsional curvatures. It can be seen that the discrepancy between
the different curves are relatively small leading to only small differences in the secant stiffness.
Therefore, this minor error resulting from the assumption can be accepted. However, it has to be
noted that this assumption can only be made as long as the curvatures in x- and y-direction have
the same sign. In this case the concrete is influenced only by lateral compression, which would
slightly increase the strength (Kupfer and Gerstle 1973). In the case of lateral tension due to an
opposite signed flexural moment, the compression zone will be drastically softened.
Consequently, the rigidity would be largely decreased and thus the assumption would not be
valid.
(a)
(b)
300
xy=0.00 m
(c)
-1
-1
-1
xy=0.015 m
xy=0.02 m
y=0.00 m-1
y=-0.04 m-1
y=-0.06 m-1
y=0.00 m-1
y=-0.04 m-1
y=-0.06 m-1
mx [kN]
250
200
150
100
y=0.00 m-1
y=-0.04 m-1
y=-0.06 m-1
50
0
0
0.05
x [m -1]
0.1
0.05
x [m -1]
0.1
0.05
x [m -1]
0.1
Figure 5.9: Moment-curvature diagram in x-direction for different lateral curvatures y and for twisting
curvatures of (a) xy = 0.00 m-1, (b) xy = 0.015 m-1, and (c) xy = 0.02 m-1
100
The torsional stiffness however cannot be assumed as independent of the moments in x- and ydirection. In order to address this influence, the moment mxy needs to be calculated by a
variation of xy and a pair of x and y. Figure 5.10 shows the moment-curvature (mxy- xy) curve
for different flexural curvatures x and y whereby the curvature y is shown as a function of xy.
Unlike the flexural curvature (x), the torsional stiffness depends largely on the flexural
curvatures as the discrepancies between the different curves show. Another interesting aspect to
note is that in the case of lateral curvature (y) the stiffness is low until a certain point at which
the slab starts to behave more rigid. This can be explained in the change of the state of stress. At
first a large depth of the slab is cracked due to the dominance of the curvature x. However, at a
certain point some cracks will be closed due to compression resulting from the twisting
moments, leading finally to an increase in stiffness. This explanation is of course only based on
a pure theoretical point of view since the actual behavior may depend as well on the effective
loading and thus the cracking history of the slab element. Nevertheless, since the load history is
neglected in the NLFEA, the theoretically obtained behavior is assumed to be accurate enough.
Another assumption was made in order to decrease the calculation time and data volume. Since
the stiffness is within two boundaries, only one calculation with the lateral curvature equal to
zero (xy 0; x 0; y = 0) and one calculation with the lateral curvature equal to the main
curvature (xy 0; x = y 0) are performed whereby values of the stiffness between these two
boundaries will be linearly interpolated.
(a)
(b)
mxy [kN]
150
-1
-1
x=-0.015 m
x=-0.030 m
y=-0.00xy
y=-0.75xy
y=-1.00xy
y=-0.00xy
y=-0.75xy
y=-1.00xy
100
50
0
0
0.01
0.02
0.03
xy [m-1]
0.04
0.05
0.01
0.02
0.03
xy [m-1]
0.04
0.05
Figure 5.10: Moment-curvature diagram for torsional moments
5.4. Comparison
The analysis of the stiffness in combination with the finite element method is independent of the
geometry and the boundary conditions of the slab. Therefore, any slab test could be used in
order to verify the proposed model. However, since the main objective is to obtain the loadrotation curve of punching tests performed within this research, similar punching tests from
literature were used in addition to the tests performed within this research in order to verify the
numerical model. Additionally, the two boundary case regarding the flexural behavior, pure
101
Chapter 5
bending and pure torsion, were investigated. Therefore, tests found in literature were used to
compare the response of the tests specimens to the response predicted by the numerical model.
5.4.1. Pure bending

The first boundary case that was analyzed was pure bending. The investigated specimens were
slab elements with dimensions of 4.00 m x 1.00 m tested by (Kenel and Marti 2001). The slab
specimens were loaded symmetrically at each end and supported on two bearings at a distance
of 0.60 m away from the center axis. Figure 5.11 shows the principal dimensions of the test setup and the loading. The measurements that were compared are the force Q, which was applied
at each side, and the displacement w that was measured at the load introduction point.
(a)
(b)
Figure 5.11: Dimensions and loading of the bending test specimens
Table 5.1 summarizes the main parameters of the test specimens. For all the slabs, the slab
dimensions were constant (4.00 m x 1.00 m x 0.20 m). The investigated parameters were the
compressive strength of the concrete (Specimen B3), the reinforcement ratio (Specimen B4) and
the bar diameter of the reinforcement (Specimen B5). Since the test program treated main input
parameters of the NLFEA such as the reinforcement ratio and the concrete strength, this test
series is well suited for a comparison to the finite element calculations. Additionally, the
measured crack spacing can be used to verify the assumed tension stiffening approach. For this,
the measured crack spacing, as indicated in Table 5.1, was used for the comparison of these
tests. In addition, the measured tensile strength was used instead of the calculated one.
Table 5.1: Parameters of bending test specimens from literature
Specimen
B1
B3
B4
B5
102
h
[mm]
200
200
200
200
x
[%]; ([mm])
0.31 (810)
0.31 (810)
0.16 (410)
0.31 (414)
fy
[MPa]
563
563
563
508
fc
[MPa]
41.5
81.4
37.3
39.4
fct
[MPa]
3.2
4.8
3.1
2.7
srm
[mm]
150
140
280
150
Figure 5.12 shows the measured and the calculated load-displacement curves of the four
specimens. The two curves of the measured values correspond to the values obtained when the
loading was halted to take measurements (higher load value) and when the manually performed
deformation measurements were completed (lower load value). The comparison shows good
agreement between the calculated and the experimentally obtained curves. Generally, the
NLFEA predicts a slightly stiffer behavior before the yielding moment. After yielding, it
somewhat underestimates the strength. This can be explained by the fact that it does not
consider the strain hardening of the reinforcing steel properly (a small strain-hardening is
implemented but only to prevent numerical problems) and that it neglects tension stiffening
after yielding of the reinforcement. Despite these small differences, it can be noted that the
behavior is calculated accurately for specimen B1, B3, and B4. The inaccuracy for specimen B5
is due to the asymmetry of the test performed. In fact, the displacements at one side were
significantly larger than at the other side. Since the NLFEA uses a displacement controlled
analysis with equal displacements at each side, the strength at yielding will be overestimated.
However, the ultimate strength should be the same. This can also be seen in the diagram, at
which for large displacements the experimentally obtained curves and the calculated curves
converge.
(a)
(b)
NLFEA
Measured
70 B1
B3
NLFEA
Measured
B5
NLFEA
Measured
60
Q [kN]
50
40
30
20
10
0
(c)
(d)
50
45
NLFEA
Measured
B4
40
35
Q [kN]
30
25
20
15
10
5
0
0
50
100
w [mm]
150
50
100
150
w [mm]
Figure 5.12: Measured and calculated load-displacement curves of the bending test specimens: (a) B1,
(b) B3, (c) B4, and (d) B5
103
Chapter 5
5.4.2. Pure torsion

The second boundary case investigated is pure torsion. The investigated specimens were square
slab elements with a side length of 1.70 m, which were tested by Marti et al. (Marti et al. 1987).
The forces were applied downwards on opposite corners of the slab. The slab was supported by
two spherical bearings at the other two opposite corners. The introduced load as well as the
reaction forces was distributed by a steel plate with dimensions of 0.150 m x 0.150 m x 0.025 m
leading to a distance of 1.55 m between the load introduction point and the reaction point. The
vertical displacement was measured at the load introduction and the reaction points. The
measurements of the vertical displacements lead to the relative vertical displacement w of the
slab, which relates to the non-displaced axis as shown in Figure 5.13.
Figure 5.13: Dimensions and loading of the torsion test specimens
The investigated parameter was the reinforcement ratio whereas all the other parameters such as
the slab dimensions were kept constant. Table 5.2 shows the main parameters of the test
specimens. For the crack spacing, a value of half of the bar spacing was chosen, which
corresponds to 50 mm for specimen ML2, ML3, and ML9 and 100 mm for specimen ML1 and
ML7. These values were chosen in accordance to the picture of the specimens shown in the
reference (Marti et al. 1987).
Table 5.2: Parameters of torsion test specimens from literature
Specimen
ML1
ML2
ML3
ML7
ML9
104
h
[mm]
200
200
200
200
200
x = y
[%]
0.25
0.50
1.00
0.25
1.00
fy
[MPa]
551
551
481
479
479
fc
[MPa]
46.7
36.2
37.5
44.4
44.4
Figure 5.14 shows the moment-displacement curves for the torsion tests. The NLFEA shows
good agreement with the test results. The small difference in the slab response between the
calculated and the measured values can be attributed to shear deformations that were not
considered in the numerical calculation. Additionally, larger differences can be observed at the
uncracked state. However, this can be explained by the step width of the calculation. Since the
peak, at which the cracking moment is reached, lies between two calculation points, it is not
accurately displayed. This problem could be overcome by using more dense calculation steps.
Nevertheless, this part was neglected and the step width was kept as it is because the principal
interest lies in the cracked behavior.
(a)
(b)
(c)
100
ML1
90
ML3
ML2
mxy [kN]
80
70
60
50
40
30
20
10
0
0
10
(d)
20
30
w [mm]
40
50
10
20
30
w [mm]
40
50
100
10
20
30
w [mm]
40
50
120
ML7
90
ML9
100
80
70
mxy [kN]
mxy [kN]
(e)
60
50
80
60
40
40
30
20
20
10
0
0
10
20
30
w [mm]
40
50
0
0
10
20
30
w [mm]
40
50
Figure 5.14: Measured (peak and end value of each load step) and calculated (continuous curve)
moment-displacement curves of the torsion test specimens: (a) ML1, (b) ML2, (c) ML3, (d)
ML7, and (e) ML9
105
Chapter 5
5.4.3. Punching of slabs without shear reinforcement

From the two investigated boundary cases, it can be seen that the NLFEA delivers results that
are in good agreement with the experimentally measured ones. This leads to the conclusion that
the NLFEA should predict the behavior of punching tests accurately. In order to verify this
expectation, punching tests of slab specimens without shear reinforcement found in literature
(Guandalini and Muttoni 2004; Guidotti et al. 2009) and from tests within this research project
were analyzed. Figure 5.15 illustrates the dimensions and the loading of the punching test
specimens from literature.
Figure 5.15: Dimensions and loading of the punching test specimens
Except for specimen PG3 (double in size), all test specimens had the same dimensions
(3.00 m x 3.00 m x 0.25 m) and were supported by the same column size (c = 0.26 m). The
variable parameter is the flexural reinforcement ratio, which ranges from a rather low ratio of
= 0.25% to a reinforcement ratio of = 1.5%. Table 5.3 presents the main parameters of the
punching test specimens without shear reinforcement from the literature. All parameters as well
as loading conditions of the punching test specimens within this research project (PL1-PL5) can
be found in Chapter 3.
Table 5.3: Parameters of specimens without shear reinforcement of punching tests from literature
Specimen
PG11
PG2b1
PG31
PG41
PG51
PG192
1
b
[mm]
3000
3000
6000
3000
3000
3000
h
[mm]
250
250
500
250
250
250
c
[mm]
260
260
520
260
260
260
bq
[mm]
1380
1380
2590
1380
1380
1380
(Guandalini and Muttoni 2004); 2 (Guidotti et al. 2009)
106
rq
[mm]
1505
1505
2846
1505
1505
1505
x = y
[%]
1.50
0.25
0.33
0.25
0.33
0.78
fy
[MPa]
573
552
520
541
555
510
fc
[MPa]
27.7
40.5
32.4
32.2
29.3
46.2
Figure 5.16 shows the measured and the calculated load-rotation curve of tests from literature.
The calculations of all specimens show good agreement with the measured behavior. Only for
low rotations, the stiffness will be slightly underestimated. This could be improved by using a
finer mesh of the constitutive surfaces. However, since this phase is of less interest in this
research, no further analysis of this phenomenon was performed. Additionally, one has to note
that the loads in most cases are relatively small. Thus, the influence of shear deformation is
small too.
(a)
(c)
(b)
2500
NLFEA
Measured
PG1
NLFEA
Measured
PG3
NLFEA
Measured
PG19
2000
V [kN]
1500
1000
500
0
0
10
15
[]
20
25
(d)
10
15
[]
20
900
NLFEA
Measured
PG2b
10
15
[]
20
25
(f)
(e)
1000
25
NLFEA
Measured
PG4
NLFEA
Measured
PG5
800
700
V [kN]
600
500
400
300
200
100
0
0
10
20
30
[]
40
50
10
20
30
[]
40
50
10
20
30
[]
40
50
Figure 5.16: Measured and calculated load-rotation curves of the punching test specimens: (a) PG1,
(b) PG3, (c) PG19, (d) PG2b, (e) PG4, and (f) PG5
Figure 5.17 shows the measured and the calculated load-rotation curve of tests within this
research project of the specimens without shear reinforcement. Generally, the slab specimens
behave similarly as the tests form literature presented previously and thus the calculated and
measured behavior show a good agreement. However, due to the flexural reinforcement ratio of
L = 1.5%, they reach higher punching strengths than the tests form literature presented
previously. This higher load level seem to influence the stiffness of the slab. Especially, in the
cases with large shear forces (Specimens PL4 and PL5), the behavior predicted by the NLFEA
is stiffer than the measured one.
107
Chapter 5
(a)
(b)
1500
PL1
NLFEA
Measured
PL4
NLFEA
Measured
PL3
NLFEA
Measured
PL5
NLFEA
Measured
V [kN]
1000
500
(c)
(d)
3000
2500
V [kN]
2000
1500
1000
500
0
0
10
15
[]
20
25
10
15
[]
20
25
Figure 5.17: Measured and calculated load-rotation curves of the punching test specimens: (a) PL1,
(b) PL3, (c) PL4, and (d) PL5
108
5.4.4. Punching of slabs with shear reinforcement

Figure 5.18 and Figure 5.19 show the measured and the calculated load-rotation curve of tests
within this research project of the specimens with studs and stirrups, respectively. As already
seen for the specimens without shear reinforcement, the curves correspond well for lower shear
forces. However, in the case of larger shear forces the slabs show a lower stiffness as already
shown in Figure 5.17. Whereas the difference between the measured and calculated behavior for
slabs without shear reinforcement is rather small, it becomes significant for slabs with shear
reinforcement. Consequently, it is necessary to consider in the analysis the reduction in stiffness
due to concentrated shear forces.
(a)
(b)
3000
NLFEA
Measured
PL6
(c)
NLFEA
Measured
PL7
NLFEA
Measured
PL8
2500
V [kN]
2000
1500
1000
500
0
(d)
(e)
6000
NLFEA
Measured
PL9
10
20
30
[]
40
50
NLFEA
Measured
PL10
5000
V [kN]
4000
3000
2000
1000
0
0
10
20
30
[]
40
50
10
20
30
[]
40
50
Figure 5.18: Measured and calculated load-rotation curves of the punching test specimens: (a) PL6,
(b) PL7, (c) PL8, (c) PL9, and (e) PL10
109
Chapter 5
(a)
(b)
3000
NLFEA
Measured
PF1
(c)
NLFEA
Measured
PF2
NLFEA
Measured
PF3
2500
V [kN]
2000
1500
1000
500
0
(d)
(e)
6000
NLFEA
Measured
PF4
10
20
30
[]
40
50
NLFEA
Measured
PF5
5000
V [kN]
4000
3000
2000
1000
0
0
10
20
30
[]
40
50
10
20
30
[]
40
50
Figure 5.19: Measured and calculated load-rotation curves of the punching test specimens: (a) PF1,
(b) PF2, (c) PF3, (d) PF4, and (e) PF5
As already illustrated in Chapter 4, the difference of the load-rotation curves for large shear
forces is not only a problem of the NLFEA. Since other models such as the analytical
Quadrilinear model (Muttoni 2008) neglect the influence of the shear forces on the flexural
behavior, the same discrepancies occur (see Chapter 4). In fact, the calculated curves of the
NLFEA and the Quadrilinear model (Muttoni 2008) show nearly the same behavior (Figure
5.20). Therefore, it is crucial to analyze the influence of the shear forces on the flexural stiffness
in order to improve the existing models and the NLFEA.
110
(a)
(b)
3000
6000
Model
Measured
PL7
Model
Measured
PL10
2500
5000
Quadrilinear
4000
Quadrilinear
NLFEA
V [kN]
V [kN]
2000
1500
NLFEA
1000
3000
2000
500
1000
0
0
10
20
30
[]
40
0
0
50
10
20
30
[]
40
50
Figure 5.20: Comparison of the calculation model Quadrilinear (Muttoni 2008) and the NLFEA to the
experimentally obtained load-rotation curve of specimens (a) PL7 and (b) PL10
In order to see whether the nonlinear finite element model is capable to predict the loaddeformation response of the slab correctly when the influence of the shear force on the
reduction of the stiffness is known, the longitudinal reinforcement ratio was reduced by the ratio
of the measured flexural strength (VR,measured) to the calculated flexural strength (Vflex,calc). By
using a reduced flexural reinforcement ratio, a good agreement with the test results was
obtained (Figure 5.21). Therefore, it is possible to estimate the global behavior of the slab
accurately with the NLFEA, if local effects such as the reduction of the flexural stiffness due to
the concentrated shear forces in the vicinity of columns are considered. However, it has to be
noted that this is only applicable for the analysis of tested slab specimens by which the loadrotation response is known.
(a)
(b)
3000
6000
NLFEA reduced
Measured
PL7
5000
2000
4000
1500
3000
1000
2000
500
1000
0
0
10
20
30
[]
40
NLFEA reduced
Measured
PL10
2500
V [kN]
V [kN]
(5.22)
50
0
0
10
20
30
[]
40
50
Figure 5.21: Load-rotation curve for specimen (a) PL7 and (b) PL10 calculated with a reduced flexural
reinforcement ratio
111
6. Analysis of the Slab Response

As discussed in the previous chapter, the shear concentration at the vicinity of the column has a
significant influence on the flexural behavior of the slab. Consequently, the interaction of
moment and shear influences the slab behavior and thus needs to be considered in the analysis
of the punching test specimens. In order to account for the influence of shear in the numerical
model, the nonlinear finite element analysis (subsequently referred to as NLFEA) described in
the previous chapter was combined with an existing finite element software which calculates inplane deformations due to in-plane forces (subsequently referred to as jconc). It has to be noted
that this method explained in this chapter is used to analyze the global and local behavior of the
slab specimens separately. Neither one of this program is able to model the slab specimens
correctly on its own. However, by combining the two programs, a good agreement with the
actual measured behavior can be achieved with respect to the displacements and deformations
leading to valuable insights in the slab behavior. Nevertheless, one has to distinguish between a
global and a local slab behavior. Therefore, the comparison to the experimentally obtained
measurements and the results of the analysis is divided into a global and a local part. Before
these results are shown, the basis of the in-plane analysis tool will be briefly explained. For a
more detailed description it is referred to literature (Fernndez Ruiz and Muttoni 2007; Kostic
2009; Muttoni et al. 2011).
113
Chapter 6
6.1. Analysis procedure

As mentioned previously, the NLFEA explained in Chapter 5 was combined with jconc, which
is a two dimensional non-linear finite element analysis program that is used to analyses structure
with respect to the in-plane forces. The program uses an elastic-plastic material behavior for
steel and concrete whereby the maximum concrete strength is reduced to account for lateral
tensile strains. Due to the fact that the program uses several assumptions such as that the
principal strain directions coincide with the principle stress directions, it requires only few and
well defined input parameter leading to a rather straightforward calculation. Despite its ease of
application, it provides good predictions of the response of beams with shear reinforcements and
walls (Muttoni et al. 2011).
The separate calculations of the global and local response were combined by certain input
parameters as explained subsequently. Besides the geometrical and the material parameters, an
additional parameter was introduced into the NLFEA in order to reduce the flexural stiffness. In
fact, this introduced factor multiplies the amount of flexural reinforcement so that the stiffness
changes. With this adjusted stiffness, the program calculates the global slab behavior leading to
the slab deformations and to the internal forces. These forces serve as input parameter for the
jconc analysis. Since jconc performs only in-plane calculation it is not capable to analyze a slab
accurately. However, if only a short, thin cut of the slab along the axis is taken, the influence of
out-of plane actions can be reduced. Therefore, a cut along the axis in the vicinity of the column
was taken.
The length of the cut lcut was chosen so that the slab portion between the center of the slab and
the third row of shear reinforcement is modeled. This distance was chosen based on the
assumption that after the third row of studs the influence of the shear forces is small, which
corresponds to the experimental observations. The width of the cut wcut was chosen as 100 mm
at the column face and stepwise increased at the outer part accounting for the distribution of the
shear forces. This leads to a slight overestimation in stiffness but since the main deformations in
radial direction occur close to the column, this effect can be neglected. The moment and the
shear force that were introduced at the end of the cut could be obtained by the NLFEA. The
moment was determined at the end of the cut length and was introduced by a force-couple. The
shear forces were determined near the column face at a distance of 0.375d in order to avoid
regions with numerical distortions resulting from the boundary conditions of the support, which
was modeled using four point supports at the corner of the column. These shear forces were also
applied at the end of the cut model.
Figure 6.1 presents the calculation method in more detail for the calculation at one applied
displacement level. In a first step, the response of a slab element is calculated with the actual
flexural reinforcement ratio. With the obtained stiffness, a nonlinear finite element analysis is
performed. From the NLFEA, the shear force and the moment are extracted and introduced in
the jconc (in-plane) analysis. The shear force and the moment are scaled by a factor , which
will be adjusted until the rotation obtained from the jconc analysis corresponds to the rotations
114
Analysis of the Slab Response
obtained by the NLFEA. This factor defines the stiffness reduction factor , which describes
the reduction in the stiffness by adjusting the flexural reinforcement ratio. Afterwards, a new
NLFEA with the adjusted stiffness was performed. This procedure is repeated for certain times.
For the calculations performed herein, the maximum number of iteration was set to imax = 3
leading to an acceptable convergence. Finally, the applied displacement is increased and the
procedure starts from the beginning.
Calculation of the
stiffness K=f()
NLFE Analysis
Extract forces
V; M
Set =1.0; i=1
In-plane analysis
with V; M
Change
No
Calculation of the
stiffness K=f( )
slabcut
Yes
=
No
i>imax
Yes
EXIT
Figure 6.1:
Flowchart of the interaction of the NLFEA and the in-plane analysis tool for the calculation
at a certain load / displacement level
115
Chapter 6
Figure 6.2: Plan view and section of the slab cut modeled in an in-plane analysis tool (jconc)
This method allows modeling the behavior of the slab near the column leading to the vertical inplane stresses within the cut plane and the vertical deformations. By using the results from the
calculation of the cut, one can calculate the rotation of the slab at the end of the cut. Since this
rotation has to be the same as the one of the NLFEA at this point, the stiffness factor used in the
NLFEA can be adjusted until both rotations are the same. If this calculation is performed for
different applied displacements, one can obtain the global load-rotation relationship. Although
this approach uses coarse simplifications and thus the results should be viewed with caution, it
provides results that correspond well to the measurements from tests leading to valuable
information regarding the global and local slab response.
Figure 6.3 shows the calculated and the measured load-rotation curve for specimens PL7, PL8,
and PL9. It can be seen that the prediction of the load-rotation curve are in good agreement with
the experimentally obtained behavior. However, it has to be noted that the stiffness reduction
factor can only be calculated until yielding of the flexural reinforcement otherwise no
equilibrium can be found for the cut model. Therefore, the factor was chosen constant for the
calculation after yielding of the flexural reinforcement occurs. Generally, the adjustment factor
is close to 1.0 for low load levels, thus nearly no change in stiffness exists. Afterwards, this
factor decreases and reaches the minimal value when yielding occurs leading to a smaller
stiffness and thus to a change in the behavior of the slab.
116
(a)
(b)
1.2
Calculated
Measured
PL7
(c)
Calculated
Measured
PL8
Calculated
Measured
PL9
MPa
1.0
0.8
0.6
0.4
0.2
0
0
Figure 6.3:
0.1
0.2
0.3 0
0.1
0.2
0.3 0
0.1
0.2
0.3
Measured and calculated load-rotation curve for specimen (a) PL7, (b) PL8, and (c) PL9
6.2. Global slab behavior

The global slab behavior was analyzed by the adjusted NLFEA. From these calculations, one
obtained the state of deformations and the internal forces. These results were validated by the
comparison to the experimentally obtained values. For this, the global behavior of slab PL7 was
analyzed and is presented in detail subsequently. The calculations were performed at three states
of deformation corresponding to the measured load of 60%, 75%, and 90% of the failure load. It
has to be noted that the analysis was not performed at the same load level as the measured load
but at the corresponding rotation in order to achieve the same state of deformation.
Table 6.1:
Shear force, rotation, and vertical displacement at each investigated load level
Load Level
Measured shear force V [kN]
Rotation [-]
Applied displacement dz [m]
60%
1060
0.103
-0.012
75%
1334
0.144
-0.017
90%
1585
0.20
-0.025
117
Chapter 6
The comparison of the calculated to the experimentally measured deformations shows the
performance of the NLFEA approach. For this comparison, the vertical displacements as well as
the strains in radial and tangential direction on the bottom surface and in radial direction on the
top surface were investigated for the three load cases. Figure 6.4 shows the calculated and
measured values for the load stage of 60% VR. The vertical displacements calculated with the
NLFEA and Jconc correspond well to the experimentally obtained values. Differences occur
only in the prediction of the shear deformations at the column face since they are not modeled in
the NLFEA. With respect to the radial strains at the top surface of the slab (tension side), it can
be noted that the measurements clearly depend on the crack opening and crack spacing. Since
the NLFEA uses a smeared crack approach, these local strains cannot be modeled properly.
However, if the average strains are considered, the NLFEA shows a good agreement. On the
bottom surface of the slab (compression side), the NLFEA leads to a good agreement with the
experimentally measured values. For this load case, the maximal strains are at around 2, thus
close to the value of the peak strain of the assumed stress-strain relationship of the concrete.
(a)
dz [mm]
0
20
NLFEA (bending)
jconc (in-plane)
measured
40
(b)
r,top []
30
r,bot []
(c)
(d)
NLFEA
measured
20
10
0
10
0
2
4
NLFEA
measured
t,bot []
0
2
4
6
1500
Figure 6.4:
118
NLFEA
measured
1000
500
0
r [mm]
500
1000
1500
Comparison of the modeled and measured deformations along the axis for a load level of
60% VR: (a) vertical displacements, (b) radial strains on the top surface, (c) radial strains at
the bottom surface, and (d) tangential strains at the bottom surface
For the load case of 75%VR, the calculated vertical displacements are again in a good agreement
with the experimentally measured values (Figure 6.5). Although the measured radial strains on
the top surface depend again on the cracking, the trend is well predicted by the NLFEA. On the
bottom side, the predictions in radial and tangential direction correspond well with the
measurements. It can be noted that at this load level the radial strains are stabilized, thus they do
not increase with increasing load until failure occurs (see Chapter 3 and Figure 6.7). This
stabilization is also well predicted by the NLFEA as it can be seen for the load level of 90% VR.
(a)
dz [mm]
0
20
NLFEA (bending)
jconc (in-plane)
measured
40
(b)
r,top []
30
r,bot []
(c)
(d)
NLFEA
measured
20
10
0
10
0
2
4
NLFEA
measured
t,bot []
0
2
4
6
1500
Figure 6.5:
NLFEA
measured
1000
500
0
r [mm]
500
1000
1500
119
Chapter 6
Similar to the previous load levels, the predicted and the experimentally obtained displacements
and strains at the load level of 90%VR are in good agreement (Figure 6.6). The calculated tensile
strains in radial direction on the top surface correspond to the tendency of the measurements.
However, due to the cracking, no definite conclusion can be drawn. The calculated compressive
strains on the bottom side in radial, and especially in tangential direction, show an excellent
agreement with the measurements. It has to be noted that the tangential strains are beyond the
peak strain of the assumed stress-strain relationship of the concrete. Thus, it can be concluded
that the softening of the concrete is properly modeled.
(a)
dz [mm]
0
20
NLFEA (bending)
jconc (in-plane)
measured
40
(b)
r,top []
30
r,bot []
(c)
(d)
NLFEA
measured
20
10
0
10
0
2
4
NLFEA
measured
t,bot []
0
2
4
6
1500
Figure 6.6:
120
NLFEA
measured
1000
500
0
r [mm]
500
1000
1500
Figure 6.4 to Figure 6.6 showed that the calculated strains at the bottom side are generally in
good agreement with the measured strains for the investigated load levels. In order to analyze
the radial and tangential strains at the bottom surface in more detail, the load-deformation
curves were calculated at the locations of the measurements and compared to the measured
values. Figure 6.7a shows the radial strains measured at two locations at the bottom surface of
the slab. It can be seen that the behavior of the strains can be accurately modeled. It is
interesting to note that the numerical model, which considers only flexural behavior (bending
and torsion), follows the curve of the strain measurements in radial direction ORI-00 although
the strains decrease with increasing load at a certain load level. This stabilization and the
decrease afterwards of the radial strain can thus be explained by the distribution of the radial
curvature and radial moment. In other words, at a certain load level, the radial moment reaches
its maximum value at this location and does not increase furthermore. The load level at which
the change in behavior occurs is assumed to be at the point at which the yielding moment is
reached at the column face. Afterwards, the moment cannot further increase at the column face.
Additionally, due to the shear forces that define the difference in moment in radial direction, the
moment distribution is given. This provides valuable information for the development of the
analytical model presented in Chapter 7. Figure 6.7b shows the tangential strains at two
locations at the bottom surface of the slab. It can be seen that the strains increase with increasing
load. The same behavior is predicted by the numerical analysis. Although small differences
between the measured and the calculated strains occur for strains near the column (OIT-03), the
general behavior is well captured by the numerical model.
(b)
(a)
2500
2000
calculated
measured
ORI-07
(r=225 mm)
ORI-00
(r=750 mm)
OIT-03
(r=230 mm)
OIT-00
(r=530 mm)
V [kN]
1500
1000
500
tangential direction
radial direction
0
3
Figure 6.7:
1
r []
10
t []
Calculated and measured load-deformation curves for specimen PL7 of surface

