386

Flat slab`column connections

W H Dilger
University of Calgary, Canada

Summary mode of slabs without shear reinforcement to

Flat slabs are common structural elements for a ductile failure for slabs with shear reinforcement.
apartments, office and institutional buildings. Their This is followed by a discussion of the different
weak point is the resistance against punching in the theoretical approaches to predict the resistance
vicinity of the supporting columns. In this article the of slab`column connections, and an assessment
experimental evidence and the parameters that of the major design codesdACI and EC2. The
influence the shear resistance of two-way slabs are article includes a discussion of some recent
presented first. Slabs without and with shear work on the shear resistance and ductility of
reinforcement are discussed. Special emphasis is slab`column connections under reversed cyclic
put on the experimental evidence how shear conditions and recommendations for seismic
reinforcement transforms the brittle failure design.

Prog. Struct. Engng Mater. 2000; 2: 386d399

Abbreviations K "size factor (EC2),

k "correlation coefficient determined from experimentation (shear friction)
ACI "American Concrete Institute
h "angle subtended by the shear failure plane and the plane of the slab (shear
ASCE "American Society of Civil Engineers
friction)
CEB "ComiteH Euro-international du BeH ton
l "loaded length along beam strip (beam strip)
EC "European Code
J "section property of the critical section analogous to the polar moment of
FIP "FeH deH ration lnternationale de la PreH contrainte
inertia (ACI)
U of C"University of Calgary
M "service bending moment applied to the connection
M0 "nominal bending moment resistance when V "0
Terminology
M pos "positive bending moment capacity of beam strip (beam strip)
A "area of shear failure plane (shear friction) M neg "negative bending moment capacity of beam strip (beam strip)
A sv "cross-sectional area of a single shear reinforcement element (ACI, EC2) Ms "summation of M pos and M neg (beam strip)
a "angle subtended by the axis of the shear reinforcement and the plane of Mu "factored bending moment (ACI)
the slab (EC2) Ps "total load carried by one beam strip (beam strip)
a0 "ratio of the distance from the face of the column to the effective depth of R "normal force on a shear failure plane (shear friction)
the slab o, ol "flexural reinforcement ratio
as "factor which adjusts vc for support locations (ACI) s "spacing between rows of shear reinforcement
b "width of shear failure plane within effective depth (shear friction) S " shear force on a shear failure plane (shear friction)
b0 "perimeter of the critical section (ACI) p "average normal stress on the shear failure plane (shear friction)
b1, b 2 "dimensions of the critical section as defined in Fig. 13 T "tension force in the longitudinal reinforcement (shear friction)
b "coefficient to account for the effect of eccentric loading (EC2) Tv "tension force in a stirrup (shear friction)
bc "aspect ratio of the column (ratio of the long side to the short side of the qR d "basic shear resistance (EC2)
column) (ACI) u "perimeter of critical section (EC2)
cc "material resistance factor for concrete (EC2) v "average shear stress on a shear failure plane (shear friction)
cs "material resistance factor for shear reinforcement (EC2) V "service axial load applied to the connection
cv "portion of the unbalanced moment transferred by shear stresses (ACI) vc "contribution to the nominal shear stress resistance from the concrete
d "effective slab depth (ACI)
DR "drift ratio Vc "contribution to the nominal shear force resistance from the concrete
e "eccentricity of the load to the centroid of the critical section (ACI)
e1 "distance from the centroid of the critical section to point where shear vmax "maximum allowable value of shear stress resistance, v, (ACI)
stress is calculated (ACI) Vmin "minimum shear force resistance of all potential shear failure planes (shear
fc "design concrete strength (ACI) friction)
fck "characteristic concrete strength (EC2) vn "nominal shear stress resistance (ACI)
fcm "mean concrete strength when analysing test data Vn "nominal shear force resistance (ACI)
fy "yield strength of the flexural reinforcement V0 "nominal shear force resistance when M"0
fyd "design yield strength of the shear reinforcement (EC2) vR d1 "shear resistance per unit length without the effects of shear
fyv "mean yield strength of the shear reinforcement reinforcement (EC2)
h "total slab thickness vR d2 "maximum value for vR d3 (EC2)

Copyright ^ 2000 John Wiley & Sons, Ltd. Prog. Struct. Engng Mater. 2000; 2:386}399
 FLAT SLABdCOLUMN CONNECTIONS 387
vR d3 "shear resistance per unit length including the effects of shear VS d "total design shear force due to factored loads (EC2)
reinforcement (EC2) vu "factored shear stress due to Vu and M u (ACI)
vs "contribution to the nominal shear stress resistance from the shear Vu "factored shear force (ACI)
reinforcement (ACI)
"performance factor (less than 1.0) (ACI)
vS d "shear force per unit length due to factored loads (EC2) w "force per unit length; ‘loading term’ (beam strip)

connections subjected to concentric loading:

