You are on page 1of 13

PACIFIC CONCRETE CONFERENCE

New ZeaZand 8-11 November 1988

PUNCHING SHEAR DESIGN IN THE NEW AUSTRALIAN CODE


FOR CONCRETE STRUCTURES

B.V. Rangan

University of New South Wales, Australia.

SUMMARY
The behaviour of a reinforced concrete slab in the vicinity of a support and
the punching shear mechanism are first explained with the aid of a physical
model. Based on this failure mechanism, design equations for the
calculation of punching shear strength of slabs in the presence of an
unbalanced bending moment are presented. These design equations are
contained in the new Australian Standard, AS 3600-1988. An example is used
to illustrate the application of the design equations.

INTRODUCTION
The region of a slab in the vicinity of a column could fail in shear by
developing a failure surface in the form of a truncated cone or pyramid.
This type of failure, called a punching shear failure, is usually the
source of collapse of flat plate and flat slab buildings. Adequate design
of this region of slab is therefore of paramount importance.

An extensive review of existing knowledge on punching shear strength of


slabs is available in various sources (1,2,3). In general, the region of a
slab near a column must transfer both shear force and unbalanced bending
moment to the column. Numerous tests have been carried out to evaluate the
punching shear strength of slabs where the moment transfer is zero. In
recent years, a significant amount of test data has also become available
for the case where both shear and moment are transferred. Several theories
have been put forward to predict the strengths observed in these tests.

The design practice usually takes a simple approach to the problem of


punching shear. In the ACI Code 318-83, as well as in AS 3600, the punching
shear strength, Vuo, where the moment transfer is zero is given by

Vuo = ud fcv (1)

In Eq. (1), u is the length of the critical shear perimeter defined by a


line geometrically similar to the boundary of the column and located at a
distance of d/2 therefrom (see Fig.9.2.1 (A) in AS 3600), d is effective
depth of the slab averaged around u, and fcv is given by AS 3600 as

(2)

where Ph is the ratio of the longest overall dimension of the column, Y, to


the overall dimension, X, at right angles to Y, and f' is the cylinder
C

strength of concrete expressed in terms of N/mm • 2

195
simulated by sixteen point loads on each panel.

Upon loading, the first cracks appear on the top surface of the slab near
the column. These are soon followed by cracking at midspan of the slab
panel. The moment fields that develop in a slab depend mostly on the
details of reinforcement provided in various parts of the panel. Because
the midspan region usually contains only light reinforcement, a yield line
first develops here. Further loading causes cracking and yield lines start
to form in the end regions of the slab panel. If the slab in the vicinity
of the columns did not fail in punching, a yield mechanism will develop and
the slab will fail in flexure. However, the flexural capacity of a slab is
significantly enhanced by the minimum reinforcement provided in the panels
for the purpose of crack control at service loads and by the in-plane forces
that develop due to the restraint provided by the unyielding parts of the
slab. Although the in-plane forces also increase the punching shear
strength, the increase in flexural strength is greater. For the above
reasons, a punching shear failure usually precedes a complete flexure
failure. In the tests conducted at the University of New South Wales [4],
all the three flat plate models have failed in this manner at a load more
than twice the design ultimate load. The flat plate model tested at the
University of Illinois [1] has also failed in a similar manner. The flat
slab model of this study would have suffered the same consequence if the
edge columns had been stronger. Further support to this argument is given
by the model tested by Simmonds [1] which has also failed in punching.
Failure Mechanism

The mechanism of punching shear failure can be explained with the aid of a
physical model shown in Fig.1. The model illustrates the situation in the
vicinity of an edge column without an edge beam. To be consistent with the
model used in the calcuation of Vuo (Eq.1), it is assumed that the critical
section for failure is at a distance d/2 away from the face of the column
* *
(or the capital). In Fig. 1, Mv and V are the bending moment and the shear
force transferred to the column centre for the limit state of collapse. The
shear force is transferred partly by V1 at the front face and the remainder
by V2 at each side face. The moment transfer occurs partly as the yield
moment, M 1, of the slab reinforcement at the front face of the critical
section, some due to the eccentricity of the shear force v 1 and the
remainder as torsional moment T2 at each side face. At an interior column,
transfer of forces also occurs at a back face of the critical section. At a
*
corner column, there is only one side face. At an edge column where Mv acts
parallel to the edge of the floor, there are front and back faces and one
side face.
Of the forces transferred at various faces of the critical section, only Mt
can be calculated fairly accurately. Where the part of Mv transferred due
to the eccentricity of the shear forces at the front and the back faces is
zero (for example, at a typical interior column), the value of T2 is also
easily calculated. But in general the value of T2 depends on the magnitude
a�d the eccentricity of shear force v,. Unfortunately, the present
knowledge regarding the actual distribution of V * between various faces of
the critical section is rather uncertain.