deformations in (a) radial direction and (b) in tangential direction (r: distance to the center
of the slab)
121
Chapter 6
Since the calculated strains show generally a good agreement with the measurements, they were
used to calculate the curvatures. Figure 6.8 illustrates the curvature in x-direction x for the three
investigated load levels. As expected, the curvatures are the largest at the column face. With
increasing load, the curvatures increase and are more distributed along the y-axis (tangential
curvatures). In the x-direction (radial direction), the curvature decreases rapidly at the column
face and are already at low levels at the mid-distance between the column and the slab edge.
(b)
(c)
.00
.0
0.005
.01
.01
04
0.08
.01
0
0.015
0.05
0.06
0
.07
0
.0
1
0.0
05
0.0
0.02
0.025
15
05
0.0
3
.0
0.02
0.03
0.04
3
0.035
0.04
0.01
15
0.0
0.03
1
0.02
.0
0.005
5
.02
0 0.
0
.02
.05
0.
0.0
0.0
0.005
01
.
0
0.01
0.02
y
0.01
05
0
.0
0.005
0.05
0.07
05
(a)
Figure 6.8: Distribution of the flexural curvature x (in m-1) for load levels of: (a) 60% VR, (b) 75% VR,
and (c) 90% VR
Figure 6.9 illustrates the distribution of the twisting curvature xy for the three investigated load
levels. As one can expect, the twisting curvature is the largest at the point between the load
introduction points and the column and zero along the x and y axis. With increasing load, the
curvature increases at the corners of the column. The presence of twisting curvatures close to
the slab edge leads to the conclusion that this part of the slab somewhat stiffens the overall slab
behavior.
0.002
1
0.
0
0.00
6
0.00
0.000
.060 4
8
0.016
08
0.0006
0.004
0.
0.002
4 0.01
0.01
0.008
0.006 0.004
0.00
0.006
0.0
0.002
0.008
1
4
0.014
0.0
16
06
0.0
0.0
02
0.0
0.0 04
06
008
0.002
0.012
0.01
0.00
04
0.0
0.002
0.01
16
0.004
0.00
4
14
01
01
08
0.0
04
6
0.00
0.004
0.002
0.002
0.0060.004
0.
0.
0.0
02
02
004
0.0
0.002
04
0.002
0.0
02
x
02
0.00 004
0.002
14
0.0
04
0.002
0.004
012
0.
0.01
0.0
0.008
0.00
1
0.0
0.0
02
0
0.
06
0.0
08
0.0
0.00
0.0
04
0.0
06
0.008
0.01
04
0.0
0.
02
0.0
0.002 0.004
0.00
004
0.01
0.006
04
04
02
0.0
0.0
06
0.
00
8
0.0
0.0
0.002
0.002
0.004
0.00
0.004
0.002
0.004
(c)
0.0
1
(b)
0.006
(a)
0.002
0.00
Figure 6.9: Distribution of the torsional curvature xy (in m-1) for load levels of: (a) 60% VR, (b) 75% VR,
and (c) 90% VR
122
A cut along the x-axis provides more detailed information about the development of the radial
and tangential curvatures and moments. Thus, Figure 6.10 shows the radial and the tangential
curvatures and moments along the axis for the three investigated load levels (60% VR, 75% VR,
and 90% VR). Generally, it can be noted that the radial curvature is maximal close to the column
face and it decreases rapidly until it reaches the value zero at the slab edge. In contrast, the axial
distribution of the tangential curvature is somewhat flatter. Especially, for the lowest
investigated load level, the tangential curvatures are nearly evenly distributed. This is mainly
due to the fact that the yielding strength was not obtained at a load level of 60% VR. At the load
level of 75% VR, the radial as well as the tangential moments reached the flexural strength.
Although plastic redistribution of the moments took place in tangential direction, the curvatures
were still at moderate levels. The radial moment distribution is similar to that of the previously
shown load level of 60% VR. However, due to the yielding of the flexural reinforcement, the
radial curvature has increased. At the load level of 90% VR, it can be noted that while the radial
moment distribution does not significantly change in comparison to the lower load levels, the
tangential moment reaches the flexural strength at a large portion of the slab. Additionally, the
radial and tangential curvatures increase significantly due to the yielding of the reinforcement.
(a)
0.1
r [1/m]
90% VR
75% VR
60% VR
0.05
(b)
m r [kN]
300
200
100
0
(c)
t [1/m]
0.1
0.05
(d)
m t [kN]
300
200
100
0
1500
1000
500
0
r [mm]
500
1000
1500
Figure 6.10: Calculated curvatures along the axis for a load level of 60% VR: (a) radial curvature, (b)
radial moment, (c) tangential curvature, and (d) tangential moment
123
Chapter 6
The axial sections clearly show the plastic moment distribution and thus they indicate the
difference between a linear-elastic and a nonlinear analysis. This difference can also be seen by
comparing the curvature and the moment distribution in the first principal (tangential) direction
as well as from the comparison of the shear fields of a linear-elastic and a nonlinear analysis.
Figure 6.11 illustrates the results of a linear-elastic and nonlinear calculation. Since the
calculations were performed at the same shear force level (90% VR), the curvatures are much
smaller in case of the linear-elastic calculation than in the case of a non-linear calculation. The
linear-elastic calculation leads to evenly distributed moments and curvatures with a decrease in
the direction of the slab corners. In contrast, the non-linear calculation leads to a distribution of
the moments that is largely influenced by the layout of the orthogonal reinforcement. The
tangential moments concentrate along the axis of the slab where the slab reaches its flexural
strength. This moment distribution leads also to a difference in the curvature distribution
compared to the linear-elastic analysis. The same phenomenon can be seen in the analysis of the
shear fields. The shear forces follow much more the orthogonal reinforcement layout in the case
of a nonlinear calculation than in the case of a linear-elastic calculation.
124
linear-elastic
nonlinear
(a)
(b)
0.01
0.0
02
0
03
02
.00
0.002
00
0.
0.0
2
0.0
8
0.01
0.07
.05
0
06
0.0
0.02
0.02
.0
03
0
0.
0.01
00
03
.
0
0.01
0.0
2
0.00
0.01
.0
.
0
3
0.00
.0
.03
5
.04 .0
0 0 .06
0
0.03
05
0.02
0.002
03
00
0.0
0.007
0.
006
0
.00
5
.00
03
.
0
0.0
06
0.0
0.
0.007
.
0
8
0.00
004
6
00
0
.04
05
.0
0.002
0.002
03
0.02
.0
.00
0.0
0.01
0.01
.0
.0
0
.0
04
3
00
02
0.01
(d)
0
8
40
1
14
40
60 0 20
114 1 00
1
240 22
00
80
100
16
1 140 0
20
00
2 220
60
0
18
60
(e)
60
00
100
8
0
0
8
0
6
1 100
20
260
80
20
180
22
0
16
0
14
6 8
0 0
1
0
800
00
0
20
0
30300
280
40
80
80
160
140
1200
0
1
240
80
80
20
300
28
80
300
60
60 0 0
118 200
2
2240 260
280
m1
200
40
2 260
20
20
260
80
00
1
00
140
160
100
2
00
1 160 180
12 40
0
20
40
80
2 0 2300
6
2
320
80
0
12 40
1 60
1 180 00 220
2
20
100
1
160
40
180
200
220
240
240
20
20
1
140
00
100
6
0
8
0
20
1
1420
0
16
0
80
100
0
18
80 0
0
1
20
1 140 0
16
0 0
6 8
(c)
18
0
1 1
20 40
0
0
1
16
40
20
1
100
80
(f)
Figure 6.11: Distribution of the curvature 1 (a,b) and the moment m1 (c,d) in the first principal
(tangential) direction and shear fields including shear distribution along the control
perimeter (e,f) for a linear-elastic (a,c,e) and a nonlinear calculation (b,d,f)
125
Chapter 6
6.3. Local slab behavior

By using the results of the jconc calculations, one can analyze the local slab behavior at the
vicinity of the support area. Besides the in-plane stresses obtained by jconc, the moment was
calculated by the integration of the horizontal stresses along several vertical sections.
Additionally, the stresses in the flexural reinforcement and the strains in the shear reinforcement
were analyzed. Figure 6.12 to Figure 6.14 show the analysis of the local slab behavior for load
levels of 60% VR, 75% VR, and 90% VR.
With respect to the general behavior it can be noted that at the load level of 60% VR, the stresses
are generally low, only at the outermost fiber at the compression zone and at the anchorage zone
of the first and second stud the stresses are close to the concrete strength. With further increase
of the load, the stresses at these zones grow. Additionally, a clear path of the load transfer can
be seen. A portion of the load is taken by the first stud and directly transferred to the column
face. The load transfer of the second row of studs can be described as a compression strut that is
deviated at the first row of studs. In other words, the load descends from the second row in a
rather steep angle to the first row of studs. At this point, a portion of the load is transferred to
the first row of studs. The rest of the load is transferred by a slightly inclined compression strut
to the column face. This is in agreement with the strain measurements in the studs discussed in
Chapter 3. It was presented that the sum of the forces calculated based on the strain
measurements on the top part of first and second row of studs exceeds the total measured shear
force. This indicates that a certain amount of the load in the second row of studs has to be
transferred to the first row of studs.
At the load levels of 75% VR and 90% VR, the general behavior does not significantly change.
However, the plot of the relative stresses clearly shows the development of the failure zone at
the compression strut between the first stud and the column. The maximum relative stresses are
first reached at the intersection of the first stud and the flexural reinforcement before the failure
zone propagates towards the column face. Additionally, large compressive stresses occur at the
bottom side of the slab near the column face.
126

(a)
(b)
400
mr [kN]
300
200
jconc (in-plane)
NLFEA adjusted (bending)
NLFEA nonadjusted (bending)
100
0
(c)
s [MPa]
600
400
200
s []
0
0.25
0.5
s []
0
0.25
0.5
0
0.25
0.5
s []
(d)
0
0.25
0.5
s []
z [m]
0.2
0.15
0.1
0.05
0
0.5
0.4
0.3
0.2
0.1
0
r [m]
0.1
0.2
0.3
0.4
0.5
Figure 6.12: Local behavior at the load level of 60% VR: (a) shear transfer, (b) radial moment, (c) stresses
in the flexural reinforcement, and (d) strains in the shear reinforcement
With respect to the stresses in the top reinforcement and to the moment calculated by the
horizontal concrete stresses and the forces in the bottom reinforcement, it can be noted that the
yielding strength of the flexural reinforcement is reached between load level 60% and 75%.
However, more interesting to note is that yielding of the flexural reinforcement occurs only
between the studs in the first row. Since the compression strut is rather steep, the moment
decreases rapidly between the column face and the first stud. This effect is not considered in the
NLFEA leading to a flatter moment curve. At the intersection point of the first stud and the
flexural reinforcement, the stresses in the flexural reinforcement decrease. This effect can be
explained by the compression strut that adds forces to the inner part of the reinforcement.
Therefore, the stresses outside the first row of stud are reduced.
127
Chapter 6
(a)
(b)
mr [kN]
400
300
200
jconc (in-plane)
100
0
(c)
s [MPa]
600
400
200
s []
0
0.25
0.5
s []
0
0.25
0.5
0
0.25
0.5
s []
(d)
0
0.25
0.5
s []
z [m]
0.2
0.15
0.1
0.05
0
0.5
0.4
0.3
0.2
0.1
0
r [m]
0.1
0.2
0.3
0.4
0.5
Unlike the analysis of the flexural reinforcement, the calculated strains of the transverse
reinforcement have to be regarded with reservations. Since the strains in the studs mainly
depend on the shear cracks and the bond conditions, which both are not properly considered in
the calculation, the results do not resemble the effective strain distribution accurately.
Nevertheless, the calculated values show a qualitative behavior leading to certain general
conclusions. It can be noted that the calculated average strains are much less than the yielding
strains (y2.5). Additionally, it can be seen that the first and the second row of studs exhibit
nearly the same strains. Thus, they transfer nearly the same amount of force. This corresponds
to the strain measurements on the top end of the first and second row of studs (refer to Figure
3.20).
128
(a)
(b)
mr [kN]
400
300
200
jconc (in-plane)
100
0
(c)
s [MPa]
600
400
200
s []
0
0.25
0.5
s []
0
0.25
0.5
0
0.25
0.5
s []
(d)
0
0.25
0.5
s []
z [m]
0.2
0.15
0.1
0.05
0
0.5
0.4
0.3
0.2
0.1
0
r [m]
0.1
0.2
0.3
0.4
0.5
The herein discussed shear transfer at the column vicinity shows also a good agreement with the
experimentally obtained cracking pattern. Figure 6.15 illustrates the cracking pattern observed
at the saw-cut after failure overlaid by the results of the calculation at a load level of 90% VR.
Especially at the right-hand side, the cracks indicate a similar behavior as predicted by the
numerical calculation. The calculated compression strut from the second row of studs seems to
follow the cracks in this area. Similarly, the calculated concrete strut from the first stud to the
column face corresponds to the experimentally observed cracking pattern. Moreover, the
numerical analysis predicts basically two failure areas: one at the bottom surface close to the
column and one close to the top surface at the intersection of the stud and the flexural
reinforcement. Again, both failure zones can be seen at the cracking pattern. At the bottom
surface, the concrete is spalling. However, it has to be noted that although the calculated radial
strains are not small, the spalling seen in the test results from the large tangential strains near the
column face, which are not modeled in jconc. At the intersection between the flexural
reinforcement and the first stud, crushing of concrete occurs at a large area due to lateral tensile
strains. This corresponds to the predicted behavior of the numerical calculation. Thus, the local
model helps to understand qualitatively the behavior of the slab in the vicinity of the column.
129
Chapter 6
Figure 6.15: Calculated stresses at a load level of 90% VR overlaid by the experimentally observed
cracking pattern
Since the global and the local model showed good agreement to the experimentally observed
and measured behavior, the results help to understand the actual response of the slab. However,
the calculations are based on rather coarse simplification and thus the application of this
approach is limited to the investigation of tested slab specimens. For the prediction of the slab
response, a more applicable approach is desired. In the following chapter, an analytical model
will be presented with which the load-rotation response of slabs can be easily predicted.
Additionally, formulations for failure criteria enable the estimation of the punching strength and
rotation capacity.
130
7. Development of an Analytical Model

The experimental observation and the results of the numerical analysis of the tests performed
within this research provide valuable information regarding the slab behavior and failure mode.
Based on these findings, an analytical model was developed allowing a simplified calculation.
Since the punching strength strongly depends on the state of deformation, the analytical model
should consider the load as well as the slab deformation. Thus, the model is based on the CSCT
which defines the punching strength as a function of the slab rotation. However, the
experimental and numerical investigation revealed that several additional aspects should be
considered in the analytical model. With respect to the load-rotation response of the slab, it was
shown that the current Quadrilinear model (Muttoni 2008) predicts a stiffer load-rotation
response compared to the experimentally observed behavior. Thus, the model should enable
more accurate predictions of the load-rotation response of slabs.
An improved load-rotation prediction necessitates also a change of the failure criteria in order to
provide similar or even better predictions for the punching strength and rotation capacity than
the existing formulation. With respect to the failure within the shear-reinforced area, the model
should be based on the mechanical approach of the CSCT (Fernndez Ruiz and Muttoni 2009)
since it showed good agreement with test results. However, the calculations for validation were
performed using a smeared shear reinforcement approach. Therefore, the proposed model
should enhance the applicability of calculations using the actual location of the shear
reinforcement by considering the load transfer path in the column vicinity. With respect to the
failure of the concrete strut, the basic hypothesis of the CSCT that the strength of the
compression strut depend on the transverse strains and thus are related to the state of
deformation should also be applied in the new model. However, additional parameter such as
the amount of compression at the column face or the spacing of the shear reinforcement should
be implemented to improve the prediction for example in the case of small column sizes.
The development of the analytical model presented within this chapter begins with the assumed
slab behavior and the kinematics in the vicinity of the column. Afterwards, equations will be
derived to estimate the slab behavior and to determine the failure mechanism. Finally, several
parameters required for the calculation will be derived and discussed. A numerical example of
the calculation procedure is shown in Appendix B.
131
Chapter 7
7.1. General slab behavior

As already discussed in Chapter 2, an analytical approach necessitates an axisymmetric slab
specimen. Therefore, the derivation of the slab behavior was performed for an axisymmetric
case. The required parameters for the transformation from an axisymmetric to a square slab
identical to the tested specimens are presented in the next chapter. Additionally, it can be noted
that the proposed approach is based on the CSCT (Fernndez Ruiz and Muttoni 2009) and the
Quadrilinear model (Muttoni 2008). Thus, the model considers that the punching failure
depends on the deformation of the slab. Therefore, a crucial aspect is to define the load-rotation
response of the slab.
The basis of such an approach results from the model of Kinnunen and Nylander (Kinnunen and
Nylander 1960) in which it is assumed that the slab is divided into two parts separated by a
shear crack (Figure 7.1a). Additionally, it is assumed that the outer part deforms following a
conical shape with a constant slab rotation. Based on this approach, Muttoni (Muttoni 2008)
derived a formulation to obtain the shear force as a function of the slab rotation by defining the
equilibrium condition at the outer slab segment and by using a quadrilinear moment-curvature
relationship (refer to Chapter 2).
(a)
(b)
(c)
Figure 7.1: Simplified slab behavior proposed by previous research: (a) Kinnunen and Nylander
(Kinnunen and Nylander 1960), (b) Andersson (Andersson 1963), and (c) Shehata and
Regan (Shehata and Regan 1989)
132
Development of an Analytical Model
In presence of shear reinforcement, this general approach has to be somewhat changed.

Andersson (Andersson 1963) proposed to change the shear crack to a flexural crack that opens
at the column face (Figure 7.1b). This approach was adopted and further developed by Gomes
(Gomes and Regan 1999a). He used an approach for slabs without punching shear
reinforcement proposed by Shehata (Shehata and Regan 1989) who introduced an additional
shear crack dividing the slab into three parts namely a column part, a wedge, and an outer
segment. However, Gomes (Gomes and Regan 1999a) assumed that the whole slab except the
column part rotates as a rigid body leading to a similar global slab model as the model proposed
by Andersson (Andersson 1963).
The model proposed within this research project and presented subsequently relies on similar
principles as the aforementioned models. The main concept of the proposed model is the
distinction between a global and a local behavior. The local behavior concerns the shear-critical
region in the column vicinity that is separated to the outer slab segment by an outer shear crack.
The inner part is itself separated by a step crack leading to a slab element over the column and
to a wedge element (Figure 7.2). The global and the local part can be combined by the
equilibrium and compatibility conditions at the outer shear crack area (subsequently referred to
it as intersection plane).
Figure 7.2: Slab divided into three parts: Column part, wedge, and outer slab segment
With respect to the local behavior in the vicinity of the column, the proposed model is based on
the experimental observations and the numerical model. Figure 7.3a illustrates the assumed
deformed shape in comparison to the undeformend shape. It is assumed that the rotations are
performed within an area exceeding the intersection plane whereby large amount of rotation is
performed within the shear crack. Besides the rotations, vertical (shear) deformations occur in
the wedge due to the inclined compression force introduced by the shear reinforcement. The
assumption of such a slab deformation leads to several implications, which can be confirmed
either by measurements or visual observations. Figure 7.3 shows principal experimental
observations such as the detachment of concrete on the top and the spalling of concrete at the
bottom surface, and considerations that are used in the proposed model such as the location of
the shear deformations.
133
Chapter 7
For certain test specimens, spalling of the concrete at the bottom surface occurred at large
rotations. This spalling results from the large tangential compressive strains acting on the
bottom surface near the column. The spalling of concrete on the top surface is more a
detachment of the concrete cover as it can be seen in Figure 7.5. The detachment results from
the rather large shear deformations that are assumed to be performed in the wedge element. The
location of these shear deformations can be determined by the detailed analysis of the vertical
measurements in the column vicinity. The vertical displacement measurements along the axis on
the bottom surface indicate that large shear deformation occurs (refer to Chapter 3).
Additionally, the change in slab thickness measured by the distance between the top and bottom
surface of the slab lead to similar deformations as it was obtained based on the vertical
displacement measurements. However, the elongation of the shear reinforcement predicts a
much smaller displacement.
(a)
Figure 7.3:
(b)
Assumed deformed shape of the slab specimen close to failure
Figure 7.4 shows these three vertical deformations as a function of the slab rotation. While the
change in thickness corresponds directly to the measured value, the shear deformations were
calculated based on the vertical displacement measurements on the bottom surface
(Equation 3.1) and the stud elongations were estimated based on the local strain gauge
measurements on the top side of the first row of studs. Although all presented graphs represent
vertical deformations, each measurement or estimate relies on different influences according to
the proposed model. The calculated shear deformations w depend solely on the shear forces
and are thus independent of the rotations. In contrast, the changes in slab thickness h depend
not only on the shear deformations, since the top surface detaches due to the dowel action of the
flexural reinforcement, but also on the slab rotation that lead to the opening of inclined shear
cracks. Therefore, the measurements of the change in thickness should be larger than the shear
deformations. This statement is confirmed by Figure 7.4.
134
Although it has to be noted that the estimations of the stud elongations are based on coarse
assumptions such as that the strains are constant along the stud and that they are equal to the
measurement on top, they provide valuable information. It can be seen that the estimated stud
elongation are much smaller than the calculated shear deformations at the column face. This
observation is especially noteworthy if one considers that the stud elongation results from the
combination of the shear deformations within the shear cracks that cross the shear reinforcement
and the opening of inclined shear cracks resulting from the slab rotation. Additionally, it can be
noted that the deformation in the studs increases nearly linearly with increasing rotations. These
observations lead to the conclusion that only few shear deformations occurred within the shear
crack. Therefore, it can be concluded that the main deformations in the shear reinforcement
result from the rotation of the slab confirming a principal hypothesis of the proposed model.
(a)
(b)
35
=w (Eq. 3.1)
30
=s1,topls
=h
=s1,topls
[]
25
=w (Eq. 3.1)
=h
20
15
10
0
0
PL9
ls=285 mm
PL7
ls=215 mm
5
2
4
[mm]
6 0
[mm]
Figure 7.4: Rotation-deformation curve for specimens (a) PL7 and (b) PL9 of deformations calculated
based on the strain measurements in the shear reinforcement s assuming a constant strain
along the stud length ls and for shear deformations w calculated based on the vertical
displacement measurements (see Eq. 3.1), and the change in slab thickness h
Generally, unconfined concrete shows a limited deformation capacity. In contrast, well confined
concrete has not only an increased compressive strength but also an increased deformation
capacity so that even strains in the range between 0.01 and 0.02 can occur. Such strains are
required to obtain the measured shear deformations. Thus, it can be concluded that these
deformations occur, without visible failure, at a location where confinement exists. The only
possible location for this is at the bottom side over the column where a triaxial state of
compressive stresses exists leading to an active confinement of the concrete. The rather large
deformations at this area result in a rigid vertical displacement of the outer slab segment and
cause a shortening of the wedge element, which leads to the detachment of the top cover
concrete as it is visible in Figure 7.5.
135
Chapter 7
(a)
(b)
Figure 7.5: Top view after failure of (a) specimen PL6 and (b) specimen PL7
As mentioned previously, it is assumed that the rotations cause the opening of the cracks
separating the three slab elements. The assumption of such a mechanical model leads to several
possible failure locations. With respect to failure, the most critical aspects regarding the
assumed load transfer path are the shear reinforcement crossing the crack, the transverse
(tensile) deformations close to the top surface in the wedge, and the direct shear transfer of the
outer slab segment to the column. The direct shear transfer can be considered as concrete
contribution to the total punching shear strength, similar to the CSCT approach, and as the
contribution of the shear reinforcement in the outer segment that transfers load directly to the
support (Figure 7.6). Besides the shear force that is directly transferred from the outer segment
to the support, a fraction of the shear force is transferred to the wedge by the shear
reinforcement crossing the outer shear crack. This force is either limited by the amount of shear
reinforcement crossing the crack or by the strength of the compression strut transferring the
force to the support. If the sum of the contribution of the concrete, the contribution of the shear
reinforcement outside the outer shear crack, and the contribution of the shear reinforcement
crossed by the outer shear crack is exceeded, the slab fails in punching. The failure mode will be
determined by whether the shear reinforcement (failure within) or the compression strut
(crushing of concrete) reaches its limit first.
Figure 7.6: Load transfer to the support by the shear reinforcement crossing the crack, by the shear
reinforcement outside the crack and by the concrete
136
7.2. Load-rotation response

In the case of slabs without shear reinforcement, the Quadrilinear approach proposed by
Muttoni (Muttoni 2008) presented in Chapter 2, in which the shear force can be calculated using
the equilibrium conditions at the outer slab segment, leads to a good agreement with test results.
The equilibrium condition is given by:
(7.1)
where is the angle of the slab segment, V is the shear force, rq the load application radius, rc
is the radius of the column, mr is the moment acting radially at a distance r0, r0 is the radius of
the shear crack, rs is the radius of the slab, and mt is the tangential moment at a certain position
r.
However, in presence of shear reinforcement additional aspects need to be considered. The

Quadrilinear approach assumes that all the force is directly transferred from the outer slab
segment to the column and thus no shear force occurs at the slab portion within a radius r0. This
assumption is valid for slabs without shear reinforcement. If shear reinforcement is present, one
has to account for shear forces within the zone between the column face rc and radius r0.
Moreover, the load transfer in the column vicinity has to be fully considered. Consequently, the
column region can be defined as a discontinuity region and thus needs to be treated differently
than the outer slab portion. Therefore, the slab behavior is divided into a global and local part.
The outer slab segment can be treated as a slab subjected to bending. In other words, the
assumptions that plane sections remain plane can be applied. In contrast, the response in the
column vicinity is highly influenced by the shear forces. In this area the response is determined
by rigid-plastic stress fields. The change in the calculation is assumed to be at the outer shear
crack at radius r2. The state of deformation at this intersection plane defines the stresses in the
reinforcement at the shear crack and thus the compression forces acting at the outer slab
segment.
7.2.1. Global slab behavior

Assumed rotation
The distribution of the slab rotation is simplified by describing three different regions. The part
over the column has a linear increase in rotation leading to a constant radial and tangential
curvature. The outermost part of the slab starting at a radius r0 is assumed to have a constant
rotation assuming that the outer part undergoes a rigid body displacement in radial direction.
Between the column face and the outer slab segment (between rc and r0), a transition part exists.
In this area, the variation of the rotation is described as a quadratic function. The quadratic
function leads to a linear variation in radial curvature and thus describes the actual behavior
137
Chapter 7
poorly. However, a polynomial function of higher order has not been considered for simplicity.
Nevertheless, for a more accurate calculation, the curvature at radius r2 can be determined
numerically using a more sophisticated rotation function.
Figure 7.7: Assumed simplified slab behavior for the outer slab segment: (a) rotations, (b) radial
curvatures, and (c) tangential curvatures
As mentioned previously, the rotations are defined by a linear relationship over the column, by a
quadratic relationship at a transition zone between the column face rc and r0, and by a constant
value 0 at the outer segment. The quadratic function can be defined as:
(7.2)
where parameter a and c can be determined so that the values of the function and the derivatives
are equal at the boundaries leading to the parameter a equal to:
2
(7.3)
and to the parameter c equal to
2
138
(7.4)
Based on the assumed rotation, one can calculate the curvatures. It has to be noted that for the
global behavior only the curvatures in tangential direction are of interest. In axisymmetric cases,
the tangential curvature is defined by the fact that the curvature is equal to the rotation at a
certain location divided by the radius of this location leading to following expressions:
(7.5)
where 0 is the constant rotation at the outer slab segment.
Flexural relationships
Generally, the pure flexural relationship can be calculated numerically using a similar approach
as it was used for the NLFEA described in Chapter 5. For the analytical calculation of the slab
response, a simplified moment-curvature relationship proposed by Muttoni was used (Muttoni
2008). The relationship is defined by four phases: an uncracked stage, a cracking stage, a
stabilized cracking stage, and a yielded stage. Figure 7.8 shows the different phases in the
moment curvature diagram.
-m
-y
t
-mR
-mR
-p -cr
TS
mt
EI1
-mcr
-mcr
EI0
r2
ry
rp
rcr
rs
-cr
Figure 7.8:
-p
-y
Moment and curvature distribution (adapted from (Muttoni 2008))
Assuming that the influence of the reinforcement can be neglected for the phase before
cracking, the cracking moment can be estimated as:
(7.6)
The cracking moment and the stiffness before cracking can be expressed as:
12
(7.7)
139
Chapter 7
leading to the cracking curvature of:

2
(7.8)
After cracking, the stiffness of the reinforced slab specimen decreases. If a linear-elastic
behavior of the concrete and the reinforcing steel is assumed, the stiffness after cracking can be
estimated as:
(7.9)
where xel is the height of the compression zone calculated with linear-elastic material behavior
given by:
(7.10)
where Es is the Youngs modulus of the reinforcing steel, Ec is the Youngs modulus of the
concrete, is the flexural reinforcement ratio, d is the effective depth and is an efficiency
factor that accounts for the orthogonal reinforcement layout.
If one assumes a rigid-plastic concrete behavior, the flexural strength can be determined by:
(7.11)
where is the flexural reinforcement ratio, d is the effective depth, fy is the yielding strength of
the reinforcement, and fc is the compressive strength of the concrete.
The consideration of a constant influence of the tension stiffening estimated as:
1
6
(7.12)
leads to the curvature in the beginning of the stabilized cracking phase of:
(7.13)
and to the curvature at yielding of:
(7.14)
140
7.2.2. Local slab behavior