Flat slabs are a very common structural element for
— concrete strength,
medium- and high-rise buildings. Flat slab
— reinforcement ratio,
construction optimizes interior space and minimizes
— aspect ratio of the supporting column,
storey height resulting from the elimination of the
— perimeter to thickness ratio,
beams. In modern construction of multi-storey
— size effect.
buildings the preferred forming system is the so-called
fly-form. This form work system is most expedient Parameters of lesser importance are:
and economical if neither drop panels nor column
— grade of the reinforcement,
capitals are used. This leads to a slab with constant
— arrangement of the reinforcement,
thickness, the flat plate. There are two aspects of flat
— concrete cover,
plate design, which need special attention: the
— boundary conditions.
deflection and the punching shear strength of the slab
in the vicinity of columns. The focus of this paper will These parameters are now discussed and the two main
be the shear strength of the connection between the design codes, ACI 318-95[7] and EC2[8], will be
slab and the supporting columns often referred to in scrutinized in the light of the experimental evidence.
the literature as punching of two-way slabs.
Punching shear failure of slabs without shear CONCRETE STRENGTH
reinforcement is brittle in nature and this resulted in Elstner and Hognestad[9] systematically varied the
the accidental collapse of some flat slab buildings and concrete strength fcm and the reinforcement ratio in
to some failures during recent earthquakes[1]. For this their 1956 study on punching shear vu . A recent study
reason the performance of flat slab structures attracted by Marzouk and Hussein[10] on high-performance
increased attention in the last decade and recent concrete adds valuable information to the study of this
research has focussed on slabs with shear parameter. The results are plotted in Fig. 1 in terms of
reinforcement, capable of sustaining large inelastic vu/(100o)1/3 and compared with two curves which
deformations without resulting in collapse. express the effect of fcm on vu in the form (fcm )n where
n"1/3 and 1/2. The overall trend is well represented
by n"1/3 and 1/2 but any value n'1/2
Experimental information
overestimates the influence of fcm on vu for higher
The large body of experimental information and strength concretes. Tests[11] on a large full-scale slab
theoretical work on punching shear of flat slabs has with realistic boundary conditions in which external
been periodically collected, digested and published by restraints did not enhance the shear strength indicate
committees as background information for the that n"1/3 best represents the trend. The value
formulation of new codes or for code revisions. The n"1/3 has been proposed by many other authors and
most notable of these documents are the state-of-the- adopted in the CEB-FIP Model Code MC 90[12].
art reports produced by ACI-ASCE Committees[1–3]
and by a subcommittee of CEB Commission IV[4]. EFFECT OF FLEXURAL REINFORCEMENT
These reports and two recent publications[5, 6] provide With respect to the flexural reinforcement, four
most of the background for this paper. parameters are of interest: Reinforcement ratio o,
The punching shear problem, like the beam shear
problem, defies a convincing theoretical solution to
this day. This means that we have to rely primarily on
experimental evidence for the formulation of
appropriate code clauses for the design of slabs
against punching shear.

Parameters affecting the punching shear

strength of concentrically loaded slabs
without shear reinforcement
The following parameters have been identified as
most important for the resistance of slab–column Fig. 1 Effect of concrete strength on shear strength

Copyright ^ 2000 John Wiley & Sons, Ltd. Prog. Struct. Engng Mater. 2000; 2:386}399
388 STRUCTURAL ANALYSIS AND CAD

grade and arrangement of reinforcement and concrete f(bc)"(0.65#0.35/bc ). It should be noted that this
cover. Results of tests[9, 13–20] in which the yields a value less than 1.0 for bc'1.
reinforcement ratio o was the main variable are
plotted in Fig. 2 in terms of vu/(fcm)1/3 together with the PERIMETER-TO-THICKNESS RATIO
function 0.7(100o)1/3. This type of function has been This effect has been evaluated in Ref.[6] assuming the
proposed by many researchers and forms the basis of critical perimeter at d/2 from the face of the column as
several major building codes. recommended in Ref.[2]. Test results[9, 14–19] relating
Tests of slabs in which the yield strength of the b0/d to vu/(fcm100o)1/3 are shown in Fig. 3. This plot
reinforcement was the main variable[20] indicated that demonstrates the strong influence of this parameter on
if the flexural reinforcement yields significantly before the punching shear resistance of slabs. The function
the shear strength has been reached, the shear strength f(d/b0)"(6.7d/b0#0.4) is proposed in Ref.[6] to reflect
is reduced. This effect can be attributed to the loss of this trend.
membrane action. However, a plot of the shear
strength vs. yield strength in Ref.[6] does not indicate SIZE EFFECT
a clear trend. The size effect on the shear strength of concrete slabs
The evaluation of slab tests[21–23] with concentration has been investigated by several researchers[25–30].
of the reinforcement near the column showed that They observed a reduction in shear strength with
such concentration of reinforcement may be increasing depth. This effect is appropriately reflected
detrimental if too much steel is concentrated near the by f(d)"1/(d)1/3.
column. This harmful effect can be attributed to a loss
of bond[23].
Tests with different concrete cover, showed no effect Parameters affecting the shear strength
of cover on the punching shear strength. The authors of slabs with shear reinforcement
concluded that the total slab thickness, h, should be
Tests of slab–column connections with shear
used to calculate the shear strength. Several
reinforcement demonstrate that shear reinforcement
authors[4, 23] recommend to use the total slab thickness
not only considerably increases the shear strength but
h to calculate the shear strength.
also the ductility, see Fig. 4. This increase depends on
many factors, the most important being the type,
amount and the detailing of the shear reinforcement.
COLUMN ASPECT RATIO
Experimental studies[24] indicate that the
concentration of concrete strains near the corner of TYPES OF SHEAR REINFORCEMENT
rectangular columns increases with increasing aspect The following main types of shear reinforcement have
ratio bc ("ratio of long side to short side of a column) been used in experiments.
and that circular columns give higher shear strength — conventional stirrups,
than rectangular ones, presumably due to the absence — hat bars,
of stress concentrations. The evaluation in Ref.[6] of the
tests reported in Ref.[24] resulted in the expression

Fig. 2 Effect of flexural reinforcement ratio on shear strength Fig. 3 Effect of perimeter to thickness ratio on shear strength

Copyright ^ 2000 John Wiley & Sons, Ltd. Prog. Struct. Engng Mater. 2000; 2:386}399
 FLAT SLABdCOLUMN CONNECTIONS 389

