You are on page 1of 9

Magazine of Concrete Research, 1999, 51, No. 2, Apr.

, 121±129

Punching strength of slabs reinforced for shear


with offcuts of rolled steel I-section beams
R. Gomes and P. Regan{

Universidade Federal de Goia s; University of Westminster

Punching shear failure may occur at columns supporting flat slabs or at concentrated loads. The rupture is brittle
and may occur clearly below the flexural capacity of the slab. This work aims to analyse the symmetric punching
resistance of reinforced concrete flat slabs with shear reinforcement. Twelve 200 mm thick slabs representing
regions around columns, loaded at the centre and reacted at the edges, were tested up to failure. Short offcuts of
steel I-beams were used as shear reinforcement. The principal variables considered in the tests were the area,
number and distribution of the shear steel elements. All the slabs failed by punching, and the failure load of the
specimen with the most reinforcement was more than twice that of the slabs without shear reinforcement.

Notation
problem of a slab with circular flexural reinforcement.
Asv shear reinforcement area Later, Kinnunen4 extended this model to slabs with
c column side orthogonal flexural reinforcement and considered dowel
d effective depth of slab 5,6
and membrane effects. Others have also made theor-
fc cylinder compressive strength of concrete etical and experimental contributions.
(' 0:8 f cu ) The use of shear steel in a reinforced concrete flat
f cu concrete cube strength slab to prevent punching failure was initiated by Graf
7
fy yield or 0´2% proof stress of steel 8
and Wheeler but only after 1971, when the first provi-
u length of control perimeter 9
sions for shearheads appeared in the ACI code, have
r ratio of reinforcement there been many tests of slabs with shear steel giving
10,11
workable empirical relationships.
Many different types of shear reinforcement have
Introduction
been used by researchers and have given some increase
1
Since Talbot's work many punching tests have been of punching strength. Although strength considerations
carried out on slabs and footings. Elstner and Hog- must be taken into account, the choice of the shear
2
nested reported tests on 39 slabs in which the concrete reinforcement type may be determined by practical
strength, flexural reinforcement ratio, column size and factors such as end anchorages, possible interference
amount and position of shear reinforcement were the with column and flexural reinforcement, facility of
main variables considered. In 1960, Kinnunen and installation within slabs of limited depth, and economy.
3
Nylander, on the basis of experimental work, pub- Studs, shear combs or, in general, individual bars with
lished the first rational theoretical approach to the end anchorages have the advantages of not interfering
with column and flexural reinforcement, being usable
in thin slabs and being relatively inexpensive.
 Universidade Federal de GoiaÂs, Escola de Engenharia Civil, Praca To investigate the effects of shear steel in rein-
UniversitaÂria s/no., Setor UniversitaÂrio, GoiaÃna GoiaÂs, Brazil 74605- forced concrete slabs, tests were performed on models
220. of reasonable scale, intended to represent the negative
{ School of Architecture and Civil Engineering, University of West- bending regions around columns or positive-moment
minster, London NW1 5LS, UK.
regions around concentrated loads, with their edges
(MCR 685) Paper received 19 January 1998; last revised 30 April representing contra-flexural lines in continuous slabs.
1998; accepted 18 June 1998 The principal variables were the area of shear steel in
121

0024-9831 # 1999 Thomas Telford Services Ltd

Downloaded by [ Imperial College London Library] on [24/04/17]. Copyright © ICE Publishing, all rights reserved.
Gomes and Regan