The local slab behavior near the support is defined by the state of deformation at section r2
(Figure 7.9) and by the shear forces. Therefore, the first step is to calculate the curvature in
radial direction at the intersection plane r2. The curvature in radial direction can be calculated as
a function of the rotation based on the global slab behavior. However, the numerical analysis
showed that at a certain rotation, the radial curvature at radius r2 stays approximately constant
(refer to Chapter 6). Therefore, the curvature is limited by a value r,lim. A formulation for the
estimation and a further discussion of this curvature will be presented subsequently in the
section of model parameters. Thus, the curvature in radial direction at the intersection plane r2
can be estimated as:
(7.15)
where r2 is the radius of the intersection plane, rc is the radius of the column, a and c are
calculation parameter determined according to Equation 7.3 and Equation 7.4, respectively, and
0 is the rotation at the outer slab segment.
From the state of deformation, one can obtain the stresses in the reinforcement at the
intersection plane r2. Using linear-elastic material relationships, one can obtain the stresses in
the reinforcement based on the curvatures in radial direction r:
(7.16)
where cr is the curvature at cracking (Equation 7.8), p is the curvature at crack stabilization
(Equation 7.13), TS is the difference in curvature due to tension stiffening (Equation 7.12), d is
the effective depth, h is the slab thickness, Es is the Youngs modulus of steel, xel is the height of
the compression zone based on a linear-elastic calculation (Equation 7.10), is an efficiency
factor that accounts for the orthogonal reinforcement layout, and fy is the yielding strength of the
reinforcement.
From the equilibrium condition regarding horizontal forces at the intersection plane, it can be
noted that the sum of the compression and the tensile forces is equal to zero at the intersection
plane for a segment of (Figure 7.9). Thus, it can be noted that:
(7.17)
where Cr2 is the compression force acting at the intersection plane at a slab segment of .
141
Chapter 7
area of compression
in radial direction
area of flexural reinforcement

(dr2)
xb
xc
rs
r2
rc
area of compression
in tangential direction
Figure 7.9:
Area of tensile and compression forces acting at the intersection plane r2
The total compression forces acting at the intersection plane can be divided into a horizontal
part Cb of an inclined compression strut due to the shear force (Fc,b in Figure 7.10b) and a
horizontal part Cc due to bending (Fc,c in Figure 7.10b). The horizontal component of the
compression force Cb depends on the amount of shear force, which is directly transferred from
the outer slab segment to the column, and on the inclination angle of the compression strut.
Thus, it can be expressed as:
1
cot
(7.18)
where is the load distribution factor (Equation 7.70) and b is the inclination angle of the
compression strut. Angle b defines only the ratio between the compression force Cb and the
compression force Cc. In other words, it does not affect the sum of the compression forces Cb
and Cc. Therefore, the variation of angle b has nearly negligible influence on the overall
response of the slab. Thus, a constant value of 26.6 (cot(b) = 2) was chosen based on the
numerical analysis which indicated generally an angle around 20.
The compression force Cc due to bending is defined as the difference of the total compression
force acting at the intersection plane and the compression force Cb. Thus, it is given by:
(7.19)
142
However, it has to be checked that the total compression force acting at the column face does
not lead to tensile stresses in the reinforcement that exceed the yielding strength. Therefore, the
following condition needs to be ensured:
(7.20)
where Ca is the horizontal component of the compression strut in the wedge element. It can be
calculated by the amount of shear force transferred to the wedge element and the inclination of
the compression strut. Thus, the horizontal component of the compression strut in the wedge can
be calculated by:
/2
(7.21)
where is the load distribution factor, V is the applied shear force, r1 the radius of the resultant
force that is transferred across the outer shear crack by the shear reinforcement, rc is the radius
of the column, d is the effective depth, and x1 is height of the compression zone at the column
face.
While the compression force due to bending acts along the whole intersection plane, the
compression force due to the compression strut acts only along the column face. This fact leads
to the different distribution of the compression force in radial and tangential direction (Figure
7.9).
(a)
(b)
(c)
Figure 7.10: Forces acting at each slab element: (a) overview; (b) detailed view of forces acting at the
column face; (c) detailed view of forces acting at the intersection plane at r2
143
Chapter 7
Based on the compression forces, the moments in radial and tangential direction (per unit
length) can be determined by:
(7.22)
and
2
(7.23)
where the heights of the compression zones (x1 and xc) can be approximated by using a rigidplastic stress distribution.
Thus, the height of the compression zone x1 can be estimated as:
(7.24)
and the height of the compression zone xc by:

(7.25)
7.2.3. Equilibrium conditions

The moment equilibrium condition at a free body enables the calculation of the acting shear
forces. Using the column face for defining the equilibrium condition, one can equilibrate the
moment due to the shear force acting at a distance rq and the radial and tangential moments
acting at the free body (Figure 7.11). The moment in tangential direction can be further
separated into the force that acts within the outer shear crack and the moment that acts exterior
the outer shear crack. Therefore, it can be stated:
144
(7.26)
V
2
-mt,ext
-mt,int
-mr
rq rs
r2
rc
V
2
Figure 7.11: Forces and moments acting on a free body segment of the slab
The moment in radial direction acting at the column face is rather straightforward and can be
stated according to Equation 7.22. The tangential part within the shear crack is assumed to be
uniformly distributed between the boundaries r2 that defines the location of the outer shear crack
(intersection plane) and the column face rc.
The integration of the moment in tangential direction exterior the shear crack has to be
separated to the part with a constant rotation (r0-rs) and the part with a quadratic rotation
function (r2-r0). Thus, the integral of the moment in tangential direction outside the shear crack
can be evaluated from:
(7.27)
The moment acting at the part with constant rotation (r0-rs) can be solved using the constitutive
relationship defined previously (see Figure 7.8). The integral of the tangential moment at the
part with constant rotation can be expressed as:
ln
ln
ln
ln
(7.28)
where the following condition applies:

if
, ,
, ,
(7.29)
145
Chapter 7
The constitute relationship defines three different regions at the part with constant rotation.
These regions are:
The region in which the reinforcement is yielding defined by
,
(7.30)
the region where the cracking is stabilized defined by

,
(7.31)
and the region where the concrete is cracked defined by
(7.32)
The tangential moment in the part where it is assumed that the rotation corresponds to a
quadratic function (r2-r0) can be integrated as follows:
(7.33)
where following conditions apply:

if
, ,
, ,
(7.34)
if
, ,
, ,
(7.35)
and
The distances ri can be calculated by:

2
(7.36)
where a and c are calculation parameter determined according to Equation 7.3 and Equation 7.4,
respectively, 0 is the rotation at the outer slab segment, and t,i is the tangential curvature for
which radius ri is searched (y, p, cr).
146
Knowing the radius ri for each change in phase, the integral can be solved independently.
Leading to the expressions:
(7.37)
ln
(7.38)
(7.39)
and
ln
(7.40)
147
Chapter 7
7.3. Failure criteria

Within the research presented herein, two failure modes were investigated: failure within the
shear-reinforced area and failure due to crushing of the compression strut. Thus, for each of this
two failure modes a failure criterion was developed. With respect to the failure within the shearreinforced area, codes such as ACI 318-11, EC2 2004, or MC 2010 and the CSCT use generally
the sum of a certain contribution of the concrete and a certain contribution of the shear
reinforcement to the punching strength. While ACI 318-11 and EC2 2004 uses a fixed
parameter for the concrete contribution, MC 2010 and the CSCT uses a contribution of the
concrete as a function of the slab rotation. Generally, the proposed model uses the same
approach as the CSCT proposed by Fernndez Ruiz and Muttoni (Fernndez Ruiz and Muttoni
2009) for failure within the shear-reinforced area. However, certain modifications were made
mainly for the calculation of the crack width and the amount of the shear reinforcement that is
contributing to the punching strength. Both modifications can be attributed to the different
assumption regarding the location and the inclination of the shear crack crossing the shear
reinforcement. The used assumptions of the location and the inclination of the shear crack are
discussed in detail in subchapter 7.4.2.
If large amount of shear reinforcement is present, failure due to crushing of the concrete strut
may occur. Generally, codes such as ACI 318-11, NAD 2011, or MC 2010 and the CSCT
increase the punching strength of slabs without shear reinforcement, in the case of MC 2010 and
the CSCT defined as a function of the rotation, by a certain factor. However, the different
provisions have different justifications for the use of such an approach. In the case of
NAD 2011, the approach is based on the conclusion that the resistance is defined by the strength
of the concrete compression zone at the column face (Beutel 2003; Husler 2009). The main
parameters defining the strength of the concrete compression zone are the effective depth, the
flexural reinforcement ratio and the compressive strength of concrete. These parameters are all
considered in the formulation of punching of slabs without shear reinforcement so that the
increase of the punching strength of slabs without shear reinforcement by a certain factor was
justified.
In the case of the CSCT and MC 2010, the approach is based on the main assumption that the
crushing strength of the concrete depends on the state of transverse strains in the compression
strut (Fernndez Ruiz and Muttoni 2009; Muttoni and Fernndez Ruiz 2010). Since the state of
transverse strains can be described as a function of the opening of cracks in the shear-critical
region, which corresponds to the hypothesis of the formulation of the CSCT for slabs without
shear reinforcement, the use of the same formulation is proposed as basis for the failure criterion
for crushing of concrete. Additionally, the width of the crushing zone is assumed to depend on
the aggregate size (Muttoni and Fernndez Ruiz 2010), for which the formulation of the CSCT
for slabs without shear reinforcement also accounts for. The multiplication factor used to
increase the punching strength of slabs without shear reinforcement takes into account of the
performance of the anchorage and the distribution of the shear reinforcement system.
148
The failure criterion for crushing of the concrete strut proposed within this research relies
basically on the assumed load transfer path in the vicinity of the column. The assumed load
transfer corresponds to the load transfer proposed for failure within the shear-reinforced area, by
which the concrete, the shear reinforcement outside the shear crack, and the shear reinforcement
crossing the shear crack contributes to the punching strength. In other words, the proposed
model considers that only a portion of the force is transferred by the compression strut close to
the column whereby this force is directly supported by the column (Figure 7.12c,d). With
increasing load, the force in the compression strut increases until it becomes too large and the
concrete starts crushing at the anchorage zone of the shear reinforcement. At this moment, the
direct support of the force is interrupted so that the compression strut has to be deviated (Figure
7.12e,f). However, this deviation leads to large tensile forces transverse to the compression strut
leading immediately to failure. This explains the immediate loss of strength that was
experimentally observed without any softening which one would expect in the case of a
compression failure.
Figure 7.12: Load transfer and failure mechanism shown for PL6 (a, c, e) and PL7 (b, d, f): (a, b)
Detailed view of the crack pattern, (c, d) load transfer until failure, and (e, f) change in load
transfer at failure
149
Chapter 7
As discussed previously, the total shear force can be defined as the sum of the contribution of
the shear reinforcement crossing the crack, the contribution of the shear reinforcement outside
the crack, and the contribution of the concrete. Therefore, it can be stated that:
(7.41)
The concrete contribution is assumed to be the same as for slabs without shear reinforcement as
proposed by the CSCT approach (Muttoni 2008; Fernndez Ruiz and Muttoni 2009). Therefore,
the maximum concrete contribution can be expressed as:
3
4
1
15
(7.42)
16 mm, dg is the aggregate size, fc is the compressive strength of concrete, and b0 is the control
perimeter set at a distance of 0.5d from the support region with circular corners.
The sum of the contribution of the shear reinforcement can be defined by the load distribution
factor for the shear reinforcement s, which defines the portion of the shear force that is
transferred across the crack by the shear reinforcement and afterwards transferred to the column
by the compression strut. The use of this factor leads to the following expression:
(7.43)
where Vs1 is the shear force transferred across the crack.
Failure occurs when the applied shear force V is larger than the shear strength defined by:
V
(7.44)
The maximum concrete contribution is defined by Equation 7.42 and the maximum shear force
transferred across the crack (VRs) is either limited by the amount of shear reinforcement (failure
within) or by the compression strut (crushing of concrete).
150
For a failure within the shear-reinforced area, an approach similar to the CSCT (Fernndez Ruiz
and Muttoni 2009) is used. Therefore, the maximum shear force transferred across the crack VRs
is calculated by the product of the area of shear reinforcement crossed by the crack (Asw) and the
stresses in the shear reinforcement as a function of the rotation w(). Thus, the maximum shear
force transferred through the crack VRs,I is given by:
(7.45)
For the calculation of the stresses in the shear reinforcement, one requires the crack width at the
location of the shear reinforcement. According to Figure 7.13a, the crack width w1 can be
estimated as:
cos
cos
(7.46)
where defines the ratio of the rotation performed at this crack to the total slab rotation.
Assuming that approximately one third of the total rotation is performed within this crack, can
be estimated as:
0.35
(7.47)
This factor can be associated to the factor used in the CSCT (Fernndez Ruiz and Muttoni
2009) that was empirically determined to be CSCT = 0.5. However, a difference in the value
exists since the CSCT approach defines the crack width as:
cos
(7.48)
where is the rotation of the slab and h1 is the vertical distance from the crack tip to the point
where the crack crosses the shear reinforcement.
Assuming a crack inclination of 45 (/4), s0 can be stated as:
tan
(7.49)
leading to the expression:
cos
0.5 cos
0.35
(7.50)
If more than one vertical branch of shear reinforcement crosses the crack, the crack opening can
be determined similarly at each intersection point (Figure 7.13b). This can occur if the spacing
of the shear reinforcement is rather close as it was the case of the tests with continuous stirrups
(PF1-PF5).
151
Chapter 7
(a)
(b)
Figure 7.13: Estimation of the crack width at the location of the shear reinforcement for (a) cases by
which only one vertical branch of the shear reinforcement crosses the crack and (b) cases by
which more than one vertical branch of the shear reinforcement crosses the crack.
The crack width at the location of the shear reinforcement leads to the stresses in the
reinforcement. These stresses can be calculated using the approach of the CSCT (Fernndez
Ruiz and Muttoni 2009) that distinguishes between three cases (described in detail in
Chapter 2).
Case 1: If
4
(7.51)
reinforcement, and lai is the shorter distance between the crack and one end of the shear
reinforcement (lai = min(lbi; lbs))
then
4
(7.52)
reinforcement, wi is the crack opening, and fyw is the yielding strength of the shear
reinforcement.
152
Case 2: If
4
(7.53)
the shear reinforcement (las = max(lbi; lbs))
then

2
(7.54)
Case 3: If
2
(7.55)
the shear reinforcement (las = max(lbi; lbs))
then
(7.56)
153
Chapter 7
For failure due to crushing of the concrete strut, the assumed failure mechanism leads to a set of
parameters that must be considered. The main parameters to consider are the transverse strain in
the compression strut, the area of the compression strut, and the anchorage conditions. Thus, the
strength of the compression strut can generally be expressed as:
sin
(7.57)
where k1 is a factor accounting for the reduction of the strength due to transverse strains, k2 is a
factor accounting for the stress distribution within the compression strut depending on the
anchorage of the shear reinforcement, fc is the compressive strength, Ac is the sectional area of
the compression strut.
Factor k1 that accounts for the transverse strains can be calculated based on the approach
proposed by Vecchio and Collins (Vecchio and Collins 1986), which is given by:
,
0.8
(7.58)
170
where 1 is the tensile strain transverse to the compression strut.

1
2
2
1
Figure 7.14: Strains acting on an element within the compression strut
Based on Equation 7.58, the reduction factor accounting for the softening due to tensile strains
transverse to the compression strut can be expressed as:
1
(7.59)
where 1 is the average tensile strain transverse to the compression strut within the outer shear
crack and 1 and 2 are constant parameters.
The average tensile strains transverse to the compression strut mainly depends on the opening of
cracks within the outer shear crack. Thus, it can be assumed that the average tensile strains are
proportional to the rotation of the slab at the location of the compression strut, which is assumed
to be proportional to the rotation at the outer slab segment, leading to following expression:
154
(7.60)
where 0 is the rotation at the outer slab segment, d is the effective depth, x2 is the height of the
compression zone at the outer slab segment (Figure 7.14), and r2 is the radius of the outer shear
crack.
By combining Equation 7.59 and 7.60 and estimating the constants 1 and 2, one can express
the reduction factor k1 as:
1
1
135
(7.61)
where 0 is the rotation at the outer slab segment, d is the effective depth, x2 is the height of the
compression zone at the outer slab segment (Figure 7.14), and r2 is the radius of the outer shear
crack.
The strength of the compressive strength depends not only on the softening of the concrete due
to transverse strains but also on the maximum area of the compression strut and the stress
distribution within the strut. The cross-sectional area of the compression strut is given by the
perimeter and the width of the compression strut. While the perimeter is geometrically defined
for each location along the compression strut, the maximum width of the compression strut
cannot be directly determined. The maximum width of the compression strut may be defined
considering material properties such as the maximum aggregate size or geometrical conditions
such as the spacing of the shear reinforcement. For example, the CSCT defines the width of the
crushing zone as a function of the aggregate size (Muttoni and Fernndez Ruiz 2010). In
contrast, the model proposed herein considers the location of the shear reinforcement as the
principal parameter defining the maximum width of the compression strut.
Generally, the maximum width of the compression strut can be expressed by the horizontal
width and the inclination angle of the strut. Thus, one can define the maximum width of the
compression strut as:
sin
(7.62)
where sc is the horizontal width of the compression strut and 1 the inclination angle of the strut
(Figure 7.15).
The horizontal width of the compression strut is defined by the location and the inclination of
the compression strut as well as by the location and the inclination of the shear crack. Generally,
two cases can occur, the width of the compression strut is limited by the outermost vertical
branch of shear reinforcement crossing the shear crack (Figure 7.15a) or by the shear crack
(Figure 7.15b). The first case occurs if the inclination angle is rather steep, as it is the case if
only one vertical branch crosses the shear crack, and the compression zone is moderate. If more
155
Chapter 7
than one vertical branch crosses the shear crack, the width is generally limited by the shear
crack. Similarly, if the compression zone becomes large as it is the case of small columns the
width of the compression strut is also limited by the shear crack. Thus, the proposed model also
accounts for the amount of compression acting along the column perimeter.
(a)
(b)
Figure 7.15: Failure of the compression strut in the case of (a) one row of shear reinforcement crossing
the crack and (b) in the case of multiple rows of shear reinforcement crosses the crack
The stresses in the compression strut near the anchorage of the shear reinforcement cannot be
assumed to be evenly distributed along the width of the strut (Figure 7.16). Thus, the maximum
strength is reduced by a factor k2. This factor depends on the anchorage conditions of the shear
reinforcement system since it is assumed that good anchorage conditions enable a better load
distribution for the load introduction into the compression strut. Therefore, the better the
anchorage is the higher the value k2 is. Good agreement with test results can be obtained for a
value of k2 = 0.75 for double headed studs and for a value of k2 = 0.50 for stirrups.
Figure 7.16: Stress distribution near the anchorage zone of the shear reinforcement
Factors k1 and k2 as well as the maximum cross-sectional area of the compression strut are not
constant along the strut length. Factor k1 varies due to the variation of the transverse strain 1. If
it is assumed that the transverse strain varies linearly, one can obtain the distribution of factor k1
(Figure 7.17a). Factor k2, representing the variation in stresses along the strut width due to the
anchorage condition, is assumed to vary linearly along the compression strut (Figure 7.17b).
Thus, k2 is minimal at the anchorage zone and equal to one at the bottom end of the strut, which
means that the stresses are assumed to be evenly distributed along the strut width at this
location.
156
The cross-sectional area of the compression strut is given by the geometry and by the applied
shear force. On the bottom end of the strut, the area is defined by the applied shear forces and
by the compressive strength of the concrete. The area at this location increases with increasing
load since generally no geometrical boundaries occur. Along the compression strut, the area
increases up to the height at which the shear reinforcement crosses the crack. Since it is
assumed that the compression forces are only deviated within the boundaries of the
reinforcement introducing the compression force (flexural reinforcement and shear
reinforcement), the area is limited by the shear reinforcement. Therefore, the area is nearly
constant above this point (Figure 7.17c).
b) k2
2
0.4
d) VRs1
1.0
1.0
c) Ah
0.7
5
a) k1
Figure 7.17: Distribution of the shear resistance VRs1 (d) and its components (a) k1, (b) k2, and (c) Ah (Ah:
Area of a horizontal section Acsin(1)) along the compression strut for specimen PL7
These parameters lead to the shear strength within the compression strut (Figure 7.17d). It has to
be noted that the shear strength on the bottom end of the strut was chosen as equal as the applied
shear force since no geometrical boundaries occur. In contrast, the shear force on the top end is
limited by the parameters k1 and k2 and the geometrical conditions. Therefore, the proposed
model assumes that the weakest part in the compression strut is at the top end close to the point
where the load is introduced by the shear reinforcement.
Based on the previously mentioned considerations, the following failure criterion can be
derived:
,
sin
2
sin
(7.63)
factor accounting for the stress distribution within the strut, fc is the compressive strength, rc is
the radius of the column, bw is the width of the compression strut, and 1 is the inclination of the
compression strut.
157
Chapter 7
Introducing Equation 7.62 into Equation 7.63, the shear strength VRs can be expressed as:
(7.64)
factor accounting for the stress distribution within the strut, fc is the compressive strength, rc is
the radius of the column, and sc is the horizontal width of the compression strut.
7.4. Definition of model parameters

7.4.1. Limitation of the radial curvature r,lim at radius r2
The numerical analysis presented in Chapters 5 and 6 showed that at a certain rotation, the
radial curvature at radius r2 stays approximately constant (Figure 7.18). In other words, the
curvature in radial direction at radius r2 does not exceed a certain limit r,lim.
PL7
90% VR
75% VR
60% VR
r [1/m]
0.1
0.05
0
1500
1000
500 -r2
-rc
0
rc
r [mm]
r2 500
1000
1500
Figure 7.18: Curvature in radial direction for specimen PL7 calculated with the NLFEA approach
presented in Chapter 6
This limitation can be explained with the analysis of the equilibrium conditions within this zone.
From the equilibrium condition of the moments acting on a discrete slab element (Figure 7.19),
one obtains the following expression:
r
viri
mtir
(mri+m)(ri+r)
mriri
(vi+v)(ri+r)
ri
r
mtir
Figure 7.19: Forces acting on a discrete element of an axisymmetric slab
158
(7.65)
Solving Equation 7.65 for the difference in radial moment leads to the expression:
(7.66)
Assuming that the moment in radial mr and tangential mt direction reached the flexural capacity
of the section mR, Equation 7.66 can be simplified to:
(7.67)
In order to fulfill Equation 7.67, it can be stated that either the difference in moment is zero and
the shear forces are zero as well or shear force is present and the difference in moment is
unequal to zero. Assuming that the resultant shear force transferred across the crack acting at
radius r1 (Figure 7.20a) is directly supported by the column, the shear force is equal to zero
within a cone of radius r1. Thus, within this cone the first statement applies leading to a constant
moment in radial direction. On the other hand, outside radius r1, shear forces are present and
thus the second statement applies. Therefore, the absolute value of the radial moment at radius
r2 is smaller than the absolute value of the radial moment at radius r1 (Figure 7.20c).
Thus, it is assumed that the limit of the curvature in radial direction r,lim at location r2 is reached
when the radial curvature at location r1 exceeds the yielding curvature (Figure 7.20b). The
curvature at r1 can be estimated using the simplified rotation distribution (and thus curvature
distribution) used for the analysis of the global slab response. Therefore, the limit of the radial
curvature can be expressed as:
,
(7.68)
where y is calculated with Equation 7.14.

a)
b)
c)
Figure 7.20: Limitation of the radial curvature at position r2 using the simplified global slab behavior
159
Chapter 7
7.4.2. Radius r0
Generally, radius r0 can be described as the location beyond which the rotation can be assumed
to be constant. Thus, the exact location should be determined at a radius at which the curvature
in radial direction becomes small. One possible location could be at a radius at which the
section remains uncracked since the curvatures of an uncracked section are very small.
However, such a distance would not lead to a good agreement with the actual response since
several other conditions of the choice of r0 need to be considered. One condition is that the
simplified rotation function (quadratic function) leads to a linear variation of the curvature in
radial direction (Figure 7.7b). Therefore, the radius r0 need to be chosen so that this
simplification represents the distribution of the curvature best. Another condition is that the
model should represent the limitation of the curvature in radial direction at position r2
accurately. In order to fulfill these three conditions most accurately, the radius r0 was chosen at
a distance of 1.5d from the column face but minimal at a distance of 0.5d from the shear crack
radius r2. Thus, radius r0 is given by:
1.5
0.5
(7.69)
In order to justify this value, it can be noted that this choice of the distance r0 corresponds also
to the experimentally observed principal area of rotations stated in literature (Husler 2009).
Additionally, the analytical approach can be compared to a semi-numerical approach. For the
analytical approach, the load-rotation response was calculated using a curvature in radial
direction at location r2 calculated using Equation 7.15 with r0 according to Equation 7.69. The
semi-numerical approach uses the curvature in radial direction at location r2 calculated by the
NLFEA described in Chapter 5 and 6. Figure 7.21 shows the two predicted load-rotation
responses of slabs with different column sizes and different thicknesses. It can be seen that for
specimens with a thickness of h = 250 mm, the analytical (Equation 7.68) and the seminumerical (NLFEA) calculations correspond well. Only for the specimen with a large thickness
(h = 400 mm), a difference occurs. However, it can be noted that the maximal difference at the
yielding plateau is rather small (<10%). Therefore, it can be concluded that the choice of the
distance r0 and the estimation of the limitation in radial curvature r,lim at radius r2 is reasonable.
160

(a)
(b)
MPa
1.5
h=250 mm
c=130 mm
(c)
h=250 mm
c=260 mm
analytical
h=400 mm
c=440 mm
analytical
semi-numerical
semi-numerical
analytical
semi-numerical
0.5
0
0
0.1
0.2
0.3 0
0.1
0.2
0.3 0
0.1
0.2
0.3
Figure 7.21: Load-rotation curve calculated by the proposed model using the radial curvature determined
by Equation 7.15 and determined by the NLFEA for (a) a specimen with h = 250 mm and
c = 130 mm (PL6), (b) a specimen with h = 250 mm and c = 260 mm (PL7), and (c) a
specimen with h = 400 mm and c = 440 mm (PL10)
7.4.3. Shear crack distance r2

The distance r2 from the center of the slab to the location where the outermost shear crack
intersects the flexural reinforcement is another main parameter of the proposed model.
Primarily, two assumptions lead to the determination of the distance r2. The first assumption is
that the outer shear crack starts at the top end of the compression zone x2 and has an inclination
of approximately 30. The second assumption is that the shear crack passes the intersection
point of the flexural reinforcement and a vertical branch of shear reinforcement. This can be
explained by the work that has to be performed. Since more work has to be performed to open a
crack that crosses shear reinforcement, the crack is supposed to run until a vertical branch of
shear reinforcement without crossing them. Based on these assumptions, it can be noted that in
the case of studs with commonly used spacing, the crack and thus the distance r2 arrives at the
second row of studs without crossing them (Figure 7.22a). In the case of continuous stirrups
with relatively close spacing, the crack may cross one or more rows of vertical branches (Figure
7.22b and c).
(a)
(b)
(c)
Figure 7.22: Comparison of the assumed outer shear crack and the experimentally obtained crack pattern
for (a) specimen PL11, (b) specimen PF1, and (c) specimen PF2
161
Chapter 7
Radius r2 basically influences the load-rotation curve and both failure criteria of the proposed
model. While the load-rotation response is only slightly influenced by the distance r2, the
influence of the failure criteria especially for failure within the shear-reinforced area may be
significant. This can be explained by the fact that radius r2 defines in conjunction with the
spacing of the shear reinforcement the amount of shear reinforcement crossing the shear crack.
Thus, the prediction of the punching strength is strongly dependent on r2 and the actual location
of the shear reinforcement defined by the spacing.
For the CSCT, the influence of the spacing and the actual location of the shear reinforcement
was limited by using a smeared shear reinforcement approach. Thus, a shear reinforcement ratio
was determined and used in the calculations. However, the drawback of such an approach is that
although the model allows considering the bond condition and the activation of each vertical
branch independently, the smeared shear reinforcement approach leads to an average activation.
Consequently, a main advantage of the mechanical model, the modeling of an individual
activation of each vertical branch of shear reinforcement, is lost.
Using the actual location of the shear reinforcement however is sensitive to the spacing of the
shear reinforcement. In fact, the change in the predicted strength is rather large if only one
vertical branch is crossed by the assumed shear crack or if the assumed shear crack crosses more
than one vertical branch of shear reinforcement. Figure 7.23 shows the normalized punching
strength as a function of the shear reinforcement spacing for a specimen corresponding to PL12.
The dashed lines show the normalized punching strength calculated with the CSCT using a
smeared shear reinforcement approach and using the actual location of shear reinforcement.
(a)
(b)
MPa
1.5
s1 =2s0
CSCT (ncut=2)
Proposed model
(ncut=2)
1.0
0.5
CSCT (smeared)
CSCT (ncut=1)
Proposed model (ncut=1)

0.0
0.5
0.6
0.7
0.8
s1 /d [-]
0.9
1.0
Figure 7.23: (a) Predicted punching strength as a function of the spacing of the shear reinforcement
calculated with the proposed model, with the CSCT using a smeared shear reinforcement
approach, and with the CSCT using the effective location of the shear reinforcement; (b)
One or two vertical branches of shear reinforcement crossing the crack assuming a crack
pattern according to the proposed model and the CSCT
162
The model proposed herein considers the spacing in the calculation of the shear force
transferred across the shear crack leading to a variation of the normalized punching strength as a
function of the spacing of the shear reinforcement. Furthermore, a variable angle of the shear
crack reduces the magnitude of change in strength if an additional row of a vertical branch of
shear reinforcement crosses the assumed shear crack (Figure 7.23). Similar to the smeared shear
reinforcement approach by the CSCT, the proposed model predicts lower normalized punching
strength for larger spacing. The punching strength decreases with increasing spacing since the
model assumes that less force is directly transferred from outside the shear crack to the column.
7.4.4. Load distribution factor