Fig. 5 Shear studs

Fig. 4 Influence of shear reinforcement on ductility

— bent-up bars,
— ladders,
— shear studs.
Many other forms of shear reinforcement have been
investigated, they are not summarized in Refs.[20, 31].
Many of these are not practical because of difficulties
during installation. For conventional stirrups to be
practicable and effective, the slab should be at least Fig. 6 Different arrangements of shear studs
250 mm thick and the stirrup cage must consist of
closed stirrups with 1353 hooks and longitudinal bars
row of studs from the column. If the first element is
at the corners[31]. Even so, the cages intersect at the
more than about 0.5d away from the column face,
column and installation is difficult. Single vertical bars
a steep shear crack missing the first row of shear
with 1803 hooks fitting tightly around longitudinal
elements may lead to premature failure[31]. On the
bars at both ends are rather effective in increasing the
other hand, if the first shear element is too close to the
punching shear strength[32], but hat bars with 903
column face, it may be ineffective as it may be
hooks are inefficient[32, 33]. Generally, normal bent radii
bypassed by the failure crack. Evaluation of numerous
for hooks lead to anchorage slip of stirrups and as
tests[37] shows that s40.75d leads to satisfactory
a result, the full yield strength of the stirrups is rarely
results.
reached, particularly in shallow slabs.
Bent-up bars provide some increase in shear
resistance, but in many tests[9] failure occurred outside
YIELD STRENGTH OF THE SHEAR REINFORCEMENT
Recent tests in Germany[36, 38] and in the UK[39, 40] on
the shear-reinforced zone.
slabs reinforced with shear studs with fy5500 MPa
From the above discussion it is obvious that the
indicate that it is not necessary to limit fy to values less
anchorage of the shear reinforcement plays a major
than 500 MPa. It is, of course important to provide
role for the efficiency of the shear reinforcement.
good anchorage for higher strength shear studs; this is
Recognizing this, the shear studs shown in Fig. 5 were
ensured with a head diameter equal to about three
developed at the University of Calgary[18, 31, 33–35].
times the shaft diameter.
These studs are provided with a large head at the top
and a weld to a steel strip at the bottom. The large
heads provide near slip-free anchorage and promote SHEAR STRENGTH OF THE CONCRETE IN THE PRESENCE
OF SHEAR REINFORCEMENT
development of the yield strength of the vertical shaft.
For the design of slabs with shear reinforcement, the
The comb-like assembly also facilitates easy
key questions are: what is the contribution of the
installation of the flexural reinforcement.
concrete to the shear resistance; what is the inclination
of the failure crack, and how far does the shear
ARRANGEMENT OF SHEAR REINFORCEMENT reinforcement have to extend from the face of the
Stirrup cages are normally arranged in two orthogonal column to prevent shear failure outside the shear-
strips framing into the column. In most of the tests the reinforced zone?
shear studs were arranged as shown in Figs 6a and b. Tests[9, 20, 32, 41, 42] show that the shear resistance of
Both arrangements are equally effective. An important slabs with conventional shear reinforcement is
aspect for the placement is the distance s of the first significantly smaller than the sum of the shear

Copyright ^ 2000 John Wiley & Sons, Ltd. Prog. Struct. Engng Mater. 2000; 2:386}399
390 STRUCTURAL ANALYSIS AND CAD

resistance of the concrete alone, Vc , and the

contribution of the shear reinforcement, Vs. This is due
to either the concrete reaching the full shear capacity
before the yield capacity of stirrups is developed, or
the stirrups not developing the yield strength because
of anchorage slip. Measurements on stirrups[31]
support the latter. If the shear reinforcement is well
anchored its contribution to the shear resistance is
increased. Evaluation[4] of tests with well-anchored
shear reinforcement shows that the reduction of the
contribution of the concrete does not exceed about
25%. The plot of the results of AndraK [43] in Fig. 7 (from
Ref.[4]) is typical for shear studs.

SHEAR RESISTANCE OUTSIDE THE

SHEAR-REINFORCED ZONE
Shear failure outside the shear-reinforced zone occurs
when the shear resistance of the concrete is exceeded.
Experimental evidence[9, 18, 19, 35] shows that the
concrete resistance decreases as the critical section
moves away from the face of the column, refer to
Fig. 8[6]. The upper limit of vu is the two-way shear
resistance of the concrete and the lower limit is the
value for one-way shear[7]. Fig. 8 indicates that the
one-way shear is reached at a distance of about four Fig. 7 Contribution of concrete to shear strength in slab with
times the effective depth of the slab. shear studs

Slab`column connections with

moment transfer
Edge and corner column connections are always
subjected to moment transfer. At interior joints,
unbalanced moments are caused by pattern loading,
unequal span lengths and lateral loads. In addition to
the parameters that affect the capacity of
concentrically loaded connections, the shear resistance
of eccentrically loaded slab–column connections also
depends on the eccentricity of the load, e, expressed by
the ratio M/V.
The information on slab–column connections with
moment transfer is extensive and cannot be fully
reviewed here. The interested reader is referred to the
comprehensive report by Braestrup and Regan[4].
As a cursory overview of the effect of load
eccentricity on punching shear resistance, results of Fig. 8 Shear strength of slabs failing outside the shear-reinforced
zone
Stamenkovic and Chapman[44] are shown in Fig. 9,
normalized for concrete strength. In these tests on
interior, edge and corner columns the eccentricity was moments applied normal to one edge. This curve
varied systematically. The results on interior suggests that there is little interaction between
connections show the same linear variation as the edge moment and shear force. However, tests with moment
columns with moment transfer in the direction parallel transfer in the diagonal direction show an increase in
to the free edge, Fig. 9a, and consequently are not moment and shear resistance when V and M are
shown. If the moment at the edge column is applied applied simultaneously.
normal to the free edge, Fig. 9b suggests an interaction
curve of the form (Vu /V0 )2#(Mu/M0)2"1, where V0 is Reversed cyclic loading
the shear resistance when M"0, and M0 the moment
resistance when V"0. The interaction diagram for For slab}column connections reversed cyclic loading
corner columns, shown in Fig. 9c was obtained for is used to simulate a seismic event. Not unlike an

Copyright ^ 2000 John Wiley & Sons, Ltd. Prog. Struct. Engng Mater. 2000; 2:386}399
 FLAT SLABdCOLUMN CONNECTIONS 391