a circumferential layer, the number of layers of shear meter of the loaded area, the nominal design shear
steel and the distribution of shear elements. stress should not exceed 0:8( f cu )1=2 or 5 N=mm2.
The shear capacity of a failure zone without shear
reinforcement (vc ) is given by equation (2):
Design rules
r 
12
According to BS 8110/85, punching failures are 0:27 p 4 400
vc ˆ 3 (100r f
cu ) (2)
supposed to occur on the inclined faces of truncated ãm d
cones or pyramids, depending on the shapes of the
loaded areas. For calculation purposes control peri- where r < 0:03, 20 N=mm2 < f cu < 40 N=mm2 ,
p 
4 (400=d) > 1, d is in millimetres and ã
meters are rectangular in all cases. The nominal design m is the partial
shear stress (v) is defined as the design ultimate value safety factor for materials, equal to 1´25.
of the concentrated load (V ) divided by the effective When the nominal design shear stress calculated
length of the outer perimeter of the zone considered (u) from equation (1) exceeds the shear capacity calculated
and by the average effective depth (d): from equation (2), shear reinforcement may be pro-
vided according to equation (3):
v ˆ V =(ud) (1)
8
> (v ÿ vc )ud
>
>
< for vc < v < 1:6vc
The first normal perimeter (u) is defined as the X 0:87 f yv
boundary of the smallest rectangle drawn at a least Asv sin á ˆ
>
>
> (3:5v ÿ 5:0vc )ud
distance of 1´5d from the edge of the loaded area. : for 1:6vc < v < 2vc
Other perimeters can be taken at distances which are 0:87 f yv
multiples of 0´75d afterwards; see Fig. 1. For the peri- (3)

Perimeter un,ef

Perimeter un . 2d

, 2d

2d
d

d
2d

First shear
1.5d
First perimeter perimeter
of reinforcement ò0.75d At least 2 perimeters
BS 8110 of reinforcement

Failure zone 1

d/2
Face of loaded 0.75d
area of support Second shear perimeter
At least 2 perimeters
of reinforcement

Failure zone 2

Required area of
reinforcement should be
provided within
failure zone

12 13
Fig. 1. Critical perimeters according to BS 8110 and CEB Model Code
122 Magazine of Concrete Research, 1999, 51, No. 2
Downloaded by [ Imperial College London Library] on [24/04/17]. Copyright © ICE Publishing, all rights reserved.
Punching strength of slabs reinforced with steel I-section beams

where f yv is the P characteristic strength of the shear structed at a distance 2´0d outside the outermost shear
steel in N=mm2 , Asv is the area of the shear rein- reinforcement. When the circumferential distance be-
forcement in mm2 , á is the angle between the shear tween the shear steel elements in the outermost layer
reinforcement and the plane of the slab and v ÿ vc exceeds 2´0d, un,ef is limited as indicated in Fig. 1.
should not be taken as less than 0:4 N=mm2 . The CEB code limits the radial spacing of the shear
For the prediction of strengths, safety factors can be reinforcement to 0´75d and the spacing between the
removed and equation (3) can be rewritten as column face and innermost shear steel to âd, where â
is given by the ratio of the capacity of a slab without
Vu ˆ Vc ‡ Vs for Vs < 0:6Vc shear steel to the required capacity (â < 0:5).
  For comparisons with test results the safety factor to
10 Vs
V u ˆ Vc 1 ÿ ‡ Vs for 0:6Vc < Vs < 2Vc be removed from the steel stress is clear but that in-
7 2Vc
volved in the concrete term is not. It may be intended
Vu ˆ Vs for Vs > 2Vc to be 1´5 as for concrete in compression, but f ck ap-
(3a) pears only in a cube root and a value of 1´33 has been
assumed.
where
X
Vs ˆ Asv f yv sin á
Experimental investigation
p  p 
Vc ˆ 0:27 3 (100r f cu ) 4 (400=d) (3b) The slabs tested were 200 mm thick with side lengths
equal to 3000 mm. The central columns (or loaded
The third expression might be replaced by Vu ˆ 2Vc areas) were 200 mm square. The load was applied up-
but this is not very clear in the code. wards from a jack placed at the centre of the slab and
The use of shear steel is accepted only in slabs acting on a 200 mm square steel plate (Fig. 2).
200 mm or more thick and should be distributed evenly All the slabs were cast in plywood forms with the
on at least two perimeters within 1´5d from the column. same concrete mix proportions, made up of rapid-hard-
The circumferential distance between shear steel ele- ening Portland cement, natural sand and irregular gravel
lents is limited to 1´5d. aggregates with a maximum size of 20 mm (in the
13
The CEB-FIP Model Code MC90 suggests that for proportions 1´0:2´0:3´0 by weight, with w=c ˆ 0:5). The
interior flab-slab connections subjected to symmetric concrete was compacted by internal poker vibrators.
load the punching resistance should be checked in three
zones: (a) immediately adjacent to the column, (b)
within the shear reinforcement region and (c) outside 3000
the shear reinforcement region. 1400 200 1400
PLAN
For the zone immediately adjacent to the column, the 75 2750 75
design force (Fsd ) is limited by
150 3 150 3 20
813