The load distribution factor depends on several factors such as the shear reinforcement layout,
the rotation, or the concrete strength. Therefore, an exact calculation of this value requires a
rather complicated approach based on the compatibility conditions in the column vicinity and
the anchorage condition of the shear reinforcement. However, although this factor is used for
several estimations within this model, the influence on the load-rotation response is fairly
limited as shown later. Thus, a rather simple estimation of the load transfer factor is proposed
that considers an estimated concrete contribution and the spacing of the studs with respect to the
effective depth.
Factor is defined as the ratio of the shear force contributed by the shear reinforcement crossing
the crack to the total amount of shear force:
(7.70)
This factor can be expressed as the product of the factors c and s:
(7.71)
where factor c is the contribution of the shear reinforcement to the total shear force. In other
words, c is the ratio of the amount of shear force without the concrete contribution to the total
amount of shear force. Thus, it can be expressed as:
(7.72)
Factor s is the ratio of the shear force contributed by the shear reinforcement crossing the outer
crack to the total amount of shear force contributed by the shear reinforcement. Thus, it can be
expressed as:
(7.73)
163
Chapter 7
Factor c can be estimated by assuming that it corresponds to the ratio of the predicted to the
shear strength at a certain rotation to the predicted total shear strength at this rotation. Thus,
factor c can be estimated based on the following expression:
(7.74)
where VRc is defined by Equation 7.42 and VR by Equation 7.44. However, since the maximum
strength VR depends also on the factor , the factor c must be determined iteratively.
The force distribution between the shear reinforcement within and outside the crack can be
estimated by accounting for the distances between the shear reinforcement. For general cases by
which only one vertical branch of shear reinforcement crosses the outer shear crack (ncut = 1),
this value can be estimated as:
2
1.0
(7.75)
where s0 is the distance between the column face and the first vertical branch of shear
reinforcement, s1 is the distance between two adjacent vertical branches of shear reinforcement
of same radius, and d is the effective depth.
However, if the spacing between the shear reinforcement is small, it is assumed that the shear
reinforcement crossing the outer shear crack provides all the contribution of the shear
reinforcement. This is proposed to be the case if more than one vertical branch of shear
reinforcement crosses the outer shear crack (ncut > 1). Thus, in these cases, factor s becomes:
1.0
(7.76)
Figure 7.24 shows the comparison between the calculated load distribution factor and the one
that was estimated based on the strain measurements on the shear reinforcement according to
the method described in Chapter 3. It can be seen that for low load and rotation levels, a
difference may occur. However, at failure the calculated factors correspond well with the
estimates that are based on the measurements. Nevertheless, it has to be noted that the estimated
load distribution factors show only a trend since they are based on local strain measurements on
the top end of the first row of shear studs (refer to Subchapter 3.6.7).
164
(a)
(b)
1
PL6
(c)
PL7
PL10
0.9
0.8
estimated based on measurements
[-]
0.7
calculated
0.6
calculated
calculated
0.5
0.4
0.3
0.2
0.1
0
0
10
20
30
[]
40
50 0
10
20
30
[]
40
50 0
10
20
30
[]
40
50
Figure 7.24: Load distribution factor estimations based on the strain measurements on the shear
reinforcement and calculated as a function of the slab rotation for specimens (a) PL6, (b)
PL7, and (c) PL10
Generally, it can be noted that the influence of value on the calculated load-rotation response
is rather small. Figure 7.25 shows the calculated load-rotation curve for specimens with
different thicknesses and column sizes for values of = 0.60, = 0.75, and = 0.90. It can be
seen that only small differences occur.
(a)
(b)
MPa
1.5
h=250 mm
c=130 mm
(c)
h=250 mm
c=260 mm
h=400 mm
c=440 mm
=0.60
=0.75
=0.90
=0.60
=0.75
=0.90
=0.60
=0.75
=0.90
0.5
0
0
0.1
0.2
0.3 0
0.1
0.2
0.3 0
0.1
0.2
0.3
Figure 7.25: Calculated load-rotation curve for different values of for (a) a specimen with h = 250 mm
and c = 130 mm (PL6), (b) a specimen with h = 250 mm and c = 260 mm (PL7), and (c) a
specimen with h = 400 mm and c = 440 mm (PL10)
165
Chapter 7
With respect to the two failure criteria, the partial factor c has a rather limited influence.
However, the influence of the partial factor s, which accounts for load distribution between the
different rows of shear reinforcement, is significant. Figure 7.26 shows the failure criteria for
failure within the shear-reinforced area and failure of the compression strut for different values
of s. The different values lead to a rather large bandwidth for both failure criteria especially for
failure within the shear-reinforced area. Therefore, the calculation of the punching strength may
be sensitive to this value.
(a)
(b)
2
h=250 mm
c=130 mm
1.8
MPa
(c)
h=250 mm
c=260 mm
h=400 mm
c=440 mm
1.6
1.4
1.2
1
0.8
0.6
0.4
VR,I
s=0.60
s=0.75
s=0.90
VR,II
s=0.60
s=0.75
s=0.90
0.2
0
0
0.1
VR,I
s=0.60
s=0.75
s=0.90
0.2
0.3 0
0.1
VR,I
s=0.60
s=0.75
s=0.90
VR,II
s=0.60
s=0.75
s=0.90
0.2
0.3 0
0.1
VR,II
s=0.60
s=0.75
s=0.90
0.2
0.3
Figure 7.26: Calculated failure criteria for different values of for (a) a specimen with h = 250 mm and
c = 130 mm (PL6), (b) a specimen with h = 250 mm and c = 260 mm (PL7), and (c) a
specimen with h = 400 mm and c = 440 mm (PL10); (VR,I: failure within shear reinforced
area; VR,II: failure of compression strut)
In order to verify the estimated value used in the calculation, one can compare this value again
with the shear force in the stud estimated with the strain measurements. However, this factor
cannot be obtained directly from the measurement but it can be validated by the force in the
shear reinforcement. Therefore, the value of the force in the first row of studs at failure was
calculated based on the local strain measurements in the shear reinforcement and based on the
proposed formulations that use factor s.
166
According to the proposed model, the total force taken by the shear reinforcement is given by:
(7.77)
where VR(R) is the measured punching strength and VRc(R) is defined as:
3
4
1
15
(7.78)
The forces in the shear reinforcement crossing the outer shear crack at failure can be calculated
by:
(7.79)
where the calculated value Vs1(R) can be compared to the value estimated from the strain
measurements in the shear reinforcement.
Table 7.1 shows the calculated values and the values estimated based on the measurements for
the specimen PL6, PL7, and PL10. It can be seen that generally the calculated forces in the
shear reinforcement correspond well to the forces that were estimated based on the strain
measurements. However, it has to be considered that the forces from the measurements are
obtained indirectly by certain assumptions (see Chapter 3) and thus they involve uncertainties
itself.
Table 7.1:
Specimen
PL6
PL7
PL10
Calculated and experimentally obtained forces in the first row of studs at failure
VR(R)
[kN]
1363
1773
5193
R
[]
18.6
32
18.0
VRc(R)
[kN]
376
371
1077
Vs(R)
[kN]
987
1402
4116
s,model
[-]
0.81
0.81
0.76
Vs1,calc(R)
[kN]
799
1135
3128
Vs1,measured(R)
[kN]
697
1124
3424
167
8. Validation of the Analytical Model

The comparison of the results from the analytical model to the experimentally obtained results
enables the verification of the performance of the proposed model. The comparison consists of
the tests performed within this research and of tests available in literature. All the test specimens
within this research and the selected test specimens from literature have an orthogonal flexural
reinforcement layout and most tests specimens are square slab. Therefore, all test specimens in
the selection for the validation were non-axisymmetric. However, since the model applies only
for fully axisymmetric slabs, the model requires a set of parameters for the slab transformation
from a non-axisymmetric to an axisymmetric slab. Therefore, the first part of this chapter covers
the transformation of the slab specimens. Afterwards, it presents the performance of the
proposed model with respect to the strength and rotation prediction and the influence of
different parameter namely the slab thickness, the column size, and the shear reinforcement
ratio. Finally, the tests were compared to the same selection of tests used in Chapter 4 to
investigate the overall performance of the proposed model. A list of all test specimens
considered can be found in Appendix A.
169
Chapter 8
8.1. Transformation of the specimen

At first, the transition from the geometry of the test specimens to an axisymmetric model will be
discussed. The main critical aspects to investigate are the influences of the orthogonal flexural
reinforcement layout, the change from a square to a circular column shape, the change from a
square to a circular slab specimen, and the application of the load. Figure 8.1 shows the
transition that has to be made from the tested specimens (a,c) to an axisymmetric slab specimens
(b,d). Afterwards, this chapter presents an approach to estimate the load-rotation behavior of a
non-axisymmetrical slab.
(a)
(b)
(c)
(d)
Figure 8.1: Transformation from a square slab as tested (a,c) into an axisymmetric slab (b,d)
Table 8.1 summarizes the geometry and the material properties used for the calculations. The
parameters of the square slab correspond to test specimen PL7 with the exception that a flexural
reinforcement of 1.0% was used instead of 1.5%. The smaller flexural reinforcement allows the
slab to undergo larger deformation so that it is assured that the calculated load-rotation curve
reaches the point where yielding of the flexural reinforcement starts to occur. Additionally,
Table 8.1 shows the parameters of an equivalent axisymmetric slab. The determination of these
parameters will be discussed subsequently.
170
Validation of the Analytical Model

Table 8.1:
Properties of the analyzed specimens unless noted otherwise
Specimen
Thickness h
Flexural reinforcement ratio L
Column size c
Effective depth d (average of both directions)
c/d ratio
Slab dimension l; ds
Concrete compressive strength fc
Yielding strength fsy
Square
250 mm
1.00%
260 mm
210 mm
1.24
3000 mm
30 MPa
500 MPa
Axisymmetric
250 mm
1.00%
333 mm
210 mm
1.59
2992 mm
30 MPa
500 MPa
8.1.1. Influence of the orthogonal reinforcement

A full axisymmetric model necessitates the use of circularly and radially arranged
reinforcement. Since such a layout is, due to obvious reasons, not used in practice, the
reinforcement is in most experimental tests orthogonally arranged. Therefore, the model should
be able to account for a non-axisymmetric reinforcement layout that leads to a softening of the
section at locations where the principle directions of the moment are not in the direction of the
reinforcement (Figure 8.2). This phenomenon corresponds to the reduced torsional stiffness
already discussed in Chapter 5. In order to account for the reduced torsional stiffness, Kinnunen
(Kinnunen 1963) introduced a reduction factor. A similar factor was later used by Guandalini
(Guandalini 2005) and Muttoni (Muttoni 2008) whereby it was either analytically or empirically
obtained. For the analytical determination, the stiffness of a tension chord was calculated as a
function of the flexural reinforcement ratio for a reinforcement layout parallel to the applied
force and for a reinforcement layout at an angle of 45 to the applied force. Afterwards, the
average of the two cases was chosen for the axisymmetric calculation. Other parameters such as
the slab geometry and the loading conditions were not included in the derivation of the factor.
(a)
(b)
Figure 8.2: Element by which the principal forces act (a) parallel to the reinforcement or (b) at an angle
of 45 to the direction of the reinforcement
171
Chapter 8
If one defines the stiffness reduction factor as -value representing the ratio between the
stiffness of an element where the principal forces are acting parallel to the reinforcement and the
stiffness of an element where the principal forces are acting in an arbitrary direction, this factor
can be expressed as:
(8.1)
II
where EI1 is the secant flexural stiffness in the principal direction (tangential) and EIIIx is the
stiffness of a cracked section in parallel direction to the reinforcement obtained by using a
linear-elastic behavior of steel and concrete. It can be expressed as:
II
(8.2)
where is the flexural reinforcement ratio, Es is the Youngs modulus of the reinforcing steel, d
is the effective depth and xel is the height of the compression zone and can be estimated as:
(8.3)
where Ec is the Youngs modulus of concrete and the other parameters are defined as in
Equation 8.2.
The secant flexural stiffness EI1 was calculated with the NLFEA approach explained in Chapter
5. The value of the flexural stiffness in principal (tangential) direction enables the determination
of the -value at any slab location according to Equation 8.1. Figure 8.3 shows the distribution
of the -value for a square slab specimen with properties according to Table 8.1 for different
load levels. One can see that generally the stiffness in tangential direction is larger along the
axis (parallel to the reinforcement) than in diagonal direction. For the load level of 60% VR,flex,
the -value is along the axis approximately 1.0 whereas it is in diagonal direction approximately
0.5. For larger load levels the -value decreases below 1.0 along the axis due to the fact that the
flexural reinforcement starts yielding and thus the secant stiffness is lower than the calculated
stiffness based on a linear-elastic material behavior.
172
0.75
5
3
1.2
3
0.
5
3
3
0.75
5
1.2
1.2
1
1.25
0.
75
0.
1.25
0.75
1.2
5
5
1
0.7
25
1.
0.75
0.5
0.5
1.25
1
0.7
0.5 5
1.250.75 1
0.715 1.25
3
1.25
5
0.7
3
3
31
1.2
1 3
1
3
0.5
25
0.5
5
1.2.7.55
00
0.
0.
5
75
1.25
5
0.
1.25
0.25
1.25
0.5
5
0.7
0.5
1.25
1.25
0.
25
1.
0.5
75
1.250.
0.7
0.5
5
0.7
0.5
1.25
1.2
0.75
1.25
0.75
0.
0.5
1.2
15
3
3
1
1.25 0.75
0.5
0.5
0
0.7 .5
5
1
1.25
1.2
0.75
0.5
3
1
0.75 1.25
1.25
0.7
3
1
0.7
5
1.21
5
1.2
0.5
0.75
0.7
3
1
1.25
0.5 75
0.
0.5
1
1.25
1.25
0.75 1
3
10.75
0.75
0.
0.7
1.25
11.25
0.75
5
1.2
05
1.2
1.25
1.
20.5 55
0.7
00.7.55
3
1
3
3
Figure 8.3: Distribution of -value for load cases of (a) 60% VR,flex, (b) 75% VR,flex, and (c) 90% VR,flex
The distribution of the value depends mainly on the geometry of the slab and on the load
application. Consequently, each geometry or loading may lead to a different distribution of the
stiffness and thus may lead to a different slab response. Figure 8.4 shows the distribution of the
-value for a circular slab with load applied at the slab perimeter. In general, the behavior looks
similar to the square slab for which the -values are smaller in the diagonals. However, due to
the equal load around the perimeter, the first principal moments in the diagonal direction are
larger and thus the stiffness in the diagonals is smaller than for square slabs with only eight load
introduction points.
1.25
0.75
0.
1.2
0.7
1
1
1
1.25
11.25
0.7
0.5
0.75
0.5
1.2
5
0.
75
11
5
0.
0.55
0.
25
0.2
31.
25
75
0.
1.25
1.
0
0.5
5
0.7
0.
75
1
3
2
1.
5 .5
0. 0
0.2
0.5
1.25
11.25
75
75
0.
75
11.25 1
1
75
0.
0.
31
0.75
0.
25
1.
1.251
75
0.
0.5
1.25 1
1.25
3 1.25
0.25
1.25
0.7
0.7
0.7
0.75
0.75
0.7
1.25
5
0.7
0.
5
25 0.5
.
1
1
1.12
1.12
3
0.751
0.7
3
1
11.25 1 1
0.7
25
1.2
3 1.
11
1.25
1.
0.7
5
1.2
3
0.75
1.25 3
0.
11
3
25
1
5
1.2 3
Figure 8.4: Distribution of -value for load cases of (a) 60% VR,flex, (b) 75% VR,flex, and (c) 90% VR,flex
For a simplified calculation, only one average -value is desired. Therefore, several different
parameters and their influence on an average -value were investigated by comparing the
Quadrilinear model to a corresponding NLFEA. The following calculations are performed for
slabs corresponding to the specimens tested within this research (square slab with loading at
eight points at the perimeter). The investigated parameters were the flexural reinforcement ratio,
the column size, and the slab thickness. The comparison of the curve obtained by the NLFEA to
the analytical calculations using different -values leads to an optimal average -value for the
calculation of the tested slab specimens.
173
Chapter 8
Flexural reinforcement ratio

Figure 8.5 shows the normalized load-rotation curves for slabs with different reinforcement
ratios. It can be seen that the larger the reinforcement ratio is the lower the value is, which
corresponds to the observations of Guandalini (Guandalini 2005). However, the differences
between the NLFEA calculations and the Quadrilinear curves are relatively small. Therefore,
the influence of the flexural reinforcement ratio can be neglected. For all the flexural
reinforcement ratio, the best agreement with the NLFEA is obtained by using a value of
= 0.75.
(a)
(c)
(b)
10
=1.00%
=0.77%
=1.50%
9
8
V/mR [-]
NLFEA
NLFEA
=0.75
=0.70
=0.65
6
5
NLFEA
=0.75
=0.70
=0.65
=0.75
=0.70
=0.65
4
3
2
1
0
0
0.01
[]
0.02
0.03
0.01
[]
0.02
0.03
0.01
[]
0.02
0.03
Figure 8.5: Normalized load-rotation curve of the NLFEA and the Quadrilinear approach for
reinforcement ratios of: (a) 0.77%, (b) 1.00%, and (c) 1.50%
Column size
Figure 8.6 shows the normalized load-rotation curves for slabs with different column sizes. It
can be seen that the column size does not have a crucial influence on the factor since the
curves of all column sizes follow nearly the same pattern. This can be explained by the fact that
the cross-sectional behavior is the same and that the principal direction of the moments do not
change. Again, the best agreement with the results of the NLFEA is obtained by using a value of
= 0.75.
174
(a)
(c)
(b)
10
c=260 mm
c=130 mm
c=520 mm
9
NLFEA
V/mR [-]
7
6
NLFEA
NLFEA
=0.75
=0.70
=0.65
5
4
=0.75
=0.70
=0.65
=0.75
=0.70
=0.65
3
2
1
0
0
0.01
[]
0.02
0.03
0.01
[]
0.02
0.03
0.01
[]
0.02
0.03
Figure 8.6: Normalized load-rotation curve of the NLFEA and the Quadrilinear approach for column
sizes c of: (a) 130 mm, (b) 260 mm, and (c) 520 mm
Slab thickness
Figure 8.7 shows the normalized load-rotation curves for slabs with different slab thicknesses. It
can be seen that before yielding of the flexural reinforcement the thicker the slab is the larger
the factor is. For these cases a good agreement with the results of the NLFEA is again
obtained by using a value of = 0.75.
(a)
(c)
(b)
10
h=320 mm
h=250 mm
h=400 mm
9
8
V/mR [-]
NLFEA
NLFEA
=0.75
=0.70
=0.65
NLFEA
=0.75
=0.70
=0.65
=0.75
=0.70
=0.65
4
3
2
1
0
0
0.01
[]
0.02
0.03
0.01
[]
0.02
0.03
0.01
[]
0.02
0.03
Figure 8.7: Normalized load-rotation curve of the NLFEA and the Quadrilinear approach for slab
thicknesses h of: (a) 250 mm, (b) 320 mm, and (c) 400 mm
Based on this investigation, it can be noted that the influences of the investigated parameters on
the -value are rather small and thus they can be neglected. Thus, the proposed value is = 0.75
since it showed the best agreement with the NLFEA. However, it has to be noted that this value
was determined based on calculations of slabs with the geometry and the loading conditions
corresponding to the tested specimens. In the case of other slab geometries and loading
conditions, this value might be different.
175
Chapter 8
8.1.2. Transformation of the column shape

Generally, if symmetrical punching is investigated, two different column shapes were used:
circular and square. While for circular columns no further transformation is needed, the square
columns have to be transformed to circular columns in order to obtain an axisymmetric case.
Guandalini (Guandalini 2005) proposed the use of an equivalent perimeter so that the shear
stresses at the column face are the same for both column shapes. Therefore, the radius of the
equivalent column rc can be calculated by:
2
(8.4)
where c is the side length of a square column.

To verify this assumption, NLFEA calculations were performed, in which square and circular
column shapes were compared. The radius of the column was calculated according to
Equation 8.4 to obtain an equivalent perimeter. The investigated parameters were the
reinforcement ratio, the column size, and the slab thickness. For each parameter and column
shape, the load-rotation curve was calculated. Additionally, The NLFEA calculations were
compared to the results of the Quadrilinear model, for which a value of 0.75 was used.

Figure 8.8 shows the load-rotation curves for different flexural reinforcement ratios. It can be
seen that the difference between a calculation with a square column and a circular column are
small. Furthermore, it can be noted that the flexural reinforcement does not influence the
transformation of the column size.
(a)
(c)
(b)
10
=1.00%
=0.77%
=1.50%
9
8
V/mR [-]
=0.75
square column
=0.75
=0.75
square column
square column
6
5
circular column
circular column
circular column
3
2
1
0
0
0.01
[]
0.02
0.03
0.01
[]
0.02
0.03
0.01
[]
0.02
176
0.03
Column size
Figure 8.9 shows the load-rotation curve for different column sizes. It can be seen that the
hypothesis of an equivalent parameter works well for small column sizes. For larger column
sizes the difference are somewhat larger especially after yielding of the flexural reinforcement.
However, the differences stay relatively small and are thus in an acceptable range.
(a)
(c)
(b)
10
9
c=260 mm / rc= 166 mm
c=130 mm / rc= 88 mm
c=520 mm / rc= 331 mm

=0.75
8
7
square column
=0.75
square column
=0.75
square column
V/mR [-]
6
circular column
5
circular column
circular column
3
2
1
0
0
0.01
[]
0.02
0.03
0.01
[]
0.02
0.03
0.01
[]
0.02
0.03
Slab thickness
Figure 8.10 shows the load-rotation curves for different slab thicknesses. It can be seen that
again after yielding small differences occurs in the case of thicker slabs. This effect is however
not due to the increased thickness but due to the increased column size since the column size to
effective depth ratio was kept constant. Therefore, the thicker slabs have larger column sizes
and thus the same effect occurs as described in the previous paragraph.
(a)
(c)
(b)
10
h=320 mm
h=250 mm
h=400 mm
9
8
V/mR [-]
=0.75
square column
=0.75
square column
=0.75
square column
6
circular column
circular column
circular column
4
3
2
1
0
0
0.01
[]
0.02
0.03
0.01
[]
0.02
0.03
0.01
[]
0.02
0.03
177
Chapter 8
Generally, the hypothesis of an equivalent perimeter works well in all the investigated cases.
For large columns however, small discrepancies can occur mainly after yielding of the flexural
reinforcement. Yet, these differences are for the investigated column sizes in an acceptable
range. Thus, it can be concluded that the approach of an equal perimeter is suitable for the
analysis.
8.1.3. Transformation of slab shape and loading conditions

The transformation of the slab shape from a square slab to a circular slab can be performed by
the assumptions that the maximum flexural strength (Vflex) should be the same for both slab
shapes. In the case of a square slab with loading at eight points, the maximum flexural strength
can be calculated by assuming a mechanism as shown in Figure 8.11a leading to the following
expression (Guidotti 2010):
8
(8.5)
where mpl is the flexural moment capacity, l is the side length of the slab, c is the side length of
the column, b is the distance between the load introduction points, and b1 is the distance
between the load introduction point and the slab edge (see Figure 8.11).
For axisymmetric slabs, the flexural strength can be calculated by assuming a mechanism as
shown in Figure 8.11b leading to the expression:
2
(8.6)
where mpl is the flexural moment capacity, rs is the radius of the slab, rq is the radius of the load
introduction points, rc is the radius of the column.
178

(a)
(b)
Figure 8.11: Assumed flexural failure mechanism for (a) a square slab and (b) a circular slab
The hypothesis that the flexural strength should be the same for both geometries leads to the
size of the corresponding circular slab. Therefore, the radius of the circular slab rs can be
calculated by:
8
4
2
(8.7)
where rq is the radius of the load introduction point that can be determined by
2
(8.8)
In order to verify this hypothesis, calculations with square and circular slabs were performed
whereby the following parameters were investigated: the flexural reinforcement ratio, the
column size, and the slab thickness.

Figure 8.12 shows the normalized load-rotation curves for different flexural reinforcement
ratios. It can be seen that generally the response of the square slab and the circular slab
correspond well before yielding of the flexural reinforcement occurs. After yielding the square
slab behaves stiffer than the axisymmetric slab. Several different influences contribute to the
different slab behavior. As seen before, the transformation of the slab geometry contributes
somewhat to the softer behavior of the slab. Additionally, the load distribution is different. In
the case of the axisymmetric slab the portion of torsional moment is higher than in the case of
the square slab as already seen by comparing Figure 8.4 to Figure 8.3. This leads to a softer
behavior due to the fact that the torsional stiffness is smaller than the flexural stiffness.
However, this change in stiffness is already accounted for by factor . With respect to the
179
Chapter 8
flexural reinforcement ratio it can be noted that the difference stays always the same. Thus, the
transformation of the slab according to the method presented does not depend on the flexural
reinforcement ratio.
(a)
(c)
(b)
10
=1.00%
=0.77%
=1.50%
9
8
7
=0.75
square slab
=0.75
=0.75
square slab
square slab
V/mR [-]
6
5
axisymmetric slab
axisymmetric slab
axisymmetric slab
3
2
1
0
0
0.01
[]
0.02
0.03
0.01
[]
0.02
0.03
0.01
[]
0.02
0.03
Column size
Figure 8.13 shows the load-rotation curve for specimens with different column sizes. The
change in the column size influences the difference between a square and a circular slab. The
larger the column size is, the larger the difference in behavior between a square and a circular
slab is. However, this influence is rather small. The main difference between a square and a
circular slab is in all three cases due to the influence of the torsional stiffness as explained
previously.
(a)
(c)
(b)
10
9
c=260 mm / rc= 166 mm
c=130 mm / rc= 88 mm
c=520 mm / rc= 331 mm

=0.75
V/mR [-]
7
6
=0.75
=0.75
square slab
square slab
square slab
5
4
axisymmetric slab
axisymmetric slab
axisymmetric slab
3
2
1
0
0
0.01
[]
0.02
0.03
0.01
[]
0.02
0.03
0.01
[]
0.02
180
0.03
Slab thickness
Figure 8.14 shows the load-rotation curve for specimens with different thicknesses. It can be
noted that the slab thickness does not influence the difference in the slab behavior between a
square and a circular slab. The main difference is again due to the different torsional behavior as
explained previously. This difference is approximately equal for all slab thicknesses. However,
it seems that the difference is slightly larger for thicker slabs. This is however not due to the
larger slab thickness but to the larger column size since the ratio of the column size to the
effective depth (c/d) was kept constant for these specimens. Therefore the thicker the slab is, the
larger the column is and the larger the differences are (Figure 8.13).
(a)
(c)
(b)
10
h=320 mm
h=250 mm
h=400 mm
9
square slab
V/mR [-]
=0.75
=0.75
square slab
=0.75
square slab
6
5
axisymmetric slab
axisymmetric slab
axisymmetric slab
4
3
2
1
0
0
0.01
[]
0.02
0.03
0.01
[]
0.02
0.03
0.01
[]
0.02
0.03
8.1.4. Load application

In general, no adjustments are needed regarding the load application since the calculation of the
slab dimensions already accounts for the location of the load application. However, another
aspect of the load application is to distinguish between equal force applied and equal
displacement applied calculations. In the first case, the loading of the specimen will be
introduced by applying an equal force, whereas by the latter, loading is provided by the
introduction of an equal displacement at the load introduction points. The tests within this
research were performed by introducing the same force at each load introduction point.
However, as described in Chapter 5, the NLFEA uses an approach in which equal displacements
are applied at each load introduction point. This leads to the advantage that the calculations are
more robust, especially in the case of yielding of the flexural reinforcement. The drawback of
this method is that the behavior of the slab is not modeled accordingly to the experimental test
procedure, which could result in wrong predictions. Nevertheless, as Figure 8.15 shows, the
difference between the two calculation methods is negligible for the geometry used in the
experimental campaign.
181
Chapter 8
(a)
(b)
10
=1.50%
=1.00%
9
8
V/mR [-]
equal force applied
=0.75
equal force applied
=0.75
6
5
4
equal displacement
applied
equal displacement
applied
3
2
1
0
0
0.01
[]
0.02
0.03
0.01
[]
0.02
0.03
Figure 8.15: Difference between equal force and equal displacement for square slab specimens
The reason for the rather small differences can be further investigated by analyzing the global
slab behavior. Figure 8.16 illustrates the behavior of a square slab for an equal displacement
applied calculation and the equal force applied calculation with respect to the curvature in the
first principal (tangential) direction, the moment in the first principal (tangential) direction, and
the shear fields. It can be seen that no significant difference occurs between the displacement
and the equal force applied calculation as it was shown by the calculated load-rotation curve
presented in Figure 8.15. The curvatures are slightly smaller along the diagonals than along the
direction of the load application points. This leads together with the smaller stiffness along the
diagonals to larger moment along the axis and smaller moments along diagonals. This moment
behavior is approximately the same for the calculation with equally applied displacement and
with equally applied force. Same conclusions can be drawn from the stress fields that are similar
for both cases. Therefore, in the case of square slabs loaded as the test specimen, nearly no
difference in behavior occurs between an equal displacement and an equal load approach. This
consideration is also important for the design of punching test set-ups. According to the
calculation, no difference occurs if the force is applied at eight points and supported at the
center, as it was performed within this research, or if the force is applied at the center and the
displacements are kept the same at eight points at the perimeter, which would correspond to an
equal displacement approach. Thus, tests performed with either one of these test set-ups are
comparable. However, it has to be noted that this is only the case for square slabs with the
previously described load application.
182
(a)
(b)
0.01
0.0
0.0
0.0
0.01
0.01
0.02
0.02
2
0.03
.0
.03
0.02
0.03
0.0
1
0.0
9
.03
.05
.01
0.0
6 0.1
0.
0.04
08 .07 .04
0 0
.02
0.0
.02
.02
1
0.0
2
.0
0
4
.0
0
.05
0
.01
.05
0
6
0.0
0.0
.0
0.04
0.02
0.03
7
5 .0 8
.04 0.0 00.0 0.
0
2
0.0
.03
0.0
3
0.0 .04.05
0 0.06 0.0
7
0
0.06
0.0
0.02
0.01
0.01
0.0
0.0
0.0
0.01
(d)
100
80
120 40
1 60
1 0
8
1
0
100
200
120 140
120
1
1860
0
40
120
80
100
0
8
10
20
1
12
40
1
120
200
40
1
120
0
0
0
2
180
160
180 60
1
100
40
1
20
200
220
14
100
80
(e)
220
200
20
200
220
220
10
200
180 0
16
14
1 100
20
100
60
80
140
20
18060
1
220
200
180
1
60
60
1 180
200
220
14
20
180
10
200
20
200
20
16
220
40 160
1 180
200
12
40
80
100
14
120
1 180
60
220
12
0
14 160
180
200
8
0
100
80
100
120
10
100
(c)
100
80
(f)
Figure 8.16: Global slab behavior of a square slab with respect to the curvature in the first principal
direction 1 (a,b), the moment in the first principal direction m1 (c,d), and the shear fields
(e,f) for an equal displacement applied (a,c,e) and an equal force applied calculation (b,d,f)
( = 1.00%; V / mR = 6.0)
183
Chapter 8
In the case of other slab geometries or loading conditions, for example for axisymmetric cases,
the difference between an equal force applied or an equal displacement applied calculation may
become important. Figure 8.17 presents the load-rotation curve for a calculation of a circular
slab with equally applied displacement and with equally applied force. Whereas the difference
between the equal displacement and equal force curve is small for square slabs, the difference
for circular slabs is noticeable.
(a)
(b)
10
=1.50%
=1.00%
9
8
V/mR [-]
7
6
=0.65
equal displacement
applied
=0.65
equal displacement
applied
5
4
equal force applied
equal force applied
3
2
1
0
0
0.01
[]
0.02
0.03
0.01
[]
0.02
0.03
Figure 8.17: Difference between equal force and equal displacement for circular slab specimens
The difference is also clearly visible in the analysis of the slab behavior. Figure 8.18 illustrates
the curvature in the first principle direction (tangential), the moment in the first principle
direction (tangential), and the shear fields for an equal displacement applied and for an equal
force applied calculation. While the behavior of the slab close to the column is similar in both
cases, the behavior at the outer part of the slab is clearly different. In the case of the equal
displacement applied calculation, the load distributes itself according to the stiffness of the slab.
Therefore, the moments are larger in the more rigid direction, thus in the direction of the
reinforcement. The curvatures are somewhat larger along the axis than along the diagonals. In
the case of the equal force applied calculation, the force is distributed equally along the slab
perimeter. Therefore, the moments are also equally distributed. In order to obtain these
moments, the curvatures have to be larger along the diagonals than along the axis due to the
smaller stiffness in these areas. This increase in curvature along the diagonals leads to a softer
slab behavior and thus to a less stiff load-rotation response.
184
(a)
(b)
0
0
0.01
0.02
0.0
01
0
.02
0
.0
2
0
0
.
.0 2
0
0
.0
0 5
.0
7
0.03
0.06
.0
4
0
0.03
0.0
.01
.01
0
.0
.0
.02
0.0
0.0
0
0
(d)
200
140
0
16
2
20
200
80
6
800
00
60
140
40
1
60
00
160
0
14
180
8100
0
4600
100
1 160
80
140
200
80
0
12
60
140 1
120
200
60
80
220
00
00
220
2
80
1 160
40
60
200
16040
1
180
200
140
120
1
00
40
160
220
100
18060
1
220
16
200
20
180
60
20
220
20
00
40
200
220
0
14 60
1
1
1
18600
200
200
200
00
4
6
80
100
120
80
180
120
200
180
0
160 180
40
120
140
16
20
200
16
14
80
40
(c)
(e)
3
0.0
0.0
3
.01
0.04
0.01
0.05
.01
.04
0.0
0.02
0.02
.09
0 07
0.
0.08
0.0
0.02
0.01
0.05
0.04
.02
0
4
0.0
.03
0.0
7
.06 0.0
.0
9 80
0.0
.0
0.04 5 .008.09
00.0.06 0 0.1
0.0
.05
0
0.03
0.01
0.03
0.02
8
.0
.060 90
0 0.0 0.1
.11
0.06
1
0.0
4
0.0
0.03
0.04
0.02
0.0
0
.0
7
2
.0
0.0
0.0
0
0.02
0.0
0 0.01
.0
0 0
1600
4
1
16
0
12 140
0
(f)
Figure 8.18: Global slab behavior of a circular slab with respect to the curvature in the first principal
direction 1 (a,b), the moment in the first principal direction m1 (c,d), and the shear fields
(e,f) for an equal displacement applied (a,c,e) and an equal force applied calculation (b,d,f)
( = 1.00%; V / mR = 6.0)
185
Chapter 8
8.2. Other failure modes considered