Fig. 9 Interaction diagrams for eccentrically loaded slab column Fig. 10 Effect of shear reinforcement on drift ratio [51]
connections [4, 44]

earthquake, this represents a severe test. In concrete slab}column frames to both uniformly
slab}column connections, like any other structural distributed gravity loads and large reversed-cyclic
element, moment reversal leads to a reduction in shear lateral displacements. The set-up includes a series of
capacity. The main focus in this section will be on slabs steel boundary frames installed along the free edges of
with shear reinforcement. the slab (parallel to the plane of the slab}column
frame), which facilitate vertical deflections of the slab
TESTS ON ISOLATED SPECIMENS but prevent rotations in the plane of the frames. Thus
In an early study on slab}column connections realistic boundary conditions exist. Recently, the
subjected to reversed cyclic loading, Hawkins et al[45] assembly has been used to test a slab}column frame,
investigated slabs reinforced against shear failure by designed for gravity loads only, under seismic
‘integral beam stirrup reinforcement’. The authors conditions. The 150 mm thick slab measured 5.0 m in
emphasized the importance of the detailing of the width and 10.25 m in overall length and was designed
shear reinforcement to include closely spaced in accordance with the Canadian code provisions[52] to
(s"d/3) closed stirrups with 1353 hooks and withstand a uniform service live load of 4.8 kPa. Shear
longitudinal bars in the corners. The specimens failed studs were used to provide the necessary punching
in a ductile manner after reaching drift ratios between capacity.
3.5 and 4.9%. The results of these and other During the test, two vertical actuators subjected the
tests[22, 46–50] were used in Ref.[51] to produce Fig. 10, slab to a constant uniform load of 2.2 kPa. Two
which shows the drift ratio DR as a function of the horizontal actuators subjected the top and bottom
ratio Vu /V0 . This figure demonstrates the excellent ends of the 1.50 m long columns to ever increasing
ductility of slabs with shear studs (SSR). reversed-cyclic displacements. The procedure was
In a recent test series[22] the main variables were the stopped when the displacement at the column ends
distribution of the flexural reinforcement and the reached 70 mm, corresponding to a lateral drift ratio of
direction of the imposed displacement. The hysteresis 4.7%. To conclude the test, the lateral loads were
curves of two tests in which the columns were loaded removed and the gravity load increased to failure. One
diagonally are shown in Fig. 11. Test SBJ7 had no shear edge column failed at 99% and the interior column at
reinforcement while test SJB 8 was provided with 119% of the nominal capacities predicted by the
shear studs. Fig. 11 confirms that slab-column joints Canadian code. This shows that the structure was able
properly reinforced against shear failure exhibit very to maintain its full factored design capacity after
ductile behaviour, while slabs without shear undergoing extremely large inelastic displacements.
reinforcement fail in a brittle manner. It is interesting
to note that in this test series the slabs loaded in the
principal direction reached exactly the same moment CODE PROVISIONS
capacity as the slabs loaded diagonally. This means For a reinforced concrete beam the shear stress is
that for square columns the direction of loading has no defined on the basis of the cross section. The same
influence on the capacity of the connection. In this approach is used for slabs, defining the shear stress as
series, concentrating the flexural reinforcement near the shear force divided by the area of a control surface
the column resulted in a slight decrease in moment (critical section) around the column. The factor of
capacity but in an increase in ductility. safety against punching is then established by
calibrating the important parameters influencing the
LARGE SLAB COLUMN TEST shear stress resistance of the concrete for the selected
Fig. 12 shows the test set-up developed at the U of control surface. This empirical approach has been
C for subjecting full-scale, two-span continuous, adopted by most building codes.

Copyright ^ 2000 John Wiley & Sons, Ltd. Prog. Struct. Engng Mater. 2000; 2:386}399
392 STRUCTURAL ANALYSIS AND CAD

Fig. 11 Hysteresis curves for connections subjected to reversed cyclic loading [22]

In (2), b0 is the length of the critical section,

c"[1/1#2/3(b1/b2] the fraction of unbalanced
moment, and J the property of critical section
analagous to polar moment of inertia. The effective
depth d is the average for the two directions and the
dimensions e1, b1 and b2 are defined in Fig. 13. The load
factors for dead and live loads are 1.4 and 1.7,
respectively.
The shear stress resistance of the concrete, vn, is the
smallest of the following three values (in SI units):

vc"0.083(2#4/bc ) (fc (3)

Fig. 12 Test set-up for two-span slabdcolumn frame subjected to

where fc is the design cylinder strength of the concrete
reversed cyclic loading and bc is the ratio of the long to the short side of the
column and should not be taken (2,

The American Code ACI 318-95[7] vc"0.083(2#as d/b0) (4)

In the following discussion only the most important with as"40 for interior columns, 30 for edge columns
features of the code are presented. The critical section and 20 for interior columns.
is assumed to be located at a distance d/2 from the face If vu/
5vn , shear reinforcement has to be
of the column as shown in Fig. 13. The design is based provided. The shear stress resistance, vn, of slabs with
on shear reinforcement is
vu/
4vn (1) vr"vc#vs(vmax (5)
where
("0.9) is a performance factor.
where vc is the shear stress resistance of concrete, vs the
The distribution of the shear stresses is based on the
shear stress resistance provided by shear
eccentric shear stress model shown in Fig. 14. The
reinforcement and vmax the maximum allowed shear
shear stress vu due to a factored shear force Vu and an
stress resistance. The values adopted by ACI for vc ,
unbalanced factored moment Mu is calculated at the
vs and vmax are as follows: in the presence of shear
critical section as
reinforcement the shear stress resistance is reduced to
Vu cv Mue1 0.166 (fc . The shear stress resistance of the shear
vu" # (2)
b0 d J reinforcement is based on a 453 truss model and is

Copyright ^ 2000 John Wiley & Sons, Ltd. Prog. Struct. Engng Mater. 2000; 2:386}399
 FLAT SLABdCOLUMN CONNECTIONS 393

Fig. 13 ACI critical sections for punching

s40.75d if vu /
40.5(fc and s40.5d if
vu /
'0.5(fc. Similar provisions have been adopted
by the Canadian Building Code[52].