197
data has revealed that the slopes of the unit moment curves changed rapidly
at high loads and that the expressions proposed earlier might have
overestimated the shear force transferred at the side faces. Moreover, the
applicability of those expressions for the cases of interior and corner
columns is not known because there are no test data available.

In the absence of reliable information, it appears simple to assume that the


shear force transferred at each face is equal to (b 0 /u)V*, as stated
earlier. With this assumption and a known value of M1 it is then possible
to calculate the forces at each face of the critical section by means of
statics.
A punching shear failure is caused either by the failure of the slab strip
at the side face (called torsion strip) in combined torsion and shear, or by
the failure of the slab strip at the front face (and the back face, if any)
in shear.

The strength of the torsion strip in combined torsion and shear is many
times the value calculated for isolated beams [3, 4]. When subjected to
torsion, an isolated beam undergoes an increase in length due to the warping
of the cross-section and opening of inclined cracks. When such a beam is
part of a floor system, the monolithic slab provides considerable resistance
to the beam expansion. The torsion strip is in a situation similar to a
beam in an integrated floor system. The tests conducted at the University
of New South Wales [3, 4] showed that because of the slab restraint the
measured cracking torque of the torsion strip was approximately six times
the value obtained for an isolated beam. At failure, the measured strength
in combined torsion and shear of the torsion strip which contained closed
ties was about four times that of a similar isolated beam. These beneficial
effects of slab restraint are included in the design equations presented in
the next section.

DESIGN EQUATIONS

Elsewhere [5], equations have been developed for the calculation of punching
shear strength of reinforced concrete slabs in the presence of an unbalanced
bending moment. The derivations of these equations are not reproduced here
for want of space. Only a summary of equations is given.
From the failure mechanism discussed above, it might be clear that one of
the important factors that govern the punching shear strength calculation is
how the torsion strip (Fig.1) resists the combined effects of torsion and
shear acting there. Where the torsion strip contains no spandrel beam and
no closed ties, the torsion and shear must be resisted by slab concrete
alone. On the other hand, where the torsion strip contains a spandrel beam
provided with closed ties or no spandrel beam but only closed ties in the
torsion strip the load carrying mechanism and hence the final strength
equations are different. Therefore, equations have been developed for the
following three cases (5].

(a) Where the torsion strip contains no spandrel beam and no closed ties
(e.g. a slab in the vicinity of an interior column):

In this case, the torsion and shear at the side face and the shear at

199
Vmax = 3 Vumin �x/y but not greater than Vuo (11)

where x and y are the shorter and longer dimensions respectively of the
cross-section of the torsion strip or spandrel beam.