Fsd < u0 d(0:5 f cd2 ) (4)


1500
458 458458

where u0 is the length of the periphery of the load or


3000

column, d is the effective depth and f cd2 is the design


compressive resistance for cracked concrete.
1500

For the zone within the shear reinforcement region,


813

d
Fsd < 0:09î(100r f ck )1=3 u1 d ‡ 1:5 Asv f yd sin á (5)
sr
p Slab Column
where î ˆ 1 ‡ (200=d), d is in millimetres, u1 is the
300 300

length of the control perimeter constructed at a distance


200

2´0d from the column face or loaded area, Asv is the


Jack
area of shear reinforcement in a layer around the col-
umn, sr is the radial spacing of the layers of shear Spreader
reinforcement, f yd is the design yield strength (ˆ f y = Frame supporting
1:15 . 300 N=mm2 ), á is the angle between the shear slab prior to loading
reinforcement and the plane of the slab and f ck is the
characteristic concrete strength (cylinder).
For the zone outside the shear steel region, Strong floor

Fsd < 0:12î(100r f ck )1=3 un,ef d (6)


where un,ef is the effective length of a perimeter con- Fig. 2. Set-up and dimensions (mm) of specimens
Magazine of Concrete Research, 1999, 51, No. 2 123
Downloaded by [ Imperial College London Library] on [24/04/17]. Copyright © ICE Publishing, all rights reserved.
Gomes and Regan

After casting, the slabs were covered with wet sacking was a cross of double straight lines of shear reinforce-
and polythene sheeting for about three days. Fourteen ment placed orthogonally to each face of the column.
150 mm cubes and three 150 3 300 mm cylinders per The second type of arrangement, radial, was used in
slab were taken for the control of concrete resistance. slabs 6±11 except for slab 9. The arrangement used in
Similar flexural reinforcement was used in all the slab 9 was also radial, but, as it had more layers and it
slabs. The yield and ultimate strengths of the top flex- was intended to avoid excessive circumferential spac-
ural steel were 680 N=mm2 and 810 N=mm2, respec- ings, intermediate radial lines were added from the
tively, for slabs 1, 1A and 2 and 670 N=mm2 and 758 sixth layer. Fig. 4 presents details of the shear steel.
N=mm2 , respectively, for other slabs. The anchorage U- Tensile tests of the shear reinforcement were made in
bars used in all slabs had yield and ultimate strengths the vertical direction ( y-axis) of the I-sections. The
equal to 650 N=mm2 and 791 N=mm2, respectively. results of these tests were 430 N=mm2 for the yield
A relatively high percentage and strength of flexural ( f y ) and 691 N=mm2 for the ultimate strength ( f u ).
steel were used to allow a reasonable shear punching The main measurements made were of the strains on
resistance improvement with the addition of shear rein- the shear reinforcement. A pair of FLK-6-11 electrical
forcement, prior to the intervention of flexural failure. strain gauges was fixed vertically at midheight of the
Slabs 1 and 1A were cast with a column at the steel and monitored by a Solartron 35300-Orion data-
centre. This column was extended 300 mm above and logging system. Every shear element of slabs 2±6 was
below the slab. The steel used in these columns con- instrumented. Slabs 7 and 8 had only one line of shear
sisted of four T25 bars (area ˆ 490:97 mm2 ), 750 mm reinforcement with strain gauges. The remaining slabs
long, and seven T8 stirrups (area ˆ 50:27 mm2 ). The had at least four lines of shear reinforcement with
other ten slabs, without columns, were loaded through strain gauges.
200 mm square steel plates at their centres. Fig. 3 The deflections of all slabs were measured by means
shows the flexural steel for all the slabs. of at least eight dial gauges, graduated in units 0:01
Small pieces of universal I-beam sections were used mm, positioned in one straight line crossing the column
as shear reinforcement. After strain gauges had been centre. No crack width measurements were taken, but
fixed to them, the shear elements were wired to pairs crack development was traced with a marker pen on
of 4 mm bars to maintain the correct spacing between the top of the slabs according to the load stages.
them and avoid movements during casting. A similar general test procedure was used for all
Three types of shear reinforcement arrangement were slabs. Each test took about 2 h until failure. After the
used in the tests. The first, used in slabs 2, 3, 4 and 5, jack picked up the slab, zero readings of the dial
gauges and strain gauges were taken and then the loads
were applied in steps of 25±100 kN. At each stage, the
load was held constant for about 5 min, while the read-
ings were taken and the cracks marked.
3000 mm 31 K 16 both ways/31 K 12 (anchor) both ways