The analytical model presented in the previous chapter covers failure within the shearreinforced area and failure of the concrete strut close to the column. However, for the validation
of the proposed model to tests from literature, other failure modes have to be considered as well
such as punching of slabs without shear reinforcement and punching outside the shearreinforced area. Therefore, simplified formulation for cases of punching of slabs without shear
reinforcement and for punching outside the shear-reinforced area are presented.
8.2.1. Punching of slabs without shear reinforcement

The simplified formulation for cases of punching of slabs without shear reinforcement is based
on the CSCT and the Quadrilinear approach proposed by Muttoni (Muttoni 2008). For the loadrotation response of the slab, the only parameter adjusted compared to the proposed model, is
the distribution of the rotation and thus the calculation of the curvature at the location of the
shear crack r2. Similar to the Quadrilinear model, it is assumed that the rotation is constant
outside the radius r2 and that the rotation increases linearly within radius r2 leading to a constant
radial and tangential curvature. Thus, the curvature at location r2 can be calculated by
(8.9)
where is the slab rotation at the outer slab segment and r2 is assumed to be at a distance d
(effective depth) from the column face.
The failure criterion for punching of slabs without shear reinforcement corresponds to the
formulation of the CSCT (Muttoni 2008; Fernndez Ruiz and Muttoni 2009) that is also used
for the concrete contribution in the proposed model. Thus, the punching strength as a function
of the rotation is given by:
3
4
1
15
(8.10)
186
8.2.2. Punching outside the shear-reinforced area

The investigation of the failure outside the shear-reinforced area is out of scope of this research
project. However, in order to compare tests from literature, this failure mode has to be addressed
as well. Therefore, a simplified approach is proposed for the use in the comparison. For the
prediction of the load-rotation response, the same formulation was applied as for the other
failure modes (punching within the shear-reinforced area and failure of the compression strut).
Nevertheless, it has to be noted that a shear crack at the outer perimeter may have an influence
in the slab response. For the failure criterion, the approach of the CSCT (Fernndez Ruiz and
Muttoni 2009) was used. This formulation correspond to the failure criterion for punching of
slabs without shear reinforcement but with an control perimeter set at a distance of 0.5d from
the outermost perimeter of the shear reinforcement. Thus, the punching strength outside the
shear reinforcement as a function of the rotation is given by:
3
15
(8.11)
where is the slab rotation, d is the effective depth, dv,ext is the distance between the flexural
reinforcement and the bottom end of the vertical branch of the shear reinforcement, dg0 is a
reference aggregate size equal to 16 mm, dg is the aggregate size, fc is the compressive strength
of concrete, and b0,out is the control perimeter set at a distance of 0.5dv from the outermost
perimeter of the shear reinforcement.
8.3. Validation with tests within this research

The proposed model was investigated with respect to the predicted punching strength and the
influence of several different parameters namely the slab thickness, the column size, and the
shear reinforcement ratio. Thus, the punching strength was calculated according to the proposed
model as a function of the investigated parameter. These calculations were performed using the
following assumptions: the concrete compression strength was chosen as 33.5 MPa (average of
the tests), the yielding strength of the flexural reinforcement as 575 MPa (average of the tests),
the yielding strength of the shear reinforcement as 550 MPa (average of the tests), the distance
between the column face and the first row of shear reinforcement s0 as 0.375d, the distance
between the shear reinforcement s1 as 0.75d, the shear reinforcement ratio as 0.9% if not varied,
the effective depth as 210 mm if not varied, and the ratio of the column size to the effective
depth as 1.24 if not varied. Additionally, the predicted punching strength was compared with the
experimentally obtained strength of punching tests from literature in order to investigate the
general performance of the proposed model.
187
Chapter 8
8.3.1. Strength and rotation predictions

In comparison to the results from the tests performed within this research, it can be noted that
the strength predicted by the analytical model is in good agreement with the experimentally
obtained punching strength (Figure 8.19a). The average is close to 1.0 (Avg.: 1.01) and the
coefficient of variation is rather small (COV: 5.4%). Furthermore, it can be noted that the
predictions of the rotations at failure are in good agreement with the measured rotations at
failure (Figure 8.19b). The predictions are generally lower than the experimentally obtained
rotation at failure leading to an average of 1.17. The scatter is rather low leading to a small
coefficient of variation of only 11.3%. This excellent agreement can also be seen subsequently
in the comparison of the predicted load-rotation response to the measured one.
(a)
(b)
2
3
Avg. = 1.01
COV = 5.4%
1.8
1.4
R,test / R,model []
VR,test / VR,model []
1.6
1.2
1
0.8
0.6
0.4
0.2
0
0
0.5
1
w [%]
h=250 mm
h=320 mm
h=400 mm
1.5
2
Avg. = 1.17
COV = 11.3%
2.5
2
1.5
1
0.5
0
0
0.5
1
w [%]
h=250 mm
h=320 mm
h=400 mm
1.5
2
Figure 8.19: Ratio of (a) the experimentally obtained punching strength to the punching strength
predicted by the analytical model and (b) the experimentally obtained rotation at failure to
the rotation at failure predicted by the analytical model

depth d. While the prediction for specimens without shear reinforcement and with studs shows
good agreement with the test results, the prediction for specimens with stirrups differ somewhat
from the test results. This can be explained by the difference between the properties of the test
specimen and the properties used for the calculation. For example, the shear reinforcement ratio
was larger for the calculation then the one used in the test specimens. Additionally, the spacing
between the column face and the first row of vertical branch of the stirrups were not constant by
the test specimens. This led to the prediction of a different failure mode. A detailed analysis of
the specimen predicts a failure within the shear-reinforced area (Figure 8.23) whereas by the
calculation using a standardized specimen leads to a failure of the concrete strut. Nevertheless,
the overall behavior is well predicted by the proposed model for cases with studs and with
stirrups. While the model predicts a small decrease in normalized strength for specimen without
shear reinforcement, the model predicts a slight increase in the normalized strength with
188
increasing thickness. This increase in normalized strength reduces with increasing thickness
leading to a nearly constant normalized strength for specimens with studs and an effective depth
larger than 300 mm.
Figure 8.21a shows the prediction of the normalized rotation at failure as a function of the
effective depth d. For specimens without shear reinforcement, the prediction shows good
agreement with the test results. Whereas for slabs without shear reinforcement the product of
the rotation at failure and the effective depth is nearly constant, for slabs with shear
reinforcement, the product of the rotation at failure and the effective depth is predicted to
increase nearly linearly. Thus, the model predicts that the thicker the slab is the larger the
deformations at the top surface are. This can be explained by the fact that the slenderness
decreases with increasing thickness since the load application radius stays constant. This
decrease of slenderness leads to the prediction of larger normalized strength for larger
thicknesses.
(b)
(a)
1.5
Model
Test (studs)
Test (stirrups)
Test (none)
1.25
MPa
(c)
studs
stirrups
studs
stirrups
0.75
studs
stirrups
0.5
none
none
0.25
0
0
100
200
300
d [mm]
400
500
0.5
1.5 2 2.5
c/d [-]
3.5
0
4 0
0.25
0.5
0.75
w [%]
1.25
Figure 8.20: Normalized punching strength predicted by the proposed model as a function of (a) the
effective depth, (b) the column size to effective depth ratio, and the shear reinforcement
ratio
8.3.3. Column size

column size to the effective depth c/d. For specimens without shear reinforcement, the predicted
strength is in good agreement with the experimental results. The normalized strength decreases
with increasing c/d ratio. Similarly, the predicted strength for specimens with studs is in good
agreement with the test results. The decrease in normalized strength with increasing c/d ratio is
slightly more dominant than for specimens without shear reinforcement. At a ratio of around 2.5
the flexural capacity is reached and thus the difference between studs and stirrups disappears.
Again, the comparison of the predicted normalized strength to the test results of the specimen
with stirrups has to be regarded with caution. Due to the difference in the amount and the
189
Chapter 8
arrangement of the calculated and tested specimen, the prediction based on the standard
specimen may lead to a different failure mode than a detailed analysis (compare to Figure
8.23h-j). This explains the difference for specimens with small column sizes by which more
accurate results can be obtained if the actual amount and arrangement of shear reinforcement are
used in the calculation (Figure 8.23h).
Figure 8.21b shows the prediction of the normalized rotation at failure as a function of the ratio
of the column size to the effective depth c/d. For specimens without shear reinforcement, for
specimens with stirrups and for specimens with studs, the predictions are in a good agreement
with the test results. While the increase in rotation for larger column sizes is rather small for
specimens without shear reinforcement, which can be explained by the rather large stiffness of
the investigated slab specimens, the increase in rotation is significant for specimens with shear
reinforcement. Especially, in the case of studs at a c/d ratio of around 2.0, the rotations increase
rapidly. Thus, it can be noted that for larger c/d ratios a flexural failure may occur before a
punching failure occurs as it was expected to be in the case of specimen PL8 (the test was
stopped before failure, thus it is not shown in Figure 8.21b).
(b)
(a)
15
Model
Test (studs)
Test (stirrups)
Test (none)
12.5
d [mm]
(c)
leading to
lexural failure
10
studs
stirrups
studs
7.5
studs
stirrups
stirrups
2.5
none
none
0
0
100
200
300
d [mm]
400
500
0.5
1.5 2 2.5
c/d [-]
3.5
4 0
0.25
0.5
0.75
w [%]
1.25
Figure 8.21: Normalized rotation at failure predicted by the proposed model as a function of (a) the
effective depth, (b) the column size to effective depth ratio, and the shear reinforcement
ratio

reinforcement ratio. Generally, it can be noted that the predicted strength follows the expected
trend and the experimental results. The curve can generally be divided into four different parts.
In a first part the strength of a calculation for slabs without shear reinforcement was taken since
it provides slightly higher loads due to the change in the assumed slab response. In a second
phase, the shear reinforcement is starting to contribute to the punching strength but it is not fully
activated yet. In other words, the shear reinforcement is not yielding at failure. In a third phase,
190
the shear reinforcement is fully activated, thus it is yielding at failure. Finally, the maximum
punching strength is reached at which a failure of the concrete strut is predicted.
Figure 8.21c shows the prediction of the normalized rotation at failure as a function of the shear
reinforcement ratio. Since the amount of shear reinforcement affects, according to the proposed
model, mainly the failure criterion and only slightly the load-rotation of the slab, the predicted
rotation at failure follows the same trend as the predicted punching strength. This trend is also
supported by the experimentally obtained results.
8.3.5. Prediction of the slab response and the failure criteria

The calculation of the punching strength and the rotation at failure led to good agreement with
the experimentally obtained results as shown in Figure 8.19. This good performance can also be
seen in the comparison of the measured and the predicted load-rotation curves. Figure 8.22
shows the measured and predicted load-rotation response for specimens without shear
reinforcement. In addition, Figure 8.22 shows the load-rotation response predicted by the
Quadrilinear approach. Both approaches propose nearly the same response since the simplified
calculation for slabs without shear reinforcement uses the same curvature at radius r2 as the
Quadrilinear approach. Therefore, the proposed model predicts nearly the same load-rotation
response as the Quadrilinear model showing good agreement with the measured response.
Additionally, since the failure criterion of the CSCT was adopted, the predicted punching
strength corresponds to the predictions obtained from the CSCT leading to a good agreement
with the test results.
191
Chapter 8
(a)
(b)
1
PL1
0.9
MPa
(c)
PV1
0.8
proposed model
0.7
Quadrilinear
proposed model
0.6
0.5
PL3
Quadrilinear
Quadrilinear
proposed model
measured
measured
measured
0.4
0.3
Vc
Vc
0.2
Vc
0.1
0
(d)
(e)
PL4
0.05
0.1
0.15
0.2
PL5
MPa
0.9
0.8
Quadrilinear
0.7
Quadrilinear
proposed model
proposed model
0.6
0.5
measured
measured
0.4
0.3
Vc
0.2
Vc
0.1
0
0
0.05
0.1
0.15
.2 0
0.05
0.1
0.15
0.2
Figure 8.22: Measured and predicted load-rotation curves and failure criterion for specimens without
shear reinforcement
Figure 8.23 shows the load-rotation curves for specimens with shear reinforcement. It shows the
measured and predicted load-rotation response from the proposed model as well as from the
Quadrilinear approach. As previously discussed in Chapter 4, the Quadrilinear approach
estimates a stiffer response compared to the experimentally obtained one. In contrast, the
proposed model accounts for the less stiff response leading to a good agreement with the
experimental measured load-rotation curves.
Additionally, Figure 8.23 shows the failure criteria investigated within this research project:
failure within the shear-reinforced area (VR,I) and failure of the concrete strut (VR,II). Generally, it
can be noted that the proposed model predicts for specimens with studs (PL6, PL7, PL9, PL10)
failure of the compression strut. On the other hand, for specimens with stirrups (PF2, PF4, PF5),
the proposed model predicts failure within the shear-reinforced area but without full activation
of all the shear reinforcement that crosses the outer shear crack. Three special cases exist: the
specimen with a small column size and stirrups (PF1), the specimens with large column sizes
(PL8, PF3) and the specimens with less amounts of shear reinforcement (PL11, PL12).
Due to the small column size of specimen PF1, the model predicts a larger compression zone of
the outer slab element (x2) compared to other specimens with stirrups such as specimen PF2.
The rather large height of the compression zone reduces the maximum width of the compression
192
strut in the wedge element used in the calculation. Thus, although less shear reinforcement
crosses the shear crack according to the model in the case of specimen PF1 than in the case of
specimen PF2, the model predicts nevertheless a failure of the compression strut instead of a
failure within the shear-reinforced area as it is predicted for specimen PF2.
For specimens with large column sizes, it can be noted that while the model predicts a failure of
the concrete strut at a normalized rotation of 0.27 for specimen PF3, which corresponds well to
the normalized measured rotation at failure (R,test d / (dg0 + dg) = 0.31), the model predicts a
failure of the concrete strut at a normalized rotation of 0.39 for specimen PL8 (the test was
stopped at normalized rotation of d / (dg0 + dg) = 0.31). Therefore, the proposed model
predicts not only the punching failure well but also the absence of a punching failure in the case
of specimen PL8.
For lower amount of shear reinforcement, the model predicts for both specimens (PL11 and
PL12) a failure within the shear-reinforced area. Whereas the model predicts for specimen PL11
a failure without full activation of the shear reinforcement, it predicts for specimen PL12 a
failure close to the point at which full activation of the shear reinforcement occurs.
193
Chapter 8
(a)
(b)
MPa
1.5
(c)
PL6
Quadrilinear
PL7
VR,II
proposed model
1
measured
PL8
VR,II
Quadrilinear
VR,I
VR,I
VR,I
measured
VR,II
Quadrilinear
measured
proposed model
0.5
proposed model
(d)
(e)
1.5
(f)
PL9
VR,II
PL10
Quadrilinear
VR,II
MPa
Quadrilinear
PL11
VR,II
Quadrilinear
1
measured
VR,I
VR,I
measured
VR,I
measured
proposed model
0.5
proposed model
proposed model
(g)
(h)
1.5
(i)
PL12
MPa
PF2
PF1
VR,II
Quadrilinear
Quadrilinear
Quadrilinear
proposed model
1
VR,I
measured
VR,I
VR,I
measured
measured
VR,II
0.5
VR,II
proposed model
proposed model
(j)
(k)
1.5
VR,II
MPa
(l)
PF3
PF4
PF5
Quadrilinear
Quadrilinear
VR,I
1
Quadrilinear
proposed model
measured
VR,I
VR,I
measured
0.5
measured
VR,II
VR,II
proposed model
proposed model
0
0
0.1
0.2
0.3 0
0.1
0.2
0.3 0
0.1
0.2
0.3
Figure 8.23: Measured and predicted load-rotation curves and failure criteria for specimens with shear
reinforcement
194
8.4. Comparison to tests from literature

8.4.1. General
In addition to the tests within this research project, the proposed model was compared to a
selection of tests from literature which was already used for the validation of the code provision
and the CSCT in Chapter 4. The selection consists of 132 test specimens whereby 28 specimens
without shear reinforcement and 104 specimens with shear reinforcement were used. More
detailed information about the selected test specimens can be found in Appendix A. Since the
proposed model is to a certain extent based on the CSCT, the results obtained by the proposed
model are directly compared to the results obtained by the CSCT.
Table 8.2 presents the average and the coefficient of variation (COV) for different selections of
test specimens. It can be seen that the proposed model shows generally a good agreement with
the test results. If all specimens are considered, the average is close to 1.0 and the coefficient of
variation is rather low. By the separation of specimens with and without shear reinforcement, it
can be noted that the average and the coefficient of variation for specimens without shear
reinforcement is the same as for CSCT. This can be explained by the fact that the calculation
uses the same failure criterion and similar assumptions for the calculation of the load-rotation
response. For specimens with shear reinforcement, it can be seen that despite the different
formulation of the load-rotation response and failure criteria compared to the CSCT, the average
and the coefficient of variation are nearly the same. In fact, the proposed model achieves also an
average that is close to 1.0 and a low coefficient of variation of 10.9% (CSCT: 11.3%). With
respect to the predicted failure modes, Table 8.2 shows that for the investigated failure modes
(failure within the shear-reinforced area and failure of the concrete strut), the predictions are in
good agreement with the test results. The average is for both failure modes close to 1.00.
Moreover, the coefficient of variation is only 8.4% and 10.4% (CSCT: 7.8% and 12.1%) for
failure of the concrete strut and failure within the shear-reinforced area, respectively. For failure
outside the shear-reinforced area, the coefficient of variation is in the same range. However, this
failure mode was not investigated and thus the calculation was performed with a simplified
approach. Consequently, a more detailed investigation of the load-rotation response and failure
criterion for failure outside the shear-reinforced area is desired.
195
Chapter 8
Table 8.2:
Statistical analysis of specimens from literature according to the proposed model
Selection
All specimens
Specimens without shear reinforcement
Specimens with shear reinforcement
Specimens with predicted failure:
crushing of concrete strut
within shear-reinforced area
# specimens
Average
COV
# specimens
Average
COV
# specimens
Average
COV
# specimens
Average
COV
# specimens
Average
COV
# specimens
Average
COV
proposed
model
132
1.02
10.9%
28
0.95
8.4%
104
1.04
10.8%
14
0.99
8.4%
47
1.04
10.4%
38
1.05
11.7%
CSCT
132
0.98
10.8%
28
0.95
8.4%
104
0.99
11.3%
22
0.91
7.8%
46
1.03
12.1%
36
1.00
8.8%
In order to investigate the influence of certain parameters on the agreement of the predictions
and the test results, the ratio of the experimentally obtained punching strength and the strength
predicted by the proposed model (figures in the center) is shown as a function of several
parameters namely the shear reinforcement ratio, the effective depth, the column size, the
flexural reinforcement ratio, the concrete compressive strength, and the maximum aggregate
size. Again, the results of the proposed model are directly compared to the CSCT. For this, the
ratio of the experimentally obtained punching strength and the strength predicted by the CSCT
is shown as well (figures on the left). In addition, the ratio of the strength predicted by the
CSCT to the strength predicted by the proposed model is shown (figures on the right) to analyze
the difference of the two models as a function of the previously mentioned parameters.

Figure 8.24 shows the ratio of the experimentally obtained punching strength and the predicted
strength as a function of the shear reinforcement ratio. Basically, the shear reinforcement ratio
defines the failure mode and the punching strength for failure within the shear-reinforcement
ratio. With respect to the load-rotation response, it can be noted that the load-rotation response
calculated with the CSCT is independent of the amount of shear reinforcement. In contrast, the
load-rotation response calculated with the proposed model depends somewhat on the amount of
shear reinforcement as the amount of shear reinforcement is considered in the calculation of the
load-distribution factor . However, since the load-distribution factor has only marginal
196
influence on the load-rotation response, the influence of the amount of shear reinforcement on
the response is rather limited.
With respect to the comparison of the test results with the proposed model, it can be noted that
specimens without shear reinforcement lead to lower strength ratios than specimens with shear
reinforcement. Additionally, it can be seen that the proposed model predicts failure of the
concrete strut only for moderate to large shear reinforcement ratios (w > 0.5%). For lower shear
reinforcement ratios, the predicted failure mode is either failure within or failure outside the
shear-reinforced area. In comparison to the CSCT, it can be noted that although the approaches
are different, the results are rather close. No tendency occurs if the two models are directly
compared (Figure 8.24c).
(a)
(b)
3
VR,test/VR,CSCT
no shear reinf.
crushing
within
outside
lexural failure
2.5
VR / VR,pred []
(c)
VR,CSCT/VR,model
VR,test/VR,model
2
1.5
1
0.5
0
0
0.5
1
w [%]
1.5
0.5
1
w [%]
1.5
0.5
1
w [%]
1.5
Figure 8.24: Punching strength ratio (a) VR,test/VR,CSCT, (b) VR,test/VR,model, and (c) VR,CSCT/VR,model as a
function of the shear reinforcement ratio
8.4.3. Effective depth

strength as a function of the effective depth. The effective depth is the main parameter defining
the slab stiffness and thus the main parameter for the calculation of the load-rotation response.
Additionally, the effective depth influences as well the failure criteria in both models.
Regarding the concrete contribution, the effective depth defines the shear resisting area (db0)
and the opening of the shear crack (wd). The crack opening is also the main parameter for
the failure criterion for failure within the shear-reinforced area in both approaches (the CSCT
and the proposed model). For the failure criterion of crushing of the concrete strut, the effective
depth is related to the strains transverse to the concrete strut and thus influences the strength of
the compression strut. With respect to the comparison of the test results with the proposed
model, it can be noted that no clear tendency exists with respect to the effective depth (Figure
8.25b). Additionally, it can be noted that the prediction of the failure mode does not depend on
the effective depth although the proposed model accounts only little for size effects. Moreover,
197
Chapter 8
the comparison to the CSCT reveals that the two models show nearly the same results and the
same tendency (Figure 8.25c).
(a)
(b)
3
VR,test/VR,CSCT
2.5
VR / VR,pred []
(c)
no shear reinf.
crushing
within
outside
lexural failure
VR,CSCT/VR,model
VR,test/VR,model
2
1.5
1
0.5
0
0
100
200
d [mm]
300
400
100
200
d [mm]
300
400
100
200
d [mm]
300
function of the effective depth
8.4.4. Column size

strength as a function of the column size. The column size influences the load-rotation response
by defining the distribution of the rotation and by the flexural capacity. Additionally, in the case
of the proposed model the column perimeter defines directly the height of the compression
zone. Generally, it can be noted that smaller column sizes lead to larger compression zones and
vice versa. Since the proposed model defines the location of the outer shear crack depending on
the height of the compression zone, it has an influence on the width of the compression strut and
to certain extent on the amount of shear reinforcement crossing the shear crack. Thus, both
failure criteria depend on the height of the compression zone and thus on the column size. In
contrast, the failure criteria of the CSCT account for the column size generally in the expression
of the control perimeter b0. Regarding the comparison to the test results, it can be noted that the
predictions of the proposed model show no clear tendency with respect to the column size to
effective depth ratio (Figure 8.26b). The comparison with the CSCT shows that the proposed
model lead generally to more conservative predictions, which improves somewhat the results of
small column to effective depth ratios (Figure 8.26c).
198
(a)
(b)
3
VR,test/VR,CSCT
2.5
VR / VR,pred []
(c)
no shear reinf.
crushing
within
outside
lexural failure
VR,CSCT/VR,model
VR,test/VR,model
2
1.5
1
0.5
0
0
c/d [-]
c/d [-]
c/d [-]
function of the column size to the effective depth ratio
8.4.5. Flexural reinforcement ratio

strength as a function of the flexural reinforcement ratio. The flexural reinforcement influences
only the calculation of the load-rotation response. Thus, the failure criteria are independent of
the flexural reinforcement ratio. This applies for the CSCT as well as for the proposed model.
Regarding the comparison to the test results, no tendency with respect to the flexural
reinforcement ratio can be noted for the comparison of the results of the proposed model to the
test results (Figure 8.27b). Additionally, it can be noted that the proposed model considers the
influence of the flexural reinforcement similarly to the CSCT since no tendency can be seen by
comparing the results of the two approaches (Figure 8.27c).
(a)
(b)
3
VR,test/VR,CSCT
2.5
VR / VR,pred []
(c)
no shear reinf.
crushing
within
outside
lexural failure
VR,CSCT/VR,model
VR,test/VR,model
2
1.5
1
0.5
0
0
2
flex [%]
2
flex [%]
2
flex [%]
function of the flexural reinforcement ratio
199
Chapter 8
8.4.6. Concrete compressive strength

strength as a function of the concrete compressive strength. The CSCT and the proposed model
consider the concrete strength for the calculation of the flexural strength influencing somewhat
the load-rotation response. Regarding the failure criteria, the CSCT relates the punching
strength to the square root of the compressive strength for punching of slabs without shear
reinforcement, for the concrete contribution for failure within the shear-reinforced area and for
the crushing of the concrete strut. The proposed model relates also the punching strength to the
square root of the compressive strength for punching of slabs without shear reinforcement, for
the concrete contribution for failure within the shear-reinforced area and for the concrete
contribution for failure of the concrete strut. However, the proposed model relates the strength
of the compression strut directly to the compressive strength. With respect to the comparison
with the test results, it can be noted that most tests were performed with normal strength
concrete leading to a cloud of points as already observed in the code validation in Chapter 4.
Consequently, no clear tendency can be seen for the proposed model or the CSCT. Although
from a theoretical point of view, an investigation of higher strength concrete may lead to
interesting observations, from a practical point of view, it has to be noted that the general gains
by using high-strength concrete for flat slabs are rather limited.
(a)
(b)
3
VR,test/VR,CSCT
VR / VR,pred []
2.5
(c)
no shear reinf.
crushing
within
outside
lexural failure
VR,CSCT/VR,model
VR,test/VR,model
2
1.5
1
0.5
0
0
20
40
60
fc [MPa]
80
20
40
60
fc [MPa]
80
20
40
60
fc [MPa]
80
function of the concrete compressive strength
8.4.7. Maximum aggregate size

strength as a function of the maximum aggregate size. The proposed model considers the
aggregate size only in the calculation of the concrete contribution to the punching strength for
both failure modes. Thus, the proposed model does not account for the aggregate size for the
strength of the compression strut. In contrast, the CSCT considers the aggregate size for all
200
failure modes. Nevertheless, the proposed model leads to good predictions for all aggregate
sizes (Figure 8.29b) and shows similar tendency as the CSCT does (Figure 8.29c).
(a)
(b)
3
VR,test/VR,CSCT
2.5
VR / VR,pred []
(c)
no shear reinf.
crushing
within
outside
lexural failure
VR,CSCT/VR,model
VR,test/VR,model
2
1.5
1
0.5
0
0
10
20
dg [mm]
30
40
10
20
dg [mm]
30
40
10
20
dg [mm]
30
40
function of the maximum aggregate size
8.4.8. Predicted rotation at failure