COMPARISON WITH TEST RESULTS


The correlation of punching shear strength predictions by the design
equations presented in the previous section with available test results are
presented elsewhere [5].' The average ratio of test strength to calculated
strength for 109 results is 1.61, with a coefficient of variation of 25
percent. The design equations appear to be rather over-conservative for two
reasons.
First, the value of Vuo given by Eq.1 is in itself generally conservative.
Note that all the other design equations are in terms of Vuo• It has been
found [1] that V uo increased with the increase in the average ratio of
tensile reinforcement provided in the slab and decreased slightly with
increasing depth of slab (scale effect). It is possible to modify Eq (1) to
take account of these factors. Although large percentages (up to 4 percent)
of tensile reiforcement have sometimes been used in laboratory specimens so
that a shear failure preceded a flexure failure, for slabs designed in
practice the average tensile steel ratio seldom varies widely. Similar
comments also apply to the depth of slabs. Considering the large scatter
that is usually inherent in punching shear data, Eq. (1) is retained as a
design equation for the present.
Secondly, the shear strengths measured in laboratory studies are
significantly influenced by the boundary conditions adopted for the test
specimens. The longitudinal restraint that develops in the slab is very
dependent on the boundary conditions of the test specimens. Tests carried
out at the University of New South Wales [3] have shown that the forces at a
given column are only slighly influenced by the loads applied more than one
panel away. Hence test specimens with boundaries extending at least one
panel away from the test region can be expected to be free from the effect
of boundary conditions and model closely the prototype floors. The test
conducted at the University of Illinois and at the University of New South
Wales satisfy the above condition. For these tests, Test Vu/Calculated Vu
ratios are 1.10 and 1.14 respectively. With reference to these tests, it
appears that the boundary conditions used in most of the other tests might
have overestimated , to a varying degree, the slab restraint that will be
present in prototype floors. The tests carried out at the University of
Illinois and at the University of New South Wales are, however, expensive
and time consuming. For this reason, only simple model tests are usually
feasible especially when a large number of parameters have to be
investigated. It is merely pointed out that when simple model tests are
used as a basis for design it is prudent to note that the punching shear
strengths measured in such tests may be larger than those that can be
expected in prototype floors.
In the light of the above remarks, the present design equations are
considered to be adequate for use in practice.

201
the short-span direction. In this example, the moment transfer in the
long-span direction is the critical one and therefore; the calculations are
given for this direction only.

(a) At Interior Column

According to the Simplified Method of Design of Slabs given by


AS 3600, the unbalanced moment at the first interior column is given
by
*
Mv = (0.70-0.65)Fd Lt L�/8 = 0.05 X 13.5 x 6 x 6.65 2 /8 = 22 kNm

But the moment transferred to an interior column should not be taken


less than the value (see Clause 7.4.5)

= 0.06 [(1.25 X 6 + 0.75 X 4)6 X 6.65 2 - 1.25 X 6 X 6 X 6.65 2

= 48 kNm
Therefore, take

Mv = 48 kNm

V* = 13.5 x 6 x 7 = 567 kN
Assume 16 mm diameter bars are used as the slab steel. Then

average d= 250 - 25 - 16 = 209 mm

b1 = b2 = 500 + 209 = 709 mm

u = 4 x 709 = 2836 mm

From Eq. (1),

Vuo = 2836 X 209 X 0.34 ,J"25 = 1008 kN


Because there are no closed ties provided in the slab, the punching
shear strength is given by Eq. (6).

Eq. (6) :
1008
Vu =
(1 + (48 X 2836)/(8 X 567 X 709 X 0.209))

= 838 kN

¢Vu = 0.7 x 838 = 587 kN > V*


Therefore, punching shear strength is adequate.

(b) At Edge Column

203
209
500 + 2 = 604.5 mm

u = (b1 + b2) = 2 x 604.5 = 1209 mm

From Eq. (1),

Vuo = 1209 X 209 X 0.34 ,f25 = 430 kN

Proceed as per the. edge column and calculate Vumin by Eq. (8).

Eq. ( 8) :

1.2(400/250)430
Vumin =
(1 + (67 X 1209)/(2 X 142 X 604.5 X 0.5)]

= 425 kN < Vuo

¢Vumin = 0.7 x 425 = 297 kN > V*

Therefore we require only minimum closed ties in the edge beam.

As before, use Y12 closed ties at 300 mm spacings.

REFERENCES

1. ACI-ASCE Committee 426. 1974 The Shear Strength of Reinforced Concrete


Member-Slabs. Proceedings. ASCE. 100 (8): 1543-1591.

2. Hawkins, N.M. 1974 Shear Strength of Slabs with Moments Transferred to


Columns. Shear in Reinforced Concrete, SP-42, American Concrete
Institute, Detroit.

3. Rangan, B. V. and Hall, A.S. 1983 Forces in the Vicinity of Edge


Columns in Flat Plate Floors. UNICIV Report No.R-203, University of
New South Wales, Kensington.

4. Rangan, B.V. and Hall, A.S. 1983 Moment and Shear Transfer Between
Slab and Edge Column. ACI Journal, 80,(3): 183-191.

5. Rangan, B.V. 1987 Punching Shear Strength of Reinforced Concrete


Slabs. Civil Engineering Transactions, Institution of Engineers,
Australia, CE29, (2): 71-78.

205

You might also like