The European Code EC 2[8]

The shear due to factored loads is calculated as a force
per unit length at the critical section at a distance 1.5d
from the face of the column, as defined in Fig. 15
VSdb
vsd" (7)
u
where, VSd is the total design shear force due to
Fig. 14 ACI eccentric shear stress model factored loads, u the perimeter of critical section and
b the coefficient to account for the effect of eccentric
expressed by loading ("1.5, 1.4 and 1.15 for corner, edge and
interior columns, respectively). EC2 states that for
nAsv fyv a more rigorous analysis other values of b may be
vs" (6)
b0 s used, as for example those obtained by the expressions
given in the British Code BS 8110. The EC2 load factors
In (6), n is the number of shear reinforcing elements for dead and live load are 1.35 and 1.5, respectively.
along a line parallel to the column perimeter, Asv the The value vSd is compared with the shear resistance
cross-sectional area of shear reinforcing element and per unit length given by
s the spacing of shear reinforcement.
The maximum allowed shear stress, vmax"0.5(fc. vRdl"qRd k(1.2#40ol) d (8)
The ACI Code stipulates that the shear stress be In (8) k is the size factor"1.60!d (d in metres).
checked at d/2 outside the shear reinforced zone (see The basic shear resistance qRd is
Fig. 13) and its value shall not exceed 0.166(fc .
qRd"0.75fck
2/3
/cc (9)
In a recent report[36] ACI Committee 421 recognizes
the better performance of slabs reinforced with shear where fck is the characteristic concrete cylinder
studs and allows to increase vc by 50% to 0.25 fc and the strength and cc ("1.5) is the material resistance factor
upper limit from 0.50(fc to 0.66(fc . The distance of for concrete. If vSd'vRdl shear reinforcement has to be
the first row of headed bars from the face of the provided but it may only be used in slabs with
column, 0.35d4s140.4d. For subsequent rows a minimum thickness of 200 mm. The shear resistance

Copyright ^ 2000 John Wiley & Sons, Ltd. Prog. Struct. Engng Mater. 2000; 2:386}399
394 STRUCTURAL ANALYSIS AND CAD

per unit length of a slab with shear reinforcement is reduced if bc'2. However, the evaluation of test
calculated as results[6] indicates a reduction for bc'1.
vRd3"vRdl# (Asv fyd sin a)/u(vRd2 (10)
PERIMETER TO THICKNESS RATIO
In (10), Asv is the area of all the shear reinforcement Fig. 16 showing the effect of the parameter c/d (instead
located inside the critical section defined in Fig. 15. of b0 /d) indicates that the ACI code is on the unsafe
fyd"fyu/cs is the design yield strength of the shear side for c/d'2.5. This value corresponds to a b0 /d
reinforcement sloped at angle a and cs"1.5. The ratio of 14, which is smaller than the limiting value of
maximum allowed value, vRd2"1.6vRdl . The distance 20, beyond which a reduction is to be considered in the
of the first row of shear reinforcement from the face of ACI code for interior columns.
the column shall not exceed 0.5d and subsequent rows
shall not be apart more 0.75d. It is also stipulated to SIZE EFFECT
check the shear resistance outside the critical section at EC2 correctly captures the trend discussed earlier. The
successive perimeters a distance 0.75d apart using (8). ACI code does not account for the decrease in shear
resistance as the slab thickness increases. For a slab
thickness exceeding about 300 mm ACI overestimates
Discussion of ACI Code and EC2
the punching shear resistance.
In view of the experimental data presented the
following comments are offered. SHEAR REINFORCEMENT
The most significant difference between the two codes
CONCRETE STRENGTH is the calculation of the punching resistance of slabs
The ACI code in which the effect of the concrete with shear reinforcement. This is due to two factors.
strength is expressed in terms of ( fc )1/2 reflects this First, in the ACI code the contribution of the concrete
effect quite well. The EC2 function does not correctly is reduced if shear reinforcement is required. Second,
present the influence of the concrete strength. This is the contribution of the shear reinforcement in ACI is
recognized in Ref.[53]. based on an inclination of the critical shear crack of
453, compared to 33.73 in EC2. As a result, much more
FLEXURAL REINFORCEMENT shear reinforcement is required for highly loaded slabs
The ACI code does not include the effect of the flexural designed according to the ACI code than for slabs
steel in the shear strength equation, and as a result designed to EC2. This is recognized in Ref.[53] where
may be unconservative for low reinforcement ratios. a factor 0.5 is advocated for the contribution of the
In EC2, the expression for the effect of ol in (8) reflects shear reinforcement in EC2 to produce a safe design.
this effect correctly. In both codes the effective depth Another concern is the resistance outside the shear-
rather than the total depth has been used for reinforced zone. As shown in Fig. 17, EC2 is unsafe
calibration. In the light of the test results[23] it would be when the critical perimeter is more than 3d from the
preferable to base the punching shear strength on the face of the column.
total depth, h.
MOMENT TRANSFER
COLUMN ASPECT RATIO The EC2 factors accounting for moment transfer must
Both codes include the effect of this variable, but in be used judiciously as the actual effect of unbalanced
different ways; ACI in (3) and EC2 by its definition of moments on the shear stresses may exceed these
the critical section. According to (3) the value of vn is factors. Even though the ACI eccentric shear stress
model has its shortcomings it is more rigorous in this
respect. For biaxial moments ACI predicts maximum
stress at one corner of the critical section (Fig. 14). For
the tests[22] mentioned earlier, the peak stress for
biaxial bending should have resulted in a smaller
failure load, than for uniaxial bending. The round

Fig. 15 EC2 critical sections Fig. 16 V Test /VACI as a function of c /d

Copyright ^ 2000 John Wiley & Sons, Ltd. Prog. Struct. Engng Mater. 2000; 2:386}399
 FLAT SLABdCOLUMN CONNECTIONS 395

Kinnunen and Nylander[55], which has subsequently

been modified and improved by others. Other models
include models based on plasticity[56], various forms of
beam strip models[9, 20] and numerous others. All of
these models are well documented. In the following
section two new models are presented. One is based
on the principle of shear friction; the other is a new
strip model.