Similar to the CSCT and the fib Model Code 2010 (MC 2011), the proposed model defines the
punching strength as a function of the rotation of the slab. Therefore, an interesting aspect is the
prediction of the rotation at failure. Figure 8.30 shows the ratio of the experimentally obtained
punching strength and the predicted strength as a function of the predicted rotation at failure for
MC 2010, the CSCT, and the proposed model. Generally, it can be noted that MC 2010 and the
CSCT predict smaller rotations at failure than the proposed model does. Since MC 2010 is a
code provision and thus lower strength are predicted, the prediction of the rotations are smaller
as well, which explains the lower predicted rotations compared to the predictions of the CSCT
and the proposed model. However, more important in this case is that MC 2010 does not show
any tendency regarding the ratio of the experimentally obtained punching strength and the
predicted strength with respect to the predicted rotations.
The CSCT seems to have a tendency of predicting lower ratios of the experimentally obtained
punching strength and the predicted strength when larger rotations are predicted. This tendency
can be explained by the overestimation of the slab stiffness. For slab specimens for which
smaller rotations are predicted, the discrepancies in the load-rotation response are smaller
leading to better estimates of the punching strength. Furthermore, it can be seen that this affects
mainly the predicted failure mode of crushing of the concrete strut since generally larger
rotations are predicted for this failure mode. This observation corresponds with the statistical
values presented in Chapter 4 by which it can be seen that for failure due to crushing of the
concrete strut the average of the ratios of the experimentally obtained punching strength and the
predicted strength is smaller (Average = 0.92) than for failure within and outside the shearreinforced area (Average = 1.03 and 1.00, respectively).
201
Chapter 8
In contrast, the proposed model shows the opposite tendency. The proposed model predicts
larger ratios of the experimentally obtained punching strength and the predicted strength for
specimens by which larger rotations are predicted. Additionally, Figure 8.30c shows that for
larger predicted rotation, a failure outside the shear-reinforced area is estimated. It can be noted
that all the tests with predictions of large rotations ( > 30) had either a ratio of the column
size to the effective depth c/d > 2.0 or a flexural reinforcement ratio < 0.5% so that the
flexural limit is obtained. Consequently, the proposed model may underestimate the flexural
strength in the case of failure outside the shear-reinforced area. In addition, it can be noted that
due to the prediction of a less stiff response of the slabs, the proposed model generally leads to
larger estimates of rotations at failure than the CSCT does.
(a)
(b)
3
MC 2010 Level II
2.5
(c)
no shear reinf.
crushing
within
outside
CSCT
Proposed model
2
1.5
1
0.5
0
0
10
20
30
R []
40
50
10
20
30
R []
40
50
10
20
30
R []
40
50
predicted punching strength as a function of the predicted rotation at failure
In addition to the ratio of the experimentally obtained punching strength and the predicted
strength, the ratio of the experimentally obtained rotation at failure and the predicted rotation at
failure was investigated. However, since the rotation was not measured or published in most
experimental tests, only a small selection of tests was considered. In fact, the set of specimens
consists of 26 test specimens from the test campaign performed within this research project and
from a test campaign performed by Gomes and Regan (Gomes and Regan 1999a).
Figure 8.31 shows the ratio of the experimentally obtained rotation at failure and the predicted
rotation at failure as a function of the shear reinforcement ratio for MC 2010, the CSCT, and the
proposed model. MC 2010 shows rather scattered results. Additionally, it can be noted that a
clear tendency exists with respect to the shear reinforcement ratio. The larger the shear
reinforcement ratio is the larger the ratio of the experimentally obtained rotation at failure and
the predicted rotation at failure is. Nevertheless, it has to be reminded that MC 2010 is a design
provision and thus it is supposed to rather deliver safe estimate for the punching strength than to
predict the actual rotation capacity.
202
Compared to MC 2010, the CSCT shows less scatter and a lower average. However, the same
tendency occurs with respect to the shear reinforcement ratio. The larger the shear
reinforcement ratio is the larger the ratio of the experimentally obtained rotation at failure and
the predicted rotation at failure is. This can be explained by the overestimation of the stiffness
as it was already discussed in Chapter 4. This overestimation is more dominant in the case of
larger load levels as it is the case for large shear reinforcement ratios. This leads to the tendency
that the rotations will be clearly underestimated if large amount of shear reinforcement is
present.
In contrast to MC 2010 and the CSCT, the proposed model leads to an average that is close to
1.0 and to a rather small scatter with a coefficient of variation of only 13.9%. Moreover, no
tendency occurs with respect to the shear reinforcement ratio. Thus, it can be concluded that the
load-rotation response of the slab and the failure criteria are properly addressed.
(a)
(b)
(c)
R,test / R,pred []
3.5
MC 2010 Level II
Avg. : 1.79
COV : 31.5%
CSCT
Avg. : 1.53
COV : 22.9%
Proposed model
Avg. : 1.11
COV : 13.2%
3
2.5
2
1.5
1
no shear reinf.
crushing
within
outside
0
0
0.5
0.5
1
w [%]
1.5
0.5
1
w [%]
1.5
0.5
1.5
w [%]
Figure 8.31: Ratio of the measured rotation at failure by tests from literature (Gomes and Regan 1999a)
and from within this research project to the predicted rotation at failure as a function of the
shear reinforcement ratio
203
9. Conclusions and Future Research

The research presented herein focused on the investigation of punching of flat slabs with large
amounts of shear reinforcement. In addition to an experimental and numerical investigation of
the test specimens, a principal objective of the research project was the development of an
analytical model that enables accurate predictions of the punching strength and the rotation
capacity of flat slabs with large amounts of shear reinforcement.
The experimental campaign consisted of sixteen full-scale flat slab specimens. The investigated
parameters were the column size, the slab thickness, the amount of shear reinforcement, and the
type of the shear reinforcement system. Various continuous measurements such as the slab
rotation, the vertical displacement and the strains in the shear reinforcement were taken.
Additionally, certain values could be estimated based on measurements such as the shear
deformations at the column face and the forces transferred by the shear reinforcement.
With respect to the parameter varied in the experimental campaign, current design codes and the
CSCT were validated. Thereby, the predicted punching strength was shown as a function of the
experimentally investigated parameters. In addition, results of 132 punching tests from within
this research and from literature were used for the validation of the design codes and the CSCT.
The predicted punching strength was compared to the experimentally obtained strength. The
ratio of the predicted and the experimentally obtained strength was investigated for slabs with
and without shear reinforcement and for slabs with different predicted failure modes. This ratio
was again shown as a function of several parameters.
The flexural slab responses of the test specimens were numerically modeled to investigate the
slab deformations. For this, an approach based on the finite element method was developed by
which the moment-curvature relationships were calculated using plane stress fields. The thereby
obtained bending and torsional stiffness served as input parameters for a linear-elastic finite
element analysis. The results of these calculations were validated with the load-displacement
response of pure bending tests and pure torsion tests found in literature. Additionally, the
predicted load-rotation curves were compared to load-rotation measurements of punching tests
of specimens with and without shear reinforcement.
The flexural finite element approach was combined with an existing in-plane analysis program.
This combination enabled the in-depth analysis of the global and local slab behavior. The
calculated deformation of the slab was compared to displacement and surface strain
measurements. In addition, the force transfer path in the vicinity of the column was analyzed in
detail.
205
Chapter 9
Based on the experimental and numerical investigation, an analytical model was developed. For
this, a general slab behavior and failure mechanisms with respect to failure within the shearreinforced area and to failure of the concrete strut close to the column were described. Based on
the assumed slab behavior and failure mechanisms, formulations for the prediction of the loadrotation response of the slab and for failure criteria were developed enabling the prediction of
the punching strength and the rotation capacity.
The proposed model was analyzed with respect to the parameters investigated in the
experimental campaign. Thereby, the predicted punching strength and the predicted rotation at
failure were shown as a function of the investigated parameters. Additionally, the proposed
model was validated with results of 132 punching tests from within this research and from
literature. Furthermore, a direct comparison to the predictions of the CSCT for the same set of
tests showed the different performance of the two approaches as a function of several
parameters such as the effective depth, the column size, the flexural reinforcement ratio, and the
shear reinforcement ratio.
9.1. Conclusions
The measurement during the experimental campaign delivered valuable information concerning
the behavior of the slab and the performance of the shear reinforcement. The measurements and
the observations of the experimental tests lead to the following conclusions:
206
Shear reinforcement improves the performance of flat slab-column connections with

respect to the punching strength and the deformation capacity. In comparison to slabs
without shear reinforcement, the punching strength as well as the rotation capacity
increased significantly. The improvement in performance of the slab is influenced by
several parameters such as the slab geometry, the amount of shear reinforcement, and
the type of the shear reinforcement system.
The strain measurements on the studs showed that the stresses in the shear
reinforcement may be below the yielding strength prior to punching failure. Thus, the
shear reinforcement may not be fully activated depending on the slab rotation and the
arrangement of the shear reinforcement.
The strain measurements on the studs indicated additionally that forces are directly
transferred from the first and from the second row of studs to the column. However, a
certain amount of force is transferred from the second row of studs to the first row of
studs. Considering this aspect, it can be noted that the amount of the forces transferred
by the shear reinforcement is less than the applied shear force, which confirms that the
concrete contributes as well to the punching strength. Moreover, it was shown that the
concrete contribution to the punching strength depends on the slab rotation.
Conclusions and Future Research
Vertical displacement measurements showed that large shear deformation occur at the
column face for specimens with shear reinforcement. Since these shear deformations are
much larger than the measured strain in the shear reinforcement, it can be concluded
that the shear deformations occur localized at the column face. Thus, the opening of the
cracks crossing the shear reinforcement can mainly be attributed to the slab rotation
confirming a main hypothesis of the CSCT.
The comparison of codes and the CSCT to tests within this research and to tests found in
literature showed generally a good agreement. From this comparison, following conclusions are
drawn:
ACI 318-11 leads generally to conservative results compared to test results within this
research and from literature. However, a large scatter of the results is obtained, which
may in certain cases lead to unsafe design. While the results for predicted failure within
the shear-reinforced area and failure due to crushing of the concrete strut are rather
conservative but showing acceptable agreement with the test results, the failure outside
the shear-reinforced area lead to conservative and largely scattered results.
EC2 2004 showed in several cases concerning overestimations of the punching strength
when compared to test results within this research and from literature. While the ratios
of the predicted to the experimentally obtained punching strength are in an acceptable
range for specimens without shear reinforcement, with predicted failure within the
shear-reinforced area, and with predicted failure outside the shear-reinforced area, the
ratios of specimens with predicted failure due to crushing of the concrete strut show a
rather large scatter in combination with a low average. This may potentially lead to
unsafe design if crushing of the concrete strut is governing.
The National German Annex (NAD 2011) improves the drawbacks of EC2 2004
leading to significant improvement of the strength predictions. Especially, for the
prediction of failure due to crushing of the concrete strut, the scatter is reduced leading
to safe estimates. With respect to failure outside the shear-reinforced area, it has to be
noted that the effective depth was reduced leading to more conservative results. This
reduction however is not in the code but proposed by Hegger et al. (Hegger et al. 2010).
If this reduction is not considered the predicted punching strength will be larger leading
to somewhat lower safety levels.
SIA262 2003 leads generally to rather conservative results compared to test results
within this research and from literature. Especially in the case of the prediction of
failure within the shear-reinforced area, the punching strength is underestimated. This is
due to the fact that SIA262 2003 neglects any concrete contribution to the punching
strength. This is however already recognized and will be improved in a future revision
207
Chapter 9
of Swiss code by which the punching shear provisions will be similar to the provisions
of MC 2010.
MC 2010 leads to the best agreement with test results within this research and from
literature in comparison to the other investigated codes. Especially, the low scatter of
the ratio of the predicted to the experimentally obtained punching strength is notable.
Additionally, it can be noted that all failure modes are equally well addressed so that no
tendency occurs for a certain failure mode or for a certain parameter.
The strength predictions of the CSCT, which is the basis of the MC 2010, are generally
in a good agreement with the experimentally obtained punching strength for tests within
this research and for tests found in literature. However, the comparison of the loadrotation response revealed that the Quadrilinear method used in the combination with
the CSCT leads to the prediction of a stiffer response than the experimentally measured
one. Consequently, the predictions of the rotation at failure are generally lower than the
experimentally measured rotation.
The development of the constitutive model for the nonlinear finite element analysis (NLFEA)
enabled the modeling of the moment-curvature relationship of a slab element subjected to
bending and torsion. Using the determined moment-curvature response in combination with the
finite element method the slab specimens could be analyzed. Thereby, the numerical analysis
leads to following conclusions:
208
The analysis of a reinforced concrete slab element shows that the state of deformation
influences significantly the bending and torsional stiffness of the element. Moreover,
the interaction between bending and torsional behavior is pertinent. Thus, the bending
stiffness is influenced by the torsional curvature and the torsional stiffness is strongly
influenced by the bending curvature.
The nonlinear finite element approach developed within this research allows the
consideration of the interaction of the bending and torsional behavior. The comparison
to slab specimen subjected to pure bending and to pure torsion showed good agreement
between the predicted and the experimentally obtained load-displacement response.
Additionally, the predictions of the load-rotation response of slab specimens without
shear reinforcement agreed well with the experimentally obtained load-rotation curve.
However, in the case of slabs with shear reinforcement, which allows larger shear
forces to develop, the purely flexural analysis did not lead to acceptable results.
Generally, the NLFEA led to a prediction of a stiffer response compared to the
experimentally measured response. Thus, in regions with large concentrated shear
forces, a numerical analysis solely considering the flexural behavior is not sufficient to
model the response of the slab accurately.
Conclusions and Future Research
The combination of the flexural response and an existing in-plane finite element
program improved the modeling of the test specimens with shear reinforcement
allowing the prediction of the slab deformations. These predictions were compared to
several test measurements such as the rotation, the vertical displacement, and the
surface deformation on the top and bottom side of the slab. This comparison showed
that the predictions correspond well to the measurements leading to the conclusion that
the general response of such a slab specimen can be well captured by the numerical
analysis.
The analytical model developed within this research project enables the prediction of the
punching strength and rotation capacity. Compared to the CSCT, improvements could be
achieved by slight changes in the calculation method of the load-rotation response and in the
formulations for the failure criteria.
The developed model for the load-rotation response considers a global slab part and a
local, shear-critical slab part. This separation allows for different calculation approaches
for the global part, which is mainly dependent on the flexural slab behavior, and for the
shear-critical part that is mainly dependent on the force transfer path in the column
vicinity. The importance of the estimated curvature at the intersection plane defining the
forces in the shear-critical region. The application of the proposed approach results in
excellent agreement with the experimentally obtained load-rotation response.
The developed failure criterion for failure within the shear-reinforced area, which is
based on the mechanical model of the CSCT, shows good agreement with the test
results. Moreover, the model allows the applicable consideration of the actual location
of the shear reinforcement so that the activation phase of the shear reinforcement can be
accurately modeled leading to a good agreement with experimentally obtained results.
The developed failure criterion for crushing of the concrete strut accounts for several
parameters such as the concrete compressive strength, the transverse strain in the
compression strut, the anchorage condition of the shear reinforcement, the location of
the shear reinforcement, and the amount of compression at the column face. The
consideration of these parameters is seen as beneficial enabling a good prediction of the
punching strength and the rotation capacity.
209
Chapter 9
9.2. Recommendations for future research

The research presented within this thesis focused on the investigation of punching of flat slabs
with large amounts of shear reinforcement. Several aspects such as non-symmetric slabs
(geometry and/or loading), slabs with inclined shear reinforcement, and prestressed slabs were
not considered in this research. Additionally, the failure modes investigated within this thesis
were limited to failure within the shear-reinforced area and failure due to crushing of the
concrete strut. Thus, other failure modes such as failure outside the shear-reinforced area were
not considered. Therefore, research on special slab configurations and on other failure modes is
recommended for a further enhancement of the understanding of punching of flat slabs with
shear reinforcement and for the validation of the applicability of the herein presented numerical
and analytical model.
In addition to the topics that were out of the scope of this thesis, certain aspects were limited to
the problematic of this research which could be extended to other problems and certain
hypothesis were made with respect to the analytical model that should be further investigated.
Generally, the following two recommendations are seen as the most interesting:
210
The nonlinear finite element approach presented within this research was developed for
the analysis of the load-deformation response of the punching test specimens. However,
this approach can be extended to calculations of deformations of more general slabs.
For this, the interaction of bending and torsional moments should be further
investigated and for more cases experimentally validated. In addition, the
implementation should be considered so that the structure can be analyzed within one
simple program.
With respect to failure due to crushing of the concrete strut, the research showed that
the consideration of the transverse strains, the spacing of the shear reinforcement, and
the anchorage conditions enables a good prediction of the punching strength and the
rotation capacity. However, the influence of each parameter is solely based on empirical
data. An experimental and numerical analysis of a local model considering the state of
deformation is recommended.
Bibliography
ACI 318, Building Code Requirements for Structural Concrete, ACI 318-11, American
Concrete Institute, ACI Committee 318, 503 pp., Detroit, USA, 2011.
Alexander, S. D. B. and Hawkins, N. M., A Design Perspective on Punching Shear, ACI, SP
232, pp. 97-108, USA, 2005.
Alvarez, M., Kppel, S., and Marti, P., Rotation capacity of reinforced concrete slabs, ACI
Structural Journal, Vol. 97, No. 2, pp. 235-242, 2000.
Andersson, J. L., Punching of Concrete Slabs with Shear Reinforcement, Transactions of the
Royal Institute of Technology, 56 pp., Stockholm, Sweden, 1963.
Andr, H.-P., Dbelleisten zur Verhinderung des Durchstanzens bei hochbelasteten
Flachdecken, Bautechnik, Nr. 56, H7, pp. 244-247, 1979.
Andr, H.-P., Dilger, W. H., and Ghali, A., Durchstanzbewehrung fr Flachdecken, Betonund Stahlbetonbau, Vol. 74, No. 5, pp. 129-132, Berlin, Germany, 1979.
Andr, H.-P., Zum Tragverhalten von Flachdecken mit Dbelleisten, Bewehrung in
Auflagerbereich, Beton- und Stahlbetonbau, Nr. 76, pp. H3.53-57, H4.100-104, 1981.
Bell, J. C. and Elms, D. G., Non-linear Analysis of Reinforced Concrete Slabs, Magazine of
Concrete Research, Vol. 24, No. 79, pp. 63-70, UK, 1972.
Beutel, R., Durchstanzen schubbewehrter Flachdecken im Bereich von Innensttzen, RWTH,
267 pp., Aachen, Germany, 2003.
Birkle, G., Punching of Flat Slabs: The Influence of Slab Thickness and Stud Layout,
Dissertation, University of Calgary, 217 pp., Calgary, Canada, 2004.
Birkle, G. and Dilger, W. H., Influence of Slab Thickness on Punching Shear Strength, ACI
Structural Journal, Vol. 105, No. 2, pp. 180-188, 2008.
Broms, C. E., Punching of Flat Plates - A Question of Concrete Properties in Biaxial
Compression and Size Effect, ACI Structural Journal, Vol. 87, No 3, pp. 292-304, 1990a.
Broms, C. E., Shear Reinforcement for Deflection Ductility of Flat Plates, ACI Structural
Journal, Vol. 87, No. 6, pp. 696-705, 1990b.
Broms, C. E., A method to avoid the punching failure mode, pp. 117-124, Stockholm, Sweden,
2000a.
Broms, C. E., Elimination of Flat Plate Punching Failure Mode, ACI Structural Journal, Vol.
97, No. 1, pp. 94-101, 2000b.
211
Broms, C. E., Concrete flat slabs and footings : Design method for punching and detailing for
ductility, KTH, Civil and Architectural Engineering, 114 pp., Stockholm, Sweden, 2005.
Broms, C. E., Ductility reinforcement for flat slabs in seismic areas, Magazine of Concrete
Research, Thomas Telford, Vol. 58, No 4, pp. 243-254, London, England, 2006.
Broms, C. E., Ductility of Flat Plates: Comparison of Shear Reinforcement Systems , ACI
Structural Journal, Vol. 104, No.6, pp. 703-711, 2007a.
Broms, C. E., Flat Plates in Seismic Areas: Comparison of Shear Reinforcement Systems , ACI
Structural Journal, Vol. 104, No.6, pp. 712-721, 2007b.
Carvalho, A. L., Melo, G. S., Gomes, R. B., and Regan, P. E., Punching Shear in PostTensioned Flat Slabs with Stud Rail Shear Reinforcement, ACI Structural Journal, Vol. 108, No.
5, pp. 523-531, Farmington Hills, USA, 2011.
Chana, P. S., A prefrabricated shear reinforcement system for flat slabs, Proceedings of the
Institution of Civil Engineers: Structures and Buildings, Vol. 99, pp. 345-358, UK, 1993.
Chana, P. S. and Desai, S. B., Design of shear reinforcement against punching, The Structural
Engineering, Vol. 70, pp. 159-164, 1992.
Condron, T. L., Principles of Design and Results of Tests on Girderless Floor Construction of
Reinforced Concrete, ACI Journal Proceedings, Vol. 9, No. 12, pp. 116-126, USA, 1913.
Cook, R. D., Malkus, D. S., Plesha, M. E., and Witt, R. J., Concepts and Application of
Finite Element Analysis, John Wiley and Sons Inc., 4th ed., 719 pp., USA, 2002.
Dechka, D. C., Response of Shear-Stud-Reinforced Continuous Slab-Column Frames to
Seismic Loads, University of Calgary (Alberta), Departement of Civil Engineering, 471 pp.,
Calgary, Canada, 2001.
Dilger, W. H. and Ghali, A., Shear Reinforcement for Concrete Slabs, Journal of the Structural
Division - ASCE, Vol. 107, pp. 2403-2420, USA, 1981.
EC2, Eurocode 2 Design of concrete structures, part 1-1 : General rules and rules for
buildings, 241 pp., 2004.
EC2 Commentary, Commentary Eurocode 2, European Concrete Platform, 168 pp., Brussels,
Belgium, 2008.
Eddy, H.T. and Turner, C.A.P., Concrete-Steel Construction , Heywood Mfg. Co., 438 pp.,
Minneapolis, USA, 1914.
Elgabry, A. and Ghali, A., Design of Stud-Shear Reinforcement for Slabs, ACI Structural
Journal, Vol. 87, No. 3, pp. 351-361, USA, 1990.
212
Fernndez Ruiz, M., Muttoni, A., and Kunz, J., Strengthening of flat slabs against punching
shear using post-installed shear reinforcement, ACI Structural Journal, Vol. 107, No. 4, pp.
434-442, USA, 2010.
Fernndez Ruiz, M. and Muttoni, A., On Development of Suitable Stress Fields for Structural
Concrete, ACI, Structural Journal, Vol. 104, No. 4, pp. 495-502, Farmington Hills, USA, 2007.
Fernndez Ruiz, M. and Muttoni, A., Applications of the critical shear crack theory to
punching of R/C slabs with transverse reinforcement, ACI Structural Journal, Vol. 106 N 4, pp.
485-494, USA, 2009.
FIB, Punching of structural concrete slabs, fib bulletin 12, 307 pp., Lausanne, Switzerland,
2001.
Ghali, A., An Efficient Solution to Punching of Slabs, Concrete International, pp. 50-54, USA,
1989.
Ghali, A. and Hammill, N., Effectiveness of Shear Reinforcement in Slabs, Concrete
International, pp. 60-65, USA, 1992.
Gomes, R. B. and Regan, P. E., Punching Resistance of RC Flat Slabs with Shear
Reinforcement, ASCE Journal of Structural Engineering, Vol. 125, No. 6, pp. 684-692, Reston,
USA, 1999a.
Gomes, R. B. and Regan, P. E., Punching strength of slabs reinforced for shear with offcuts of
rolled steel I-section beams, Magazine of Concrete Research, Vol. 51, No. 2, pp. 121-129,
London, UK, 1999b.
Guandalini, S., Poinonnement symtrique des dalles en bton arm, Thse de doctorat No.
3380, 289 pp., Lausanne, Switzerland, 2005.
Guandalini , S. and Muttoni , A., Essais de poinonnement symtrique des dalles en bton
arm sans armature l'effort tranchant, Rapport d'essai, 85 pp., Lausanne, 2004.
Guidotti , R., Fernndez , Ruiz M., and Muttoni , A., Essais de poinonnement de dalles en
bton vibr et bton autocompactant avec diffrentes tailles maximales du granulat, Rapport
d'essai, 67 pp., Lausanne, 2009.
Guidotti, R., Poinonnement des planchers-dalles avec colonnes superposes fortement
sollicites, Thse EPFL No. 4812, 230 pp., Lausanne, Switzerland, 2010.
Hammill, N. and Ghali, A., Punching Shear Resistance of Corner Slab-Column Connections,
ACI Structural Journal, Vol. 91, No. 6, pp. 697-707, Detroit, USA, 1994.
Hand, F. R., Pecknold, D. A., and Schnobrich, W. C., Nonlinear Layered Analysis of RC
Plates and Shells, Journal of the Structural Division, Proceedings of the American Society of
Civil Engineers, Vol. 99, pp. 1491-1505, USA, 1973.
213
Husler, F., Zum maximalen Durchstanzwiderstand von Flachdecken mit und ohne
Vorspannung, RWTH, 213 pp., Aachen, Germany, 2009.
Hawkins, N. M., Shear Strength of Slabs with Shear Reinforcement, ACI SP-42-34, Vol. 2, pp.
785-815, Detroit, USA, 1974.
Hegger, J., Ricker , M., Ulke, B., and Ziegler, M., Untersuchungen zum Durchstanzverhalten
von Stahlbetonfundamenten, Beton- und Stahlbetonbau, Vol. 101, No 4, pp. 233-243, Germany,
2006.
Hegger, J., Husler, F., and Ricker , M., Zur maximalen Durchstanztragfhigkeit von
Flachdecken, Beton - und Stahlbetonbau, Vol. 102, pp. 770-777, Berlin, Germany, 2007.
Hegger, J., Ricker , M., and Sherif, A., Punching Strength of Reinforced Concrete Footings,
ACI Structural Journal, Vol. 106 , pp. 706-716, USA, 2009.
Hegger, J., Walraven, J. C., and Husler, F., Zum Durchstanzen von Flachdecken nach
Eurocode 2, Beton - und Stahlbetonbau, Vol. 105, pp. 206-215, Berlin, Germany, 2010.
Jofriet, J. C. and McNeice, G. M., Finite Element Analysis of Reinforced Concrete Slabs,
Journal of the Structural Division, Proceedings of the American Society of Civil Engineers, Vol.
97, pp. 785-806, USA, 1971.
Kaufmann, W. and Marti, P., Structural concrete: cracked membrane model, Journal of
Structural Engineering, Vol. 124, No. 12, pp. 1467-1475, USA, 1998.
Kenel, A. and Marti, P., Faseroptische Dehnungsmessungen an einbetonierten
Bewehrungsstben, IBK Bericht, 93 pp., Zrich, Switzerland, 2001.
Kinnunen, S., Punching of Concrete Slabs with two-way Reinforcement, Transactions of the
Royal Institute of Technology, No. 198, 108 pp., Stockholm, Sweden, 1963.
Kinnunen, S. and Nylander, H., Punching of Concrete Slabs Without Shear Reinforcement,
Transactions of the Royal Institute of Technology, No. 158, 112 pp., Stockholm, Sweden, 1960.
Kostic, N., Topologie des champs de contraintes pour le dimensionnement des structures en
bton arm, Thse EPFL, No. 4414, 235 pp., Lausanne, Switzerland, 2009.
Kupfer, H. and Gerstle, H., Behavior of Concrete under Biaxial Stresses, ASCE Journal of
Engineering Mechanics, Vol. 99, No. 4, pp. 853-866, USA, 1973.
Langohr, P. H., Ghali, A., and Dilger, W. H., Special Shear Reinforcement for Concrete Flat
Plates, ACI Journal, Vol. 73, pp. 141-146, Detroit, USA, 1976.
Lips, S., Muttoni, A., and Fernndez-Ruiz, M., Durchstanzversuche an Deckenausschnitten
mit FIDECA Durchstanzbewehrung, EPFL-IBton, 55 pp., Lausanne, Switzerland, 2010.
Lips, S., Muttoni, A., and Fernndez-Ruiz, M., Durchstanzversuche an Deckenausschnitten
mit Durchstanzbewehrung, EPFL-IBton, 88 pp., Lausanne, Switzerland, 2012a.
214
Lips, S., Muttoni, A., and Fernndez-Ruiz, M., Experimental Investigation on the Punching
Strength and the Deformation Capacity of Shear-Reinforced Slabs, ACI Structural Journal,
USA, 2012b. (Accepted for publication)
Loov, R. E., Review of A23.3-94 Simplified Method of Shear Design and Comparison with
Results using Shear Friction, Canadian Journal of Civil Engineering, Vol. 25, pp. 437-450,
Canada, 1998.
Maillart, R., Zur Entwicklung der unterzuglosern Decke in der Schweiz und in Amerika,
Schweizerische Bauzeitung, Vol. 87, No. 21, pp. 263-267, Zrich, Switzerland, 1926.
Marti , P., Leesti , P., and Khalifa , W. U., Torsion Tests on Reinforced Concrete Slab
Elements, ASCE Journal of Structural Engineering, Vol. 113, pp. 994-1010, 1987.
Marti, P., Alvarez, M., Kaufmann, W., and Sigrist, V., Tension chord model for structural
concrete, Structural Engineering International, Vol. 8, No. 4, pp. 287-298, USA, 1998.
MC, Model Code 2010 (Final Draft), Fdration internationale du bton fib, Lausanne,
Switzerland, 2011.
Megally, S. H. and Ghali, A., Design Considerations for Slab-Column Connections in Seismic
Zones, ACI Structural Journal, Vol. 91, No. 3, pp. 303-314, Detroit, USA, 1994.
Mokhtar, A. S., Ghali, A., and Dilger, W. H., Stud Shear Reinforcement for Flat Concrete
Plates, ACI Journal, No. 82-60, pp. 676-683, Detroit, USA, 1985.
Muttoni, A., Punching shear strength of reinforced concrete slabs without transverse
reinforcement, ACI Structural Journal, Vol. 105, No. 4, pp. 440-450, 2008.
Muttoni, A., Fernndez Ruiz, M., and Kostic, N., Champs de contraintes et mthode des
bielles et tirants - Applications dans la conception et le dimensionnement des structures en
bton arm, EPFL - IBETON, 117 pp., Lausanne, Switzerland, 2011.
Muttoni, A. and Fernndez Ruiz, M., The Critical Shear Crack Theory as a mechanical
model for punching shear design and its application to code provisions, Fdration
Internationale du Bton, Bulletin 57, pp. 31-60, Lausanne, Switzerland, 2010.
NAD, DIN EN 1992-1-1/NA:2011 Nationaler Anhang - Eurocode 2: Bemessung und
Konstruktion von Stahlbeton- und Spannbetontragwerken, 101 pp., Berlin, Germany, 2011.
Oliveira, D. R., Melo, G. S., and Regan, P. E., Punching Strengths of Flat Plates with Vertical
or Inclined Stirrups, ACI Structural Journal, Vol. 97, pp. 485-491, USA, 2000.
Polak, M. A., Nonlinear Analysis of Reinforced Concrete Shells, University of Toronto, 195
pp., Toronto, Canada, 1992.
Polak, M. A. and Vecchio, F. J., NonLinear Analysis of Reinforced-Concrete Shells, ASCE
Journal of Structural Engineering, Vol. 119, No. 12, pp. 3439-3462, 1993.
215
Regan, P. E., Punching Shear in Prestressed Concrete Slab Bridges, Engineering Structures
Research group, Polytechnic of Central London, 230 pp., London, UK, 1983.
Regan, P. E., Shear Combs, Reinforcement against Punching, The Structural Engineering, Vol.
63B, No. 4, pp. 76-84, 1985.
Regan, P. E., Symmetric Punching of Reinforced Concrete Slabs, Magazine of Concrete
Research, Vol. 38, pp. 115-128, England, 1986.
Regan, P. E. and Samadian, F., Shear Reinforcement against punching in reinforced concrete
flat slabs, The Structural Engineer, pp. 24-31, London, England, 2001.
Seible, F., Ghali, A., and Dilger, W. H., Preassembled shear reinforcing units for flat plates,
ACI Journal, Proceedings 77, No. 1, pp. 28-35, 1980.
Shehata, I. A. E. M, Theory of punching in RC slabs, Ph.D Thesis, Polytechnic of Central
London, UK, 1985.
Shehata, I. A. E. M, Rational Method for Designing RC Slabs to Resist Punching, Journal of
Structural Engineering, Vol. 116, No. 7, pp. 2055-2060, USA, 1990.
Shehata, I. A. E. M and Regan, P. E., Punching in R.C. Slabs, Journal of Structural
Engineering, Vol. 115, No. 7, pp. 1726-1740, USA, 1989.
SIA 262, SIA 262 - Construction en bton, Socit Suisse des Ingnieurs et des Architectes, 94
p., Zrich, Switzerland, 2003.
Siburg, C. and Hegger, J., Punching of flat slabs - comparison of models, fib Symposium
Prague 2011, pp. 151-154, Prague, Czech Republic, 2011.
Sigrist, V., Zum Verformungsvermgen von Stahlbetontrgern, Dissertation, 160 pp., Zrich,
Switzerland, 1995.
Smulski, E., A Test of the S-M-I System of Flat-Slab Construction, ACI Journal Proceedings,
Vol. 14, No. 6, pp. 206-232, USA, 1918.
Talbot, A. N. and Gonnerman, H. F, Test of a Flat Slab Floor of the Western Newspaper
Union Building, University of Illinois Bulletin, Vol. XV, No. 38, 62 pp., USA, 1918.
Tassinari, L., Poinonnement symtrique des dalles en bton arm avec armature de
poinonnement, Thse EPFL No. 5030, 197 pp., Lausanne, Switzerland, 2011.
Taylor, F. W. and Thompson, S. E., A treatise on Concrete Plain and Reinforced, John Wiley
and Sons Inc., 3rd ed., 885 pp., New York, USA, 1916.
Thorenfeldt, E., Tomaszewics, A., and Jensen, J. J., Mechanical properties of high-strength
concrete and application in design, Proceeding on Utilisation of High Strength Concrete,
Stavanger, Norway, 1987.
216
Tozser, O., Shear Analysis Using Shear Friction - Size Effect and Prestressing, University of
Calgary (Alberta), Departement of Civil Engineering, 194 pp., Calgary, Canada, 1998.
Trautwein, L. M., Bittencourt, T. N., Gomes, R. B., and Bella, J. C. D., Punching Strength
of Flat Slabs with Unbraced Shear Reinforcement, ACI Structural Journal, Vol. 108, pp. 197205, USA, 2011.
Van der Voet, A. F., Dilger, W. H., and Ghali, A., Concrete Flat Plates with Well-Anchored
Shear Reinforcement Elements, Canadian Journal of Civil Engineering, Vol. 9, No. 1, pp. 107114, Canada, 1982.
Vaz Rodrigues, R., Shear Strength of Reinforced Concrete Bridge Deck Slabs, Thse EPFL
No. 3739, 289 pp., Lausanne, Switzerland, 2007.
Vecchio, F. J., Nonlinear Finite Element Analysis of Reinforced Concrete Membranes, ACI
Structural Journal, Vol. 86-S4, pp. 26-35, USA, 1989.
Vecchio, F. J. and Collins, M. P., The modified compression-field theory for reinforced
concrete elements subjected to shear, ACI Journal, Vol. 83, pp. 219-231, USA, 1986.
Vollum, R. L., Abdel-Fattah, T., Eder, M., and Elghazouli, A.Y. , Design of ACI-type
punching shear reinforcement to Eurocode 2, Magazine of Concrete Research, Vol. 62, No. 01,
pp. 3-16, London, UK, 2010.
217
Appendix A Test Database

Parameters
l
[mm]
Slab dimensions and shape where l correspond to the side length of a square slab
(sq) or to the diameter of a circular or octagonal slab(ci)
[mm]
Thickness of the slab
[mm]
Effective depth of the slab (distance from extreme compression fiber to the
centroid of the longitudinal tensile reinforcement)
[mm]
Column size and shape where c correspond to the side length of a square column
(sq) or to the diameter of a circular column (ci)
[%]
Average flexural reinforcement ratio
[%]
Shear reinforcement ratio calculated according to Appendix C
fc
[MPa]
Mean cylindrical concrete compressive strength

(Conversion factors fc=0.83fc,cube150mm; fc=0.79fc,cube100mm)
dg
[mm]
Maximum diameter of the aggregate
fy
[MPa]
Mean yielding strength of the flexural reinforcement steel
fyw
[MPa]
Mean yielding strength of the shear reinforcement steel

Shear reinforcement system:
(MC: ksys; CSCT: ; mod: k2)
(a)
corrugated double headed shear studs
(2.8; 3.0; 0.75)
(b)
smooth double headed shear studs
(2.8; 3.0; 0.75)
(c)
steel offcuts
(2.8; 3.0; 0.75)
(d)
headed stirrups
(2.8; 3.0; 0.75)
(e)
stirrups with lap at the vertical branch
(2.4; 2.5; 0.50)
(f)
stirrups or shear links
(2.4; 2.5; 0.50)
(g)
continuous stirrups or cages of shear links (2.4; 2.5; 0.50)
system
(a)
(b)
(c)
(d)
(e)
(f)
(g)
A1
Vtest/Vpred
[-]
Ratio of the experimentally obtained punching strength to the predicted

punching strength where following definitions are used:
ACI: Calculated according to ACI 318-11 (ACI 318 2011)
EC:
Calculated according to Eurocode 2 2004 (EC2 2004)
NAD: Calculated according to German National Annex (NAD 2011)
(Values in parenthesis: Changes if factor 1.9 is used instead of 1.4 for
double headed studs)
SIA: Calculated according to Swisscode 262 2003 (SIA 262 2003)
MC:
Calculated according to Modelcode 2010 (MC 2011)
CSCT: Calculated according to the Critical Shear Crack Theory
(Muttoni 2008, Fernndez-Ruiz and Muttoni 2008)
mod: Calculated according to the proposed model
and the predicted failure mode where following definitions are used:
p:
punching without shear reinforcement
c:
crushing of the concrete strut
w:
failure within the shear-reinforced area
o:
failure outside the shear-reinforced area
f:
flexural failure
A2
c: crushing of concrete strut
o: failure outside shear-reinforced area
w: failure within shear-reinforced area
f: flexural failure
Andersson, J. L., Punching of Concrete Slabs with Shear Reinforcement, Transactions of the
Royal Institute of Technology, Stockholm, Sweden, 1963.
l
[mm]
A-S2-82
A-S2-83
A-S2-82
A-S2-83
h
[mm]
dm
[mm]
150
150
120
119
c [mm]
fc
[MPa]
dg
[mm]
fy
[MPa]
fyw
[MPa]
system
1.08
1.09
0.58
0.58
28.1
24.3
32
32
442
442
294
294
g
g
(ci)
(ci)
VR,test
[kN]
VR,test /VR,ACI
(2011)
VR,test /VR,EC
(2004)
VR,test /VR,NAD
(2011)
VR,test /VR,SIA
(2003)
VR,test /VR,MC
(2010)
459
459
1.64
1.78
1.23
1.30
1.39
1.47
1.34
1.42
1.22
1.29
o
o
(ci)
(ci)
w
[%]
1840
1840
o
o
300
300
L
[%]
o
o
w
w
o
o
VR,test
/VR,CSCT
0.98
1.04
VR,test /VR,mod
o
o
0.98
1.02
o
o
A3
Beutel, R. and Hegger, J., The effect of anchorage on the effectiveness of the shear
reinforcement in the punching zone, Cement and Concrete Composites, 11 pp., 2002.
Beutel, R., Durchstanzen schubbewehrter Flachdecken im Bereich von Innensttzen, RWTH,
267 pp., Aachen, Germany, 2003.
l
[mm]
Z1
Z2
Z3
Z4
Z5
Z6
V1
V2
V3
V4
P1
P1-I
P1-II
P2-I
P2-II
P2-III
P3-I
P4-III
P5-I
P6-I
P7-I
2980
2980
2980
2980
2980
2980
2750
2750
2980
2980
2820
2820
2820
2820
2750
2750
2750
2750
2750
2750
2750
VR,test
[kN]
Z1
Z2
Z3
Z4
Z5
Z6
V1
V2
V3
V4
P1
P1-I
P1-II
P2-I
P2-II
P2-III
P3-I
P4-III
P5-I
P6-I
P7-I
A4
1323
1442
1616
1646
2024
1954
1250
1424
1182
1680
615
1187
1091
1362
1109
1276
1624
1522
1936
2349
2117
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(ci)
(ci)
(sq)
(sq)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
h
[mm]
dm
[mm]
285
285
285
285
285
285
290
290
285
383
230
230
230
230
230
230
270
267
270
273
275
250
250
250
250
250
250
250
250
250
350
190
190
190
190
190
190
220
222
220
223
229
200
200
200
200
264
200
200
200
200
200
400
400
400
400
400
400
320
320
320
320
320
VR,test
/VR,EC
(2004)
VR,test
/VR,ACI
(2011)
1.13
1.19
1.40
1.24
1.42
1.36
0.96
1.00
1.04
0.98
0.88
1.52
1.43
1.28
1.36
1.29
1.42
1.20
1.21
1.43
1.33
c [mm]
c
c
c
c
c
c
c
c
c
o
p
o
o
o
o
o
c
c
c
c
c
1.26
1.30
1.57
1.27
1.31
1.31
0.98
0.98
1.19
1.09
0.87
1.32
1.24
0.97
1.01
0.97
1.12
0.82
0.89
0.92
0.94
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
L
[%]
w
[%]
fc
[MPa]
dg
[mm]
fy
[MPa]
fyw
[MPa]
system
0.80
0.80
0.80
0.80
1.25
1.25
0.80
0.80
0.80
0.50
0.80
0.80
0.80
0.80
0.81
0.81
1.15
1.13
1.35
1.75
1.30
0.65
0.65
0.70
0.75
0.80
0.91
0.57
0.54
0.57
0.44
0.33
0.33
0.60
0.50
0.64
0.33
0.53
0.55
0.93
0.79
24.8
26.3
24.1
31.5
28.0
37.3
30.8
36.2
23.2
26.0
21.9
27.3
26.2
37.9
29.8
37.5
23.2
27.8
45.3
46.3
40.0
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
889
889
889
889
562
562
917
917
901
917
572
572
572
549
549
549
557
557
557
532
557
580
580
580
580
540
540
544
544
544
544
570
570
592
592
592
597
597
596
563
596
a
a
a
a
a
a
a
a
a
a
e
f
e
f
f
e
f
e
e
e
VR,test /VR,NAD
(2011)
c
c
c
c
c
c
c
c
c
c
p
o
o
f
o
o
w
o
w
o
o
1.28 (0.94)
1.36 (1.00)
1.57 (1.16)
1.47 (1.10)
1.40 (1.03)
1.42 (1.04)
1.12 (0.83)
1.21 (0.89)
1.17 (0.86)
1.18 (1.12)
0.87
1.56
1.45
1.15
1.17
1.12
1.24
1.09
1.12
1.21
1.22
c
c
c
c (o)
c
c
c
c
c
c
p
o
o
o
o
o
c
c
c
c
c
VR,test
/VR,MC
(2010)
VR,test
/VR,SIA
(2003)
1.13
1.20
1.39
1.29
1.35
1.28
1.14
1.38
1.08
1.15
0.97
1.50
1.37
1.27
1.11
1.19
1.90
1.12
1.36
1.23
1.25
c
c
c
c
c
c
w
w
w
w
p
w
w
c
o
c
w
c
w
c
c
1.25
1.33
1.54
1.42
1.56
1.49
1.17
1.26
1.23
1.19
1.01
1.48
1.19
1.34
1.23
1.25
1.59
1.06
1.21
1.16
1.18
VR,test
/VR,CSCT
w
w
w
w
w
w
w
w
w
w
p
o
o
o
o
o
w
c
w
c
c
0.90
0.96
1.07
0.94
1.09
1.03
0.85
0.95
0.91
0.84
0.76
1.09
0.90
1.04
0.93
0.97
1.14
0.80
0.83
0.89
0.90
w
w
w
w
w
w
w
w
w
w
p
o
o
o
o
o
w
c
c
c
c
VR,test
/VR,mod
1.02
0.98
1.14
1.01
1.08
1.04
0.88
0.89
0.94
0.89
0.77
1.13
0.95
1.18
1.03
1.08
1.14
0.82
0.82
0.83
0.84
c
c
c
w
w
w
w
w
w
w
p
o
o
o
o
o
w
c
w
o
o
Birkle, G., Punching of Flat Slabs: The Influence of Slab Thickness and Stud Layout,
Dissertation, University of Calgary, 217 pp., Calgary, Canada, 2004.
l
[mm]
S1
S2
S3
S4
S5
S6
S7
S8
S9
S10
S11
S12
S1
S2
S3
S4
S5
S6
S7
S8
S9
S10
S11
S12
h
[mm]
dm
[mm]
160
160
160
160
160
160
230
230
230
300
300
300
124
124
124
124
124
124
190
190
190
260
260
260
c [mm]
fc
[MPa]
dg
[mm]
fy
[MPa]
fyw
[MPa]
system
1.53
1.53
1.53
1.53
1.53
1.53
1.35
1.35
1.35
1.15
1.15
1.15
0.45
0.45
0.68
0.68
0.68
0.32
0.21
0.35
0.23
36.2
29.0
31.6
38.0
36.3
33.4
35.0
35.0
35.2
31.4
30.0
33.5
14
14
14
14
14
14
20
20
20
20
20
20
488
488
488
488
488
488
531
531
531
524
524
524
393
393
465
465
465
460
460
409
409
b
b
b
b
b
b
b
b
b
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
VR,test
[kN]
VR,test /VR,ACI
(2011)
VR,test /VR,EC
(2004)
VR,test /VR,NAD
(2011)
VR,test /VR,SIA
(2003)
VR,test /VR,MC
(2010)
483
574
572
634
624
615
825
1050
1091
1046
1620
1520
1.30
1.27
1.18
1.66
1.21
1.20
1.12
1.21
1.20
0.88
1.22
1.00
1.11
1.19
1.12
1.21
1.21
1.22
0.93
0.97
1.05
0.77
0.98
0.88
1.11
1.42
1.37
1.43
1.43
1.45
0.93
1.13
1.17
0.77
1.09
0.99
1.20
1.54
1.49
1.32
1.31
1.41
1.07
1.36
1.41
0.90
1.42
1.28
1.23
1.22
1.13
1.23
1.23
1.19
1.11
1.22
1.27
0.94
1.28
1.16
p
o
o
o
o
o
p
o
w
p
o
o
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
w
[%]
2300
2300
2300
2300
2300
2300
3260
3260
3260
4000
4000
4000
p
o
o
o
o
o
p
o
w
p
o
w
250
250
250
250
250
250
300
300
300
350
350
350
L
[%]
p
o
o
o
o
o
p
o
o
p
o
o
p
w
w
o
o
o
p
w
w
p
w
w
p
o
o
o
o
o
p
w
w
p
w
w
VR,test
/VR,CSCT
0.98
0.97
0.89
0.98
0.98
0.94
0.89
0.87
1.00
0.75
0.90
0.90
p
o
o
o
o
o
p
w
w
p
w
w
VR,test /VR,mod
0.98
0.97
0.92
0.95
0.94
0.91
0.90
1.00
1.06
0.76
1.06
0.98
p
o
w
o
o
o
p
w
w
p
w
w
A5
Chana, P. S. and Desai, S. B., Design of shear reinforcement against punching, The Structural
Engineering, Vol. 70, pp. 159-164, 1992.
Chana, P. S., A prefrabricated shear reinforcement system for flat slabs, Proceedings of the
Institution of Civil Engineers: Structures and Buildings, Vol. 99, pp. 345-358, 1993.
l
[mm]
S1
S2
S3
S4
S5
S6
S7
S8
S1
S2
S3
S4
3000
3000
3000
3000
3000
3000
3000
3000
3000
3000
3000
3000
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
h
[mm]
dm
[mm]
240
240
240
240
250
250
250
250
228
228
228
228
200
200
200
200
210
210
210
210
188
188
188
188
c [mm]
300
300
300
300
400
400
400
400
300
300
300
300
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
L
[%]
w
[%]
fc
[MPa]
dg
[mm]
fy
[MPa]
fyw
[MPa]
system
0.79
0.79
0.79
0.79
0.86
0.86
0.86
0.86
0.83
0.83
0.83
0.83
0.13
0.13
0.26
0.16
0.26
0.26
0.10
0.50
0.50
0.50
31.8
35.1
32.5
35.9
30.3
34.3
31.9
31.4
34.0
32.5
34.8
33.6
20
20
20
20
20
20
20
20
20
20
20
20
5001)
5001)
5001)
5001)
5001)
5001)
5001)
5001)
5001)
5001)
5001)
5001)
520
520
520
520
520
520
520
520
520
520
f
f
f
f
f
f
f
g
g
f
1)
The steel for flexural reinforcement was not tested. Values were estimated based on the
indicated steel quality.
S1
S2
S3
S4
S5
S6
S7
S8
S1
S2
S3
S4
A6
VR,test
[kN]
VR,test /VR,ACI
(2011)
VR,test /VR,EC
(2004)
VR,test /VR,NAD
(2011)
VR,test /VR,SIA
(2003)
VR,test /VR,MC
(2010)
805
1094
1139
1302
1382
1283
1492
1324
851
981
1286
1135
1.07
1.28
1.36
1.30
1.37
1.20
1.44
1.38
1.19
1.41
1.40
1.26
1.03
1.07
1.13
0.94
1.02
0.85
1.01
1.17
1.16
1.07
1.01
0.90
1.03
1.07
1.13
1.14
1.08
0.97
1.15
1.17
1.16
1.30
1.24
1.11
1.14
1.51
1.61
1.78
1.59
1.37
1.17
1.51
1.28
1.12
1.47
1.30
1.19
1.24
1.31
1.19
1.22
1.01
1.20
1.29
1.34
1.21
1.27
1.13
p
w
w
o
o
o
o
w
p
o
o
o
p
w
w
o
w
o
o
w
p
o
o
o
p
w
w
c
o
o
o
w
p
o
c
c
p
w
w
w
w
w
c
w
p
w
w
w
p
w
w
w
w
o
o
w
p
o
o
o
VR,test
/VR,CSCT
0.88
0.96
1.02
0.98
0.94
0.77
0.91
0.97
0.99
0.92
1.03
0.92
p
w
w
w
w
o
o
w
p
o
o
o
VR,test /VR,mod
0.89
1.05
1.11
1.06
1.04
0.84
0.98
1.08
1.00
0.93
1.11
0.99
p
w
w
w
w
o
o
w
p
o
w
w
Elstner, R. C. and Hognestad, E., Shearing Strength of Reinforced Concrete Slabs, ACI
Materials Journal, Vol. 53-2, pp. 29-58, 1956.
l
[mm]
B12
B12
h
[mm]
dm
[mm]
152.4
114
c [mm]
fc
[MPa]
dg
[mm]
fy
[MPa]
fyw
[MPa]
system
3.00
0.73
44.8
38.1
331.6
323
(sq)
VR,test
[kN]
VR,test /VR,ACI
(2011)
VR,test /VR,EC
(2004)
VR,test /VR,NAD
(2011)
VR,test /VR,SIA
(2003)
VR,test /VR,MC
(2010)
787
2.10
1.22
1.56
1.64
1.14
(sq)
w
[%]
1828
254
L
[%]
VR,test
/VR,CSCT
0.96
VR,test /VR,mod
o
1.08
Etter, S., Heinzmann, D., Jger, T., and Marti, P., Versuche zum Durchstanzverhalten von
Stahlbetonplatten, 64 pp., Zurich, Switzerland, 2009.
l
[mm]
SP1
SP2
SP3
4100
4100
4100
(sq)
(sq)
(sq)
dm
[mm]
350
350
350
294
294
294
1710
3080
3350
1.34
2.37
1.36
c [mm]
400
400
400
VR,test
/VR,EC
(2004)
VR,test
/VR,ACI
(2011)
VR,test
[kN]
SP1
SP2
SP3
h
[mm]
p
o
c
1.01
1.63
1.04
(ci)
(ci)
(ci)
L
[%]
w
[%]
fc
[MPa]
dg
[mm]
fy
[MPa]
fyw
[MPa]
system
1.26
1.26
1.26
0.79
0.79
35.5
34.5
33.4
32
32
32
601
601
601
529
529
529
a
a
p
o
c
1.01
1.83
1.44 (1.06)
VR,test
/VR,MC
(2010)
VR,test
/VR,SIA
(2003)
VR,test /VR,NAD
(2011)
p
o
c
1.08
1.30
1.43
p
c
c
1.15
1.47
1.29
VR,test
/VR,mod
VR,test
/VR,CSCT
p
o
c
0.92
1.11
0.95
p
o
c
0.92
1.15
1.00
p
o
c
Fernndez-Ruiz, M., Muttoni, A., and Kunz, J., Strengthening of flat slabs against punching
shear using post-installed shear reinforcement, ACI Structural Journal, Vol. 107, pp. 434-442,
2007.
l
[mm]
PV1
PV1
h
[mm]
dm
[mm]
250
210
c [mm]
w
[%]
fc
[MPa]
dg
[mm]
fy
[MPa]
fyw
[MPa]
system
1.50
34.0
16
709
3000
(sq)
VR,test
[kN]
VR,test /VR,ACI
(2011)
VR,test /VR,EC
(2004)
VR,test /VR,NAD
(2011)
VR,test /VR,SIA
(2003)
VR,test /VR,MC
(2010)
974
1.27
0.96
0.96
1.12
1.16
260
(sq)
L
[%]
VR,test
/VR,CSCT
1.01
VR,test /VR,mod
p
1.02
A7
Gomes, R. B. and Regan, P. E., Punching Resistance of RC Flat Slabs with Shear
Reinforcement, ASCE Journal of Structural Engineering, Vol. 125, Issue 6, pp. 684-692, 1999.
l
[mm]
S1
S1A
S2
S3
S4
S5
S6
S7
S8
S9
S10
S11
S1
S1A
S2
S3
S4
S5
S6
S7
S8
S9
S10
S11
A8
h
[mm]
dm
[mm]
200
200
200
200
200
200
200
200
200
200
200
200
159
159
153
158
159
159
159
159
159
159
154
154
c [mm]
fc
[MPa]
dg
[mm]
fy
[MPa]
fyw
[MPa]
system
1.26
1.26
1.31
1.27
1.26
1.26
1.26
1.26
1.26
1.26
1.31
1.31
0.15
0.20
0.26
0.40
0.40
0.58
0.58
0.58
0.15
0.20
41.7
42.7
35.8
40.7
33.3
36.0
38.8
35.1
35.4
41.5
36.7
35.9
20
20
20
20
20
20
20
20
20
20
20
20
680
680
680
670
670
670
670
670
670
670
670
670
430
430
430
430
430
430
430
430
430
430
c
c
c
c
c
c
c
c
c
c
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
VR,test
[kN]
VR,test /VR,ACI
(2011)
VR,test /VR,EC
(2004)
VR,test /VR,NAD
(2011)
VR,test /VR,SIA
(2003)
VR,test /VR,MC
(2010)
560
587
693
773
853
853
1040
1120
1200
1227
800
907
1.14
1.18
1.61
1.61
1.94
1.75
1.55
1.99
1.87
1.28
1.49
1.66
0.93
0.97
1.25
1.19
1.26
1.22
1.23
1.37
1.46
1.08
1.27
1.29
0.93
0.97
1.28
1.31
1.45
1.41
1.41
1.56
1.67
1.46
1.27
1.33
1.07
1.11
1.47
1.51
1.77
1.72
2.04
2.17
2.33
2.35
1.67
1.91
1.10
1.15
1.37
1.36
1.30
1.22
1.38
1.41
1.43
1.37
1.56
1.71
p
p
w
w
o
o
o
o
o
o
w
w
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
w
[%]
3000
3000
3000
3000
3000
3000
3000
3000
3000
3000
3000
3000
p
p
o
o
o
o
o
o
o
o
w
o
200
200
200
200
200
200
200
200
200
200
200
200
L
[%]
p
p
o
o
o
o
o
o
o
c
w
o
p
p
w
w
w
w
w
w
w
w
w
w
p
p
w
w
o
o
o
o
o
c
w
w
VR,test
/VR,CSCT
0.91
0.94
1.10
1.08
1.20
1.03
1.24
1.21
1.30
1.29
1.25
1.37
p
p
w
w
w
w
w
w
w
w
w
w
VR,test /VR,mod
0.91
0.94
1.15
1.13
1.14
1.03
1.17
1.20
1.23
1.13
1.26
1.33
p
p
w
w
w
o
o
o
o
w
w
w
Hegger, J., Husler, F., and Ricker , M., Zur maximalen Durchstanztragfhigkeit von
Flachdecken, Beton - und Stahlbetonbau, Vol. 102, pp. 770-777, 2007.
l
[mm]
EM1
EM1
h
[mm]
dm
[mm]
200
160
c [mm]
fc
[MPa]
dg
[mm]
fy
[MPa]
fyw
[MPa]
system
1.96
1.21
29.8
16
558
584
(sq)
VR,test
[kN]
VR,test /VR,ACI
(2011)
VR,test /VR,EC
(2004)
VR,test /VR,NAD
(2011)
VR,test /VR,SIA
(2003)
VR,test /VR,MC
(2010)
1213
1.54
0.83
1.27
1.26
1.18
(sq)
w
[%]
2800
290
L
[%]
VR,test
/VR,CSCT
0.90
VR,test /VR,mod
c
0.91
Ladner, M., Untersuchungsbericht Durchstanzversuche an Flachdeckenausschnitten, 38 pp.,