THE SHEAR FRICTION MODEL

Many design codes have adopted the principle of
shear friction as a means to transfer shear across an
interface along which slip can occur. Recently, shear
Fig. 17 V Test /VEC2 as a function of a0 friction was used to accurately predict the shear
capacity of beams[57]. The basis for the model is
a simple relationship given by
corners of the EC2 critical section or a plastic stress
distribution used in MC 90[12] better reflect the actual l"k( fcp)1/2 (11)
behaviour as they eliminate singular points of where, l is the average shear stress on shear failure
maximum stress. Also, the factor cv may overestimate plane, p the normal shear stress on shear failure plane,
the contribution of the unbalanced moment to vu as and k the correlation coefficient to reflect the effect of
discussed in Ref.[54]. the concrete strength on shear friction, determined
from tests, see (14).
Recommendations for seismic design of The free-body diagram of Fig. 18 describes the
forces acting on an idealized faceted failure surface[59].
flat slabs for punching
The location of each facet is defined by an origin at
Currently, design codes do not have provisions for the a distance, x, from the inside edge of the support, and
shear design of flat slabs systems in seismic zones. The an angle from the horizontal, h. Without considering
results of the aforementioned tests suggest that shear a tensile stress across the failure surface, the shear
reinforcement, particularly shear studs, provide resistance of slabs without shear reinforcement
a suitable means of maintaining the integrity of approaches zero. To offset this effect, a nominal tensile
slab}column connections when subjected to strength, ft , was introduced across the failure surface.
displacement reversals well into the inelastic range. Across each facet, l"S/A and p"R/A, where
The experimental work[22, 45, 50, 51] confirmed that A"bh/sin h. Substituting into (11) and using force
load reversals reduce the shear resistance of the equilibrium yields the shear capacity
connection. On the basis of this research, the following

 
 
recommendations can be made[22, 36, 51]—in the ACI K T rt
Vsf" [1#cot2 h] 4 # #cot2 h!cot h
format—for the shear design of flat slab}column 2 K k2
connections subjected to seismic conditions:
— The shear resistance of the concrete may be taken
as vc"0.1(fc for slabs with shear studs and
!T cot h#Tv
 (12)

where
vc"0.06(fc reinforced with conventional stirrups.
— Minimum shear reinforcement should be provided K"0.25k fcbh (13)
to resist a shear stress vs"0.17(fc. This shear
k"0.53[30(MPa)/fc]0.45 (14)
reinforcement should be arranged as shown in
Fig. 6a and extended at least 4d from the faces of rt"ft /fc (15)
the column.
— Except for low-rise buildings flat slab systems In these equations, S is the shear force on shear
should not be considered part of the lateral load failure plane, R the normal force on shear failure
resisting system. plane, T the tension force in longitudinal
reinforcement, Tv the tension force in the shear
These recommendations should produce safe and reinforcement crossing the failure plane and
ductile slab}column connections in seismic zones. b"(btop#bbot)/2.
Fig. 19 shows the a plot of the predicted punching
New theoretical models resistance of specimen SC12[33].The two curves are for
two failure surfaces, one originating at the face of the
Over the years, many theoretical models have been column, the other at the outer edge of the shear
proposed. One of the first model was suggested by reinforcement. The resistance is plotted as a function

Copyright ^ 2000 John Wiley & Sons, Ltd. Prog. Struct. Engng Mater. 2000; 2:386}399
396 STRUCTURAL ANALYSIS AND CAD

Fig. 18 Facetted failure surface for punching shear

Fig. 19 Plot of punching shear resistance for failure planes rotating about (a) and (b)

of the relative distance a0"x/d. The lowest point on will fail outside the shear-reinforced zone at
each curve represents the punching capacity for that Vmin"657 kN which compares favourably with
particular origin. According to Fig. 19 the specimen Vu"595 kN recorded in the test.

Copyright ^ 2000 John Wiley & Sons, Ltd. Prog. Struct. Engng Mater. 2000; 2:386}399
 FLAT SLABdCOLUMN CONNECTIONS 397

Fig. 20 V Test /Vpred for slabs without shear reinforcement

Fig. 21 V Test /Vpred for slabs with shear reinforcement

Data of 61 tests on concentrically loaded specimens reaction, Ps , at the column end. Bending moments,
without, and 30 tests with shear reinforcement are Mneg and Mpos are applied at the ends of each strip. The
compared with the predictions of the shear friction authors assume that the loading term, 2w, is the
model in Figs. 20 and 21, respectively. Using both limiting internal shear that can be transferred through
shear friction and yield-line procedures, the mode of the edge of each radial strip from the quadrants. The
failure was correctly predicted for 56 of the 61 tests flexural strength of the radial strip, Ms , is the sum of
without shear reinforcement, and for 25 of the 30 tests the positive and negative flexural capacities at the
with shear reinforcement. The average test-to-predicted strip ends. The loaded length of the strip is l, and the
value is 1.01 and the coefficient of variation for both total load carried by one strip is Ps . Equilibrium of the
sets of data is 14.2%. The angle of the failure surface, h, radial strip leads to
was given for five tests without shear reinforcement to
lie between 20 and 30 degrees. For these tests the l"(Ms/w)1/2 (16)
predicted angles are between 22 and 263. Ps"2wl"2(Ms/w)1/2 (17)

THE STRIP BEAM METHOD The shear capacity of a strip is assumed to be limited
Beam analogies have been suggested by a number of by the intensity of the internal shear that can be carried
authors, including Hawkins and Corley[58]. A more along each edge and the flexural capacity of the strip
explicit method has recently been developed by itself. Since the slab behaves as a slender flexural
Alexander and Simmonds[23] and expanded by member in the direction perpendicular to a radial
others[60, 61]. According to this approach, the slab is strip, it is assumed that the limiting value for the
subdivided by two orthogonal radial strips into four internal shear between the strips and the quadrants of
quadrants as shown in Fig. 22a. Each radial strip is the slab is the design value for one-way shear
loaded along each edge and supported by a vertical resistance of the slab.

Copyright ^ 2000 John Wiley & Sons, Ltd. Prog. Struct. Engng Mater. 2000; 2:386}399
398 STRUCTURAL ANALYSIS AND CAD