Luzern, 1998.
l
[mm]
V1
V2
V4
V1
V2
V4
h
[mm]
dm
[mm]
280
280
280
240
240
240
c [mm]
fc
[MPa]
dg
[mm]
fy
[MPa]
fyw
[MPa]
system
1.40
1.40
1.40
0.79
0.79
25.4
30.1
27.4
32
32
32
510
510
510
646
646
g
g
(ci)
(ci)
(ci)
VR,test
[kN]
VR,test /VR,ACI
(2011)
VR,test /VR,EC
(2004)
VR,test /VR,NAD
(2011)
VR,test /VR,SIA
(2003)
VR,test /VR,MC
(2010)
1095
1784
1799
1.60
1.60
1.69
1.02
0.99
1.09
1.03
1.13
1.17
1.12
1.10
1.15
1.19
1.06
1.11
p
c
c
(ci)
(ci)
(ci)
w
[%]
3300
3300
3300
p
c
c
300
300
300
L
[%]
p
c
c
p
c
c
p
c
c
VR,test
/VR,CSCT
0.98
0.83
0.86
VR,test /VR,mod
p
c
c
0.99
1.07
1.13
p
c
c
A9
Lips, S., Muttoni, A., and Fernndez-Ruiz , M., Durchstanzversuche an Deckenausschnitten

mit FIDECA Durchstanzbewehrung, EPFL-IBton, 55 pp., Lausanne, Switzerland, 2010.
Lips, S., Muttoni, A., and Fernndez-Ruiz , M., Durchstanzversuche an Deckenausschnitten
mit Durchstanzbewehrung, EPFL-IBton, 88 pp., Lausanne, Switzerland, 2012.
Lips, S., Muttoni, A., and Fernndez-Ruiz , M., Experimental Investigation on the Punching
Strength and the Deformation Capacity of Shear-Reinforced Slabs, ACI Structural Journal,
2012.
l
[mm]
PL1
PL3
PL4
PL5
PL6
PL7
PL8
PL9
PL10
PL11
PL12
PF1
PF2
PF3
PF4
PF5
3000
3000
3000
3000
3000
3000
3000
3000
3000
3000
3000
3000
3000
3000
3000
3000
VR,test
[kN]
PL1
PL3
PL4
PL5
PL6
PL7
PL8
PL9
PL10
PL11
PL12
PF1
PF2
PF3
PF4
PF5
A10
1135
1324
1625
2491
1363
1773
2256
3132
5193
1176
1633
1043
1567
2242
2667
4717
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
h
[mm]
dm
[mm]
250
250
320
400
250
250
250
320
400
250
250
250
250
250
320
400
193
197
267
353
198
197
200
266
343
201
201
209
208
209
274
354
130
520
340
440
130
260
520
340
440
260
260
130
260
520
340
440
VR,test
/VR,EC
(2004)
VR,test
/VR,ACI
(2011)
1.37
1.16
1.36
1.18
1.30
1.23
0.98
1.29
1.26
1.11
1.12
1.32
1.46
1.21
1.39
1.45
c [mm]
p
p
p
p
c
c
c
c
c
w
c
c
c
c
c
c
0.86
1.06
1.06
0.99
1.41
0.94
0.87
1.03
1.00
1.03
1.05
1.17
0.90
0.82
0.84
0.87
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
L
[%]
w
[%]
fc
[MPa]
dg
[mm]
fy
[MPa]
fyw
[MPa]
system
1.63
1.59
1.58
1.50
1.59
1.59
1.57
1.59
1.55
1.56
1.56
1.50
1.51
1.50
1.54
1.50
1.06
1.03
0.97
0.95
0.85
0.25
0.51
0.79
0.79
0.79
0.79
0.79
42.5
36.5
30.5
31.9
36.6
35.8
36.0
32.1
33.0
34.2
34.6
31.1
30.4
37.1
32.5
33.4
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
16
583
583
562
580
583
583
583
562
580
554
554
583
583
583
562
580
519
519
519
516
563
592
592
536
536
536
550
550
a
a
a
a
a
a
a
g
g
g
g
g
VR,test /VR,NAD
(2011)
p
p
p
p
c
c
f
c
c
w
w
c
c
f
c
c
0.99
1.06
1.06
0.99
1.36 (1.00)
1.32 (0.97)
1.27 (0.94)
1.44 (1.06)
1.51 (1.11)
1.03
1.20 (1.05)
1.04
1.15
1.20
1.18
1.32
p
p
p
p
c
c
c
c
c
w
c (w)
c
c
c
c
c
VR,test
/VR,MC
(2010)
VR,test
/VR,SIA
(2003)
1.12
1.22
1.19
1.05
1.38
1.47
1.41
1.50
1.49
1.46
1.75
1.03
1.28
1.32
1.22
1.29
p
p
p
p
c
c
c
c
c
w
w
c
c
c
c
c
1.15
1.27
1.23
1.09
1.19
1.27
1.24
1.30
1.71
1.32
1.49
1.16
1.20
1.25
1.19
1.42
VR,test
/VR,CSCT
p
p
p
p
w
c
c
c
w
w
w
w
c
c
w
w
1.00
1.04
1.07
0.96
0.96
1.03
0.97
1.02
1.04
1.01
1.03
0.79
0.96
0.97
0.91
0.98
p
p
p
p
c
c
c
c
w
w
w
c
c
c
c
c
VR,test
/VR,mod
1.01
1.04
1.08
0.98
0.95
0.99
1.06
0.99
0.95
1.11
1.10
0.98
1.02
0.96
0.94
1.06
p
p
p
p
c
c
f
c
c
w
w
c
w
c
w
w
Marti, P., Pralong, J., and Thrlimann, B., Schubversuche an Stahlbeton-Platten, Institut fr
Baustatik und Konstruktion, Nr. 7305-2, 123 pp., Zrich, 1977.
l
[mm]
1)
P2
P3
2750
2750
(ci)
(ci)
h
[mm]
dm
[mm]
180
180
152
152
c [mm]
300
300
(ci)
(ci)
L
[%]
w
[%]
fc
[MPa]
dg
[mm]
fy
[MPa]
fyw
[MPa]
1.40
1.40
0.00
0.35
34.3
33.5
16
16
596
596
563
system
f1)
special shear reinforcement system
P2
P3
VR,test
[kN]
VR,test /VR,ACI
(2011)
VR,test /VR,EC
(2004)
VR,test /VR,NAD
(2011)
VR,test /VR,SIA
(2003)
VR,test /VR,MC
(2010)
628
824
1.49
1.32
1.11
1.03
1.11
1.05
1.26
1.66
1.30
1.34
p
c
p
w
p
c
p
w
p
w
VR,test
/VR,CSCT
1.06
1.01
VR,test /VR,mod
p
w
1.07
1.00
p
w
Marzouk, H. and Jiang, D., Experimental Investigation on Shear Enhancement Types for
High-Strength Concrete Plates, ACI Structural Journal, Vol. 94, pp. 49-58, 1997.
l
[mm]
HS17
HS22
HS23
HS17
HS22
HS23
h
[mm]
dm
[mm]
150
150
150
120
120
120
c [mm]
fc
[MPa]
dg
[mm]
fy
[MPa]
fyw
[MPa]
system
1.09
1.09
1.09
1.47
1.47
67.0
60.0
60.0
20
20
20
490
490
490
400
400
a
a
(sq)
(sq)
(sq)
VR,test
[kN]
VR,test /VR,ACI
(2011)
VR,test /VR,EC
(2004)
VR,test /VR,NAD
(2011)
VR,test /VR,SIA
(2003)
VR,test /VR,MC
(2010)
511
605
590
1.05
1.32
1.21
1.13
1.17
1.13
1.13
1.38
1.35
1.16
1.15
1.12
1.19
1.19
1.16
p
o
o
(sq)
(sq)
(sq)
w
[%]
1950
1950
1950
p
o
o
250
250
250
L
[%]
p
o
o
p
o
o
p
o
o
VR,test
/VR,CSCT
0.96
1.02
1.00
VR,test /VR,mod
p
o
o
0.96
1.14
1.11
p
o
o
Mueller, F. X., Muttoni, A., and Thrlimann, B., Durchstanzversuche an Flachdecken mit
Aussparungen, IBK - Bericht (ETHZ), Institut fr Baustatik und Konstruktion der ETH Zrich,
Birkhuser Verlag, 7305-5, 118 pp., Zrich, 1984.
l
[mm]
P22
P22
h
[mm]
dm
[mm]
185
153.5
c [mm]
fc
[MPa]
dg
[mm]
fy
[MPa]
fyw
[MPa]
system
1.31
0.89
32.9
16
575
565
(ci)
VR,test
[kN]
VR,test /VR,ACI
(2011)
VR,test /VR,EC
(2004)
VR,test /VR,NAD
(2011)
VR,test /VR,SIA
(2003)
VR,test /VR,MC
(2010)
1044
1.33
0.86
1.34
1.43
1.25
(ci)
w
[%]
2824
300
L
[%]
VR,test
/VR,CSCT
1.00
VR,test /VR,mod
c
0.99
A11
Regan, P. E. and Samadian, F., Shear Reinforcement against punching in reinforced concrete
flat slabs, The Structural Engineer, pp. 24-31, 2001.
l
[mm]
1
1A
R1
R2
R3
R4
A1
A2
S1
S2
1
1A
R1
R2
R3
R4
A1
A2
S1
S2
h
[mm]
dm
[mm]
200
200
200
200
200
200
200
200
200
200
164
164
164
164
164
164
160
160
160
160
c [mm]
fc
[MPa]
dg
[mm]
fy
[MPa]
fyw
[MPa]
system
1.26
1.26
1.26
1.26
1.26
1.26
1.64
1.64
1.64
1.64
0.57
0.49
0.49
0.57
0.40
0.34
0.40
0.70
41.7
42.7
35.2
39.0
34.7
40.8
38.8
44.7
41.3
45.7
20
20
20
20
20
20
20
20
20
20
670
670
670
670
670
670
670
670
540
540
442
442
442
442
519
519
635
350
a
a
a
a
a
a
f
f
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
VR,test
[kN]
VR,test /VR,ACI
(2011)
VR,test /VR,EC
(2004)
VR,test /VR,NAD
(2011)
VR,test /VR,SIA
(2003)
VR,test /VR,MC
(2010)
560
587
1050
950
850
950
1000
950
900
950
1.09
1.13
1.56
1.42
1.35
1.31
1.45
1.37
1.35
1.35
0.88
0.92
1.20
1.05
0.98
1.04
1.07
1.02
0.99
0.80
0.88
0.92
1.44
1.26
1.17
1.24
1.28
1.16
0.99
0.99
1.01
1.05
1.92
1.76
1.64
1.73
1.78
1.61
1.59
1.62
1.04
1.09
1.21
1.21
1.11
1.04
1.05
1.25
1.12
1.11
p
p
o
o
o
o
o
w
w
o
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
w
[%]
3000
3000
3000
3000
3000
3000
3000
3000
3000
3000
p
p
o
o
o
o
o
o
o
o
200
200
200
200
200
200
200
200
200
200
L
[%]
p
p
o
o
o
o
o
o
w
c
p
p
w
w
w
w
w
w
w
w
p
p
o
w
w
o
o
w
w
w
VR,test
/VR,CSCT
0.85
0.89
1.06
0.99
0.91
0.93
1.03
0.99
0.88
0.92
VR,test /VR,mod
p
p
w
w
w
w
w
w
w
w
0.86
0.89
1.10
1.05
0.96
0.97
1.08
1.06
0.89
0.93
p
p
w
w
w
w
w
w
w
w
Rizk, E., Marzouk, H., and Hussein, A., Punching Shear of Thick Plates with and without
Shear Reinforcement, ACI Structural Journal, Vol. 108, No. 5, pp. 581-591, 2011.
l
[mm]
HSS2
HSS2
A12
h
[mm]
dm
[mm]
300
217.5
c [mm]
fc
[MPa]
dg
[mm]
fy
[MPa]
fyw
[MPa]
system
1.42
0.41
79.0
19
460
400
(sq)
VR,test
[kN]
VR,test /VR,ACI
(2011)
VR,test /VR,EC
(2004)
VR,test /VR,NAD
(2011)
VR,test /VR,SIA
(2003)
VR,test /VR,MC
(2010)
2172
1.22
1.09
1.36
1.53
1.34
(sq)
w
[%]
2650
400
L
[%]
VR,test
/VR,CSCT
1.05
VR,test /VR,mod
o
1.05
Seible, F., Ghali, A., and Dilger, W. H., Preassembled shear reinforcing units for flat plates,
ACI Journal, Proceedings 77, No. 1, pp. 28-35, 1980.
l
[mm]
SC7
SC8
SC9
SC10
SC11
SC12
SC13
SC7
SC8
SC9
SC10
SC11
SC12
SC13
h
[mm]
dm
[mm]
150
150
150
150
150
150
150
121
121
121
121
121
121
121
c [mm]
fc
[MPa]
dg
[mm]
fy
[MPa]
fyw
[MPa]
system
1.17
1.17
1.17
1.17
1.17
1.17
1.17
0.38
0.42
0.28
0.31
0.15
0.33
0.73
33.6
33.6
33.6
33.6
33.6
33.6
33.6
13
13
13
13
13
13
13
450
450
450
450
450
450
450
350
490
490
490
500
500
500
c
g
g
g
b
b
b
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
VR,test
[kN]
VR,test /VR,ACI
(2011)
VR,test /VR,EC
(2004)
VR,test /VR,NAD
(2011)
VR,test /VR,SIA
(2003)
VR,test /VR,MC
(2010)
623
592
594
537
596
595
580
1.56
1.49
1.49
1.35
1.50
1.49
1.46
1.10
1.04
1.05
0.95
1.34
1.05
1.02
1.37
1.30
1.30
1.18
1.34
1.30
1.27
1.69
1.48
1.61
1.45
1.61
1.61
1.12
1.32
1.26
1.26
1.14
1.32
1.31
1.23
o
o
o
o
w
o
o
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
w
[%]
1800
1800
1800
1800
1800
1800
1800
o
o
o
o
o
o
o
305
305
305
305
305
305
305
L
[%]
o
o
o
o
w
o
o
w
w
w
w
w
w
o
o
o
o
o
w
w
o
VR,test
/VR,CSCT
1.11
1.06
1.06
0.96
1.12
1.06
1.04
VR,test /VR,mod
o
o
o
o
w
o
o
1.36
1.28
1.17
0.99
1.33
1.11
1.08
o
o
o
o
w
o
o
Stein , T., Ghali , A., and Dilger , W. H., Distinction between Punching and Flexural Failure
Modes of Flat Plates, ACI Structural Journal, Vol. 104, pp. 357-365, 2007.
l
[mm]
V2
V3
V2
V3
h
[mm]
dm
[mm]
150
150
118
118
c [mm]
fc
[MPa]
dg
[mm]
fy
[MPa]
fyw
[MPa]
system
0.98
0.62
0.62
0.62
26.2
25.7
14
14
438
438
457
457
a
a
(sq)
(sq)
VR,test
[kN]
VR,test /VR,ACI
(2011)
VR,test /VR,EC
(2004)
VR,test /VR,NAD
(2011)
VR,test /VR,SIA
(2003)
VR,test /VR,MC
(2010)
438
365
1.39
1.17
1.18
1.16
1.41
1.38
1.19
1.22
1.25
1.28
o
o
(sq)
(sq)
w
[%]
1900
1900
o
o
250
250
L
[%]
o
o
o
o
o
o
VR,test
/VR,CSCT
1.00
1.16
VR,test /VR,mod
o
o
1.00
1.24
o
o
A13
Tolf, P., Influence of the slab thickness on the strength of concrete slabs at punching. Tests with
circular slabs. (In Swedish: Plattjocklekens inverkan p betongplattors hllfasthet vid
genomstansning. Frsk med cikulra plattor.), Royal Institute of Technology, Dep. of
Structural Mechanics and Engineering, Bulletin 146, 64 pp., 1988.
l
[mm]
S1.1
S1.2
S1.3
S1.4
S1.1s
S1.2s
S1.3s
S1.4s
S2.1
S2.2
S2.3
S2.4
S2.1s
S2.2s
S2.3s
S2.4s
S1.1
S1.2
S1.3
S1.4
S1.1s
S1.2s
S1.3s
S1.4s
S2.1
S2.2
S2.3
S2.4
S2.1s
S2.2s
S2.3s
S2.4s
A14
h
[mm]
dm
[mm]
120
120
120
120
120
120
120
120
240
240
240
240
240
240
240
240
100
99
98
99
97
99
100
99
200
199
200
197
c [mm]
125
125
125
125
125
125
125
125
250
250
250
250
250
250
250
250
w
[%]
fc
[MPa]
dg
[mm]
fy
[MPa]
fyw
[MPa]
system
0.80
0.81
0.35
0.34
0.82
0.81
0.34
0.34
0.80
0.80
0.34
0.35
0.55
0.54
0.54
0.54
-
30.1
23.8
26.0
25.5
24.1
24.8
23.4
24.0
19.4
18.9
19.3
21.6
16
16
16
16
16
16
16
16
32
32
32
32
706
701
720
712
706
711
709
710
657
670
668
664
610
630
620
620
-
f
f
f
f
-
0.82
0.82
0.34
0.34
0.27
0.27
0.27
0.27
19.3
19.7
21.8
21.7
32
32
32
32
669
673
671
669
700
640
670
670
f
f
f
f
1270
1270
1270
1270
1270
1270
1270
1270
2540
2540
2540
2540
2540
2540
2540
2540
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
VR,test
[kN]
VR,test /VR,ACI
(2011)
VR,test /VR,EC
(2004)
VR,test /VR,NAD
(2011)
VR,test /VR,SIA
(2003)
VR,test /VR,MC
(2010)
216
194
145
148
261
259
144
147
603
600
489
444
1.67
1.71
1.24
1.26
2.27
2.16
1.22
1.25
1.45
1.48
1.18
1.04
p
p
p
p
o
o
o
o
p
p
p
p
1.26
1.24
1.21
1.23
1.34
1.29
0.96
0.99
1.02
1.03
1.10
0.98
p
p
p
p
o
o
o
o
p
p
p
p
1.27
1.24
1.21
1.24
1.34
1.29
0.97
0.99
1.03
1.04
1.11
0.98
p
p
p
p
o
o
o
o
p
p
p
p
1.22
1.22
1.19
1.21
1.16
1.09
0.84
0.87
1.08
1.09
1.17
1.04
p
p
p
p
w
w
f
f
p
p
p
p
1.26
1.25
1.23
1.26
1.09
1.04
0.84
0.87
1.14
1.16
1.26
1.13
p
p
p
p
o
o
f
f
p
p
p
p
894
851
562
542
2.15
2.03
1.25
1.21
o
o
o
o
1.23
1.16
0.97
0.94
o
o
o
o
1.23
1.16
0.98
0.94
o
o
o
o
1.66
1.56
1.11
1.07
w
w
w
w
1.36
1.28
0.97
0.94
w
w
o
o
195
195
198
198
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
(ci)
L
[%]
VR,test
/VR,CSCT
1.02
1.03
0.97
1.00
0.89
0.85
0.84
0.87
0.91
0.93
0.95
0.84
1.03
0.97
0.88
0.86
p
p
p
p
o
o
c
c
p
p
p
p
w
w
c
c
VR,test /VR,mod
1.03
1.04
0.99
1.01
0.98
0.94
0.96
0.99
0.92
0.94
0.96
0.86
0.98
0.93
1.00
0.97
p
p
p
p
o
o
o
o
p
p
p
p
w
w
o
o
Vollum, R. L., Abdel-Fattah, T., Eder, M., and Elghazouli, A.Y. , Design of ACI-type
punching shear reinforcement to Eurocode 2, Magazine of Concrete Research, 62-01, pp. 3-16,
2010.
l
[mm]
S1
S2
S3
S4
S5
S6
S1
S2
S3
S4
S5
S6
h
[mm]
dm
[mm]
220
220
220
220
220
220
174
174
174
174
174
174
c [mm]
fc
[MPa]
dg
[mm]
fy
[MPa]
fyw
[MPa]
system
1.28
1.28
1.28
1.28
1.28
0.64
0.29
0.29
0.18
0.37
0.37
24.0
24.0
27.2
27.2
23.2
23.2
20
20
20
20
20
20
567
567
567
567
567
567
560
560
485
485
485
f
f
f
f
f
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
VR,test
[kN]
VR,test /VR,ACI
(2011)
VR,test /VR,EC
(2004)
VR,test /VR,NAD
(2011)
VR,test /VR,SIA
(2003)
VR,test /VR,MC
(2010)
614
843
903
906
872
740
1.22
1.11
1.33
1.33
1.17
1.09
0.96
1.18
1.22
1.22
1.24
1.32
0.96
1.32
1.35
1.36
1.38
1.47
1.11
1.52
1.56
1.56
1.59
1.47
1.15
1.22
1.25
1.35
1.27
1.43
p
o
o
o
o
o
(sq)
(sq)
(sq)
(sq)
(sq)
(sq)
w
[%]
3000
3000
3000
3000
3000
3000
p
c
o
w
c
f
270
270
270
270
270
270
L
[%]
p
o
o
o
o
o
p
w
w
w
w
w
p
o
o
w
o
o
VR,test
/VR,CSCT
0.96
1.00
1.03
1.08
1.05
1.21
p
o
o
w
o
o
VR,test /VR,mod
0.97
1.01
1.04
1.15
1.05
1.24
p
o
o
w
o
o
A15
Appendix B Example PL7

Given slab parameters
h=
d=
l=
c=
rq=
b=
b1=
250 mm
197 mm
3000 mm
260 mm
1505 mm
1200 mm
120 mm
=
dw=
Asw1=
s0=
s1=
lai=
las=
ncut=
1.59%
14 mm
2463 mm2
80 mm
160 mm
104 mm
106 mm
1
fc=
E c=
fct=
dg=
fy=
Es=
fyw=
Esw=
35.8 MPa
33000 MPa
3.26 MPa
16 mm
583 MPa
200000 MPa
519 MPa
200000 MPa
Constant parameters
2
165.5 mm
4
2
2
2
max
Eq. 8.4
1484 mm
405.5 mm
1.5 ;
Eq. 8.7
245.5 mm
0.5
504 mm
Eq. 7.69
0.75
2
1659 mm
2
Eq. 7.75
0.81
0.5
tan
80 mm
,
0.10 mm
Eq. 7.51
2
0.10 mm
Eq. 7.55
0.75
B1
0.35
Moment-curvature relationship
12
4.29 10
Eq. 7.7
62.1 mm
Eq. 7.10
1.12 10
Nmm
Eq. 7.9
33.9 kNm/m
Nmm
Eq. 7.6
314 kNm/m
7.91 10
1
6
Eq. 7.11
mm
9.08 10
Eq. 7.8
mm
Eq. 7.12
2.12 10
mm
Eq. 7.13
2.71 10
mm
Eq. 7.14
-m
-y
t
-mR
-mR
-p -cr
TS
mt
EI1
-mcr
-mcr
EI0
ry
r2
rp
rcr
rs
-cr
-p
-y
Parameters for quadratic rotation function

2
B2
2.99 10
mm
4.41 10
Eq. 7.3
mm
Eq. 7.4
This example shows the calculation for a rotation of 0=25. In order to obtain the punching
strength, the rotation needs to be adjusted until the shear force of the calculated load-rotation
response V corresponds to the punching strength VR.
V
2
-mt,ext
-mt,int
-mr
rq rs
r2
rc
V
2
In a first step the tangential moments at the outer perimeter can be calculated:
@
4.96 10
mm
Eq. 7.5
922 mm
Eq. 7.30
1484 mm
Eq. 7.31
1484 mm
Eq. 7.32
Eq. 7.27
As the tangential moment within the area between r2 and r0 is equal to the flexural
strength, the integration of the tangential moment becomes:
30.9 kNm
Eq. 7.33
Since rp and rcr is equal to rs, the general equation

ln
ln
Eq. 7.28
can be simplified to
B3
ln
270 kNm
Thus, the integral of the tangential moment at the outer slab segment becomes
301 kNm
Inner slab element

,
1.03 10
mm
1.03 10
Eq. 7.68
mm
227 MPa
Eq. 7.15
Eq. 7.16
1820 kN
Eq. 7.17
Since the radial and the tangential moment need to be found by iteration, a start value of them
need to be chosen. However, for the first step, it is easier to estimate the height of the
compression zone. Thus, the assumption of
62.1 mm
and
20.7 mm
leads to the moment
369 kNm/m
Eq. 7.22
and
138 kNm/m
Eq. 7.23
With this, a first approximation of the shear force is obtained.
B4
1855 kN
Eq. 7.26
Failure criterion: failure within shear reinforced area

3
15
443 kN
Eq. 7.42
0.70 mm
1278 kN
Eq. 7.46
519 MPa
Eq. 7.56
Eq. 7.45
2017 kN
Eq. 7.44
Failure criterion: crushing of the concrete strut

2
41.4 mm
1
1.0
0.436
135
1931 kN
Eq. 7.61
1209 kN
Eq. 7.63
Eq. 7.44
B5
From the previously calculated values, one can determine a new set of parameters:
0.626
Eq. 7.71
560 kN
Eq. 7.21
2
1
cot
Eq. 7.19
4.7 mm
Eq. 7.25
57.1 mm
B6
0.437
135
Eq. 7.23
1718 kN
42.0 mm
1.0
Eq. 7.22
33.0 kNm/m
Eq. 7.24
344 kNm/m
Eq. 7.18
431 kN
1389 kN
1935 kN
1211 kN
Eq. 7.61
Eq. 7.63
Eq. 7.44
Calculation of new set of parameters
0.626
Eq. 7.71
511 kN
Eq. 7.21
2
1
cot
Eq. 7.19
5.87 mm
Eq. 7.25
54.1 mm
0.434
135
Eq. 7.23
1715kN
40.4 mm
1.0
Eq. 7.22
40.8 kNm/m
Eq. 7.24
329 kNm/m
Eq. 7.18
536 kN
1284 kN
1926 kN
1204 kN
Eq. 7.61
Eq. 7.63
Eq. 7.44
B7
Already after two iterations, the difference in results between the iteration steps is small. Thus,
the rotation can be changed to calculate the next points of the load-rotation curve and the failure
criterion curves.
3000
Calculated
Measured
VR,II
2500
V [kN]
2000
VR,I
1500
1000
500
0
0
B8
10
20
30
[]
40
50
Appendix C Shear Reinforcement Ratio

Within this thesis the shear reinforcement ratio is for regularly distributed shear reinforcement
defined by the cross sectional area of shear reinforcement within a unit area or for radially or
orthogonally arranged shear reinforcement calculated at a control perimeter set at d/2 of the
border of the support region where d is the effective depth. In the latter case, the shear
reinforcement is given by:
(C.1)
where b0 is the control perimeter set at d/2 of the border of the support region, Asw,i is the cross
sectional area of one vertical branch of the shear reinforcement, nr is the number of vertical
branches of shear reinforcement in the first perimeter and s is defined by:
2
(C.2)
Where s0 is the distance between the border of the support region and the first perimeter of shear
reinforcement and s1 is the distance between the first and the second perimeter of shear
reinforcement.
Figure C.1: Definitions of parameters used for the calculation of the shear reinforcement ratio
C1
Curriculum Vitae
Personal
Name
First name
Date of birth
Birthplace
Nationality
E-Mail
Lips
Stefan
December 4, 1980
Zurich, Switzerland
Swiss
stlips@bluewin.ch
Education
MS in Civil Engineering, Georgia Institute of Technology, Atlanta GA, USA, 2007
BS in Civil Engineering, University of Applied Sciences Rapperswil, Switzerland, 2004
Vocational Degree as a Structural Drafter, Zurich, Switzerland, 2000
Work Experience
Research Assistant, Ecole Polytechnique Fdrale de Lausanne, Switzerland, 2008-2012
Project Engineer, University of Applied Sciences Rapperswil, Switzerland, 2004-2006
Structural Drafter Apprentice, WEWO AG Zurich, Switzerland, 1996-2000
Languages
German, native language
English, fluent
French, intermediate
Professional Affiliations
Swiss Engineering, STV
American Concrete Institute, ACI
Fdration internationale du bton fib
Gesellschaft fr Ingenieurbaukunst
Selected Publications
Lips, S., Muttoni, A., and Fernndez-Ruiz, M., Experimental Investigation on the Punching
Strength and the Deformation Capacity of Shear-Reinforced Slabs, ACI Structural Journal, USA,
(Accepted for publication)
Tassinari L., Lips S., Muttoni A., Fernndez Ruiz M., Applications of bent-up bars as shear and
integrity reinforcement in R/C slabs, Proceedings of the fib Symposium Prague 2011, pp. 631-634,
Prague, Czech Republic, 2011.
Lips S., Muttoni A., Fernndez Ruiz M., NMC2010 - Punching shear of flat slabs: Design example,
Ecole Polytechnique Fdrale de Lausanne, 13 pp., Lausanne, Switzerland, 2010.
Lips S., Muttoni A., Experimental investigation of reinforced concrete slabs with punching shear
reinforcement, Proceedings of the 8th fib-PhD Symposium, pp. 105-110, Copenhagen, Denmark,
2010.

FIB Bulletin - Bond of Reinforcement in Concrete

Transféré par

Droits d'auteur :

Formats disponibles

FIB Bulletin - Bond of Reinforcement in Concrete

Transféré par

Informations du document

Description originale:

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

FIB Bulletin - Bond of Reinforcement in Concrete

Transféré par

Droits d'auteur :

Formats disponibles

Punching of Flat Slabs with Large Amounts of Shear

THSE NO 5409 (2012)

COLE POLYTECHNIQUE FDRALE DE LAUSANNE

accepte sur proposition du jury:

In loving memory of my Dad

Prof. Dr. Aurelio Muttoni

Keywords: punching shear, shear reinforcement, flat slabs, flexural response

Mots-cls: poinonnement, armature de poinonnement, planchers-dalles, comportement la

Research significance .................................................................................................. 3

Personal contributions ................................................................................................. 6

Literature Review ................................................................................................................ 7

Overview of previous research .................................................................................... 8

ACI 318-11 (ACI 318 2011)...................................................................................... 11

Eurocode 2 (EC2 2004) ............................................................................................. 13

German National Annex (NAD 2011) ....................................................................... 16

SIA 262 (SIA 262 2003) ............................................................................................ 18

Model Code (MC 2011)............................................................................................. 21

2.6.1. Slabs without shear reinforcement ................................................................21

Critical shear crack theory (CSCT)............................................................................25

Load-rotation response ..............................................................................................34

Geometry and reinforcement .....................................................................................44

Test set-up ..................................................................................................................47

Discussion of the results ............................................................................................52

Validation of Code Provisions and the CSCT .................................................................63

ACI 318-11 (ACI 318 2011) ......................................................................................64

Eurocode 2 (EC2 2004) .............................................................................................66

4.2.3. Column size .................................................................................................. 66

German National Annex to Eurocode 2 (NAD 2011) ............................................... 68

SIA 262 (SIA 262 2003) ............................................................................................ 70

fib Model code (MC 2011) ........................................................................................ 72

Critical shear crack theory (CSCT) ........................................................................... 74

Overview of the performance of the codes ................................................................ 79

Development of a Nonlinear Finite Element Approach.................................................. 89

5.1.3. Equilibrium conditions..................................................................................96

Shear stiffness ............................................................................................................99

Analysis of the Slab Response .........................................................................................113

Analysis procedure ..................................................................................................114

Global slab behavior ................................................................................................117

Local slab behavior ..................................................................................................126

Development of an Analytical Model .............................................................................131

General slab behavior ..............................................................................................132

Load-rotation response ............................................................................................137

Failure criteria ..........................................................................................................148

Definition of model parameters ...............................................................................158

Validation of the Analytical Model .................................................................................169

Transformation of the specimen ..............................................................................170

Other failure modes considered ...............................................................................186

Validation with tests within this research ................................................................ 187

Comparison to tests from literature ......................................................................... 195

Conclusions and Future Research .................................................................................. 205

Conclusions ............................................................................................................. 206

Recommendations for future research ..................................................................... 210

Bibliography ............................................................................................................................. 211

Critical shear crack theory

Eurocode 2 (refer to reference EC2 2004)

fib Model Code 2010 (refer to reference MC 2011)

German National Annex to Eurocode (refer to reference NAD 2011)

Nonlinear finite element analysis

Swiss code 262 (refer to reference SIA 262 2003)