** [3] ACI-ASCE COMMITTEE 426. The shear strength of reinforced

concrete slabs. ASCE Journal of the Structural Division 1974: 100(ST8):
1543d1591.
** [4] Regan PE & Braestrup MW. Punching shear in reinforced concrete:
a state-of-the-art report. Bulletin d’Information No. 168, ComiteH Euro-International
du BeH ton, Lausanne. 1985. 232.
Comprehensive state-of-the-art report.
[5] Dilger WH & Sherif A. Shear strength of slab-column connections
according to ACI code, EC2 and DIN 1045. Proceedings of the International ECSN
Symposium on European Concrete Standards in Practice, Copenhagen, Denmark,
May 1997. 131d140.
* [6] Sherif A & Dilger WH. Critical review of the punching shear provisions
for interior columns. Canadian Journal of Civil Engineering 1996: 23: 998d1011.
[7] ACI Committee 318. Building code requirements for reinforced
concrete. Detroit, MI: American Concrete Institute. 1995.
[8] Eurocode 2 (EC2). Design of concrete structures, Part 1: general rules
and rules for buildings. 1991.
* [9] Elstner RC & Hognestad E. Shearing strength of reinforced concrete
slabs. ACI Journal 1956: 28(1): 29d57.
Systematic tests investigating many variables.
[10] Marzouk H & Hussein A. Experimental investigation on the behaviour
of high-strength concrete slabs. ACI Structural Journal 1991: 88(6): 701d713
[11] Sherif A & Dilger WH. Punching failure of a full-scale high strength
concrete flat slab. Proceedings, International Workshop on Punching Shear Capacity
of RC Slabs, Stockholm, June 2000.
[12] MC 90 ComiteH Euro-International du BeH ton (CEB)-FeH deH ration
Internationale de Precontrainte (FIP). Model code for concrete structures,
1990 CEB. Lausanne: Swiss Federal Institute of Technology.
[13] Base GD. Some tests on the punching shear strength of reinforced
concrete slabs. Technical Report TRA 321, London: Cement and Concrete
Fig. 22 The strip beam model Association. July 1959.
[14] Manterola MJ. Punzonamiento de placas sin armadura de esfuerzo
cortante hormigon y acoro. No. 83, Paper 457-8-30. 1967. 71d98.
[15] Criswell ME. Strength and behaviour of reinforced concrete slab-column
Using the one-way shear expressions of ACI[7], the connections subjected to static and dynamic loadings. US. Army Engineer Waterways
model was compared with the test results of 115 Experiment Station, Vicksburg, December 1970. Technical Report N-70-1.
concentrically loaded slabs that failed by punching. [16] Petcu V, Stanculescu G, Pancaldi U & Ioneseu P. Studiu de sinteza
The model was found to produce a conservative privind comportarea la stapungere a placilor armate pe dona durectii (synthesis study
concerning the punching behaviour of two-way reinforced concrete slabs). St. cerc.
prediction of the failure load, with an average test-to- Incerc, H. 3, 1973.
predicted value of 1.29 and a coefficient of variation [17] Ghali A, Elmasri MZ & Dilger YM. Punching of flat plates under static
of 12.3%. and dynamic horizontal forces. ACI Journal 1976: 73(5): 566d572.
This approach has recently been extended[62] to [18] Van der Voet AF, Dilger WH & Ghali A. Concrete flat plates with
well-anchored shear reinforcement elements. Canadian Journal of Civil Engineering
include interior slab}column connections subjected to
1982: 9(1): 107d114.
combined shear and unbalanced bending moments [19] Mokhtar A, Ghali A & Dilger WH. Stud shear reinforcement for flat
and to slab}edge column joints. plates. ACI Journal, Proceedings 1985: 82(5): 676d683.
One weakness of the model is that it does not [20] Hawkins NM & Mitchell D. Progressive collapse of flat plate structures.
ACI Structural Journal 1979: 76(7): 775d808.
predict a typical punching failure surface. Shear [21] Sherif A. Behaviour of reinforced concrete flat slabs. PhD Thesis,
reinforcement has not been incorporated in the model University of Calgary, 1996. 1d399.
to date. [22] Brown S & Dilger WH. Seismic response of flat plate column
In time, the research community expects connections. Proceedings of the Canadian Society for Civil Engineering Annual
Conference, Winnipeg, June 1994 (Vol 2) 388d397 (also submitted to ACI Structural
a theoretical model to replace the empirical provisions Journal for publication).
of the present codes. [23] Alexander SD & Simmonds SH. Tests of columneflat plate
connections in monolithic reinforced concrete structures. ACI Structural Journal
1992: 89(5): 495d502.
[24] Hawkins NM, Fallsen HB & Hinojosa RC. Influence of column
References and recommended reading rectangularity on the behaviour of flat plate structures. ACI Special Publication SP-30.
Detroit: ACI. 1971: 127d146.
Papers of particular interest have been marked:
[25] Taylor R & Hayes B. Some tests on the effect of edge restraint on
* Special interest
punching shear in reinforced concrete slabs. Magazine of Concrete Research 1965:
** Exceptional interest
17(50): 39d40.
[26] Graf O. Versuche uK ber die WiderstandsfaK higkeit von allseitig aufliegenden
** [1] ACI-ASCE COMMITTEE 352. Recommendations for design of
dicken Eisenbetonplatten unter Einzellasten. Deutscher Ausschuss fuK r Eisenbeton.
slab-column in monolithic reinforced concrete structures. ACI Structural Journal
1938. Heft 88.
1988: 85(6): 675d696.
[27] Kinnunen S, Nylander H & Tolf P. UndersoK kningar roK rande vid
References [1d3] provide information about the background and development genomstansning. Institutionen foK r Byggnadsstatik, Kungliga Tekniska HoK gskolan,
of punching of two-way slabs in the ACI code. Nordisk Betong. 1978: 25d27.
* [2] ACI-ASCE COMMITTEE 326. Shear and diagonal tension, Part 3 slabs [28] Kinnunen S, Nylander H & Tolf P. Plattjocklekens inverkan pas betong
and footings. ACI Journal 1962: 58(3): 353d396. plattors has llfasthet vid genomstansning, FoK rsoK k med rektangulaK ra plattor. Meddelande

Copyright ^ 2000 John Wiley & Sons, Ltd. Prog. Struct. Engng Mater. 2000; 2:386}399
 FLAT SLABdCOLUMN CONNECTIONS 399

No. 137. 1983. Stockholm, Sweden: Institutioen foK r Byggnadsstatik, Kungliga Tekniska [45] Hawkins NM, Mitchell D & Hanna SN. Effects of shear reinforcement
HoK gskolan. on the reversed cyclic loading behaviour of flat plate structures. Canadian Journal of
[29] Tolf P. Plattjocklekens Inverkon PAs Betongplattors Has llfasthet Vid Civil Engineering 1975: 2: 575d582.
Genomstansning. Stockholm, Sweden: Institutionen foK r Byggnadsstatik, Kungliga [46] Robertson IN & Durrani AJ. Gravity load effect on seismic behaviour of
Tekniska HoK gskolan. 1988. interior slab-column connections. ACI Structural Journal 1992: 89(1): 37d45.
[30] Regan PE. Behaviour of reinforced concrete flat slabs. Report [47] Pan AD & Moehle JP. Lateral displacement ductility of reinforced
No. 89. Construction Industry Research and Information Association. London: concrete flat plates. ACI Structural Journal 1989: 86(3): 250d258.
UK, 1981. [48] Pan AD & Moehle JP. Experimental study of slabdcolumn connections.
** [31] Dilger WH & Ghali A. Shear reinforcement for concrete slabs. ASCE ACI Structural Journal 1992: 89(6): 626d638.
Journal of the Structural Division 1981: 107(ST12): 2403d2420. [49] Wey EH & Durrani AJ. Seismic response of interior slabdcolumn
Summary paper on punching of flat slabs with stud shear reinforcement. connections with shear capitals. ACI Structural Journal 1992: 89(6): 682d691.
[50] Dilger WH & Cao H. Behaviour of slabdcolumn connections under
[32] Yamada T, Nanni A & Eudo K. Punching shear resistance of flat slabs:
reversed cyclic loading. Proceedings, 2nd International Conference on High-Rise
influence of reinforcement type and ratio. ACI Structural Journal 1992: 88(4):
Buildings, China, 1991.
555d563.
* [51] Megalli S & Ghali A. Design considerations for slabdcolumn connections
[33] Seible F, Ghali A & Dilger YM. Preassembled shear reinforcing units
in Seismic zones. ACI Structural Journal 1994: 91(3).
for flat plates. ACI Structural Journal 1980: 77(1): 28d35.
[34] AndraK HP. Ductile slab-column connections. [Link], University of Good discussion of seismic design aspects of slab-column connections.
Calgary, March 1970. 1d203.
[52] CSA Standard A 23.3-94. Design of concrete structures. Canadian
[35] AndraK HP, Dilger WH & Ghali A. Durchstanzbewehrung fuK r
Standards Association, Rexdale Ontario, 1994.
Flachdecken. Beton-und Stahlbetonbau 1979: 5: 129d132.
[53] Kordina K. Zum Tragsicherheitsnachweis gegenuK ber Schub, Torsion und
[36] Durchstanzversuche an Stahlbetonplatten mit RippenduK beln und
Durchstanzen nach EC2. Beton-und Stahlbetonbau 1994: 4: 97d100.
vorgefertigten GrossflaK chentafeln. Report No. 21-21634, Otto-Graf-Institut,
[54] Elgabri A & Ghali A. Moment transfer by shear in slabdcolumn
University of Stuttgart, July 1996.
connections. ACI Structural Journal 1996: 93(2): 187d196.
[37] ACI-ASCE COMMITTEE 421. Shear reinforcement for slabs.
[55] Kinnunen S & Nylander H. Punching of concrete slabs without shear
American Concrete Institute Report ACI 421.1R-99. 1999. 1d15.
reinforcement. Meddelande No. 38, Institution foK r Byggnadsstatik, Kungliga Tekniska
[38] Bericht uK ber Versuche an punktgestuK tzten Platten bewehrt mit DEHA
HoK gskolan, Stockholm, 1960.
Doppelkopfbolzen und mit DuK belleisten, Report No. AF 96/6-402/1, Otto-Graf-
[56] Braestrup MW, Nielsen MP, Jensen BC & Bach F. Axisymmetric
Institut, UniversitaK t-Stuttgart, DEHA 1996.
punching of plain and reinforced concrete. Report R75. Copenhagen: Structural
[39] Regan PE. Double headed studs as shear reinforcementetests of slabs
Research Laboratory, Technical University of Denmark. 1976.
and anchorages. London: University of Westminster. August 1996.
[57] Loov RE. Review of A 23.3-94 simplified method of shear design and
[40] Regan PE. Punching test of slabs with shear reinforcement. London:
comparison with results using shear friction. Canadian Journal of Civil Engineering
University of Westminister. November, 1996.
1998: 25(3): 437d450.
[41] Ghali A & Hammill N. Effectiveness of shear reinforcement in slabs.
[58] Hawkins NM & Corley WG. Transfer of unbalanced moment and shear
Concrete International 1992: 14(1): 60d65.
from flat plates to columns. ACI Special Publication SP 30, Detroit, 1974. 147d176.
* [42] NoK lting D. Das Durchstanzen von Platten aus Stahlbeton -Tragverhalten,
[59] Dechka DC, Loov RE & Dilger WH. Prediction of punching shear
Berechnung, Bemessung. Bulletin No. 62. Braunschweig: Institut fuK r Baustoffe,
capacity by shear friction. Proceedings, International Workshop on Punching Shear
Massivbau und Brandschutz, Technischen UniversitaK t. 1984.
Capacity of RC Slabs, Stockholm, June 2000.
This thesis presents a complete collection of test data published before 1984. [60] Siao WB. Punching shear resistance of flat slabs: a beam-strip analogy.
[43] AndraK HP. Zum Tragverhalten von Flachdecken mit DuK belleistene ACI Structural Journal 1994: 91(5): 595d604.
Bewehrung im Auflagerbereich. Beton d und Stahlbetonbau 1981: 3: 53d57 and 1981: [61] Desayi P & Seshadri HK. Punching shear strength to of flat slab corner
4: 100d104. column connections. Proceedings of the Institution of Civil Engineers and Structures
[44] Stamencovic A & Chapman JC. Local strength of flat slab at column and Buildings 1997: 122: 10d20.
heads. Report 39. London: Construction Industry Research and Information [62] Alexander SDB. A strip method of design for punching shear. Detroit:
Association. 1972. ACI Special Publication, 2000, in press.

Dr. Walter Dilger

Faculty Professor
Department of Civil Engineering
University of Calgary, 2500 University Drive NW
Calgary, Alberta, Canada T2N IN4
E-mail: dilger@[Link]

Copyright ^ 2000 John Wiley & Sons, Ltd. Prog. Struct. Engng Mater. 2000; 2:386}399