Top reinforcement

Results
Crack pattern, and mode and load of failure
Radial cracks spreading from the centre of the slab
and circumferential cracks surrounding the loaded area
Bottom
were the basic crack pattern which appeared in the
reinforcement
tests. The first visible cracks appeared at approximately
200 kN in all slabs. Under increasing load the radial
cracks progressed to the edges while further circumfer-
ential cracks developed at greater radii.
All the slabs failed by punching. In general, the top-
3000 mm 21 T8 both ways side shape of the punching failure was circular. Slab 5
Detail of anchor bar K 12 had a slightly different shape of failure surface, prob-
674 ably due to the shear reinforcement distribution
152
300 200 300

674 adopted. The shear failure crack was not visible in any
of the tests until failure.
Figure 5 shows a sketch of the rupture surfaces and
Column stirrups Table 1 summarizes the loads and modes of failure of
Column reinforcement
4 3 T25 length 750 mm 7 3 T8 length 968 mm all the slabs, with their principal characteristics.
Slabs 1 and 1A, without shear steel, failed at 560 kN
Fig. 3. Flexural reinforcement of all slabs (dimensions in and 587 kN, respectively. The failure loads of slabs 2,
millimetres) 10 and 11 were 693 kN, 800 kN and 907 kN, respec-
124 Magazine of Concrete Research, 1999, 51, No. 2
Downloaded by [ Imperial College London Library] on [24/04/17]. Copyright © ICE Publishing, all rights reserved.
Punching strength of slabs reinforced with steel I-section beams

Slabs 2 and 3 with 2 layers RADIAL Slab 6 with 4 layers


CROSS DOUBLE 4 with 3 layers PATTERN 8 with 6 layers
LINE PATTERN 5 with 4 layers 7,10,11 with 5 layers
1400 458
Spacing: mm Spacing: mm
s1 5 141 s1 5 139
s6
s
s 5 255 4 3 80
2 s1 4 20
s2 5 200 s2
s 458
s3 5 262 s2 6
s3 5 368 160 200 s4 5 323
s4 5 481 20 s5 5 385 s9 s6
s4 6 3 80
s6 5 447 308
Centre I-sections
s6 5 300
column 20 160 20 s7 5 342 158
1400 s8 5 383 158
s9 5 425 308
SLAB 9 WITH
9 LAYERS

1400 200 1400 1400 200 1400

SHEAR REINFORCEMENT

I-sections S S (mm) 5 6.0 for slabs 2 and 10


S
8.0 3 and 11
7.9
10.7 4

16.7 5 and 6
177.8 162.0 4.7
24.1 7 and 8

7.9 25.0 for the five innermost


and 16.7 mm for the other layers
101.6 of slab 9

Radial direction CROSS-SECTION

Wired 178 200


Bottom flexural
reinforcement
15
∅ 4 mm

Fig. 4. Details of the shear reinforcement (dimensions in millimetres)

tively, and the failure surfaces of these slabs crossed Deflection


the shear reinforcement elements. The rupture in slab 3 The deflection profiles of all the slabs were very
occurred at 773 kN and the failure surface was not well nearly symmetric, and along the slab radius the profile
defined. was, in general, almost linear from the early stages
The remaining six slabs presented failure surfaces until the last measurements. There was apparently no
outside the shear reinforcement region. Slabs 4 and 5 influence of the type of failure mechanism on the
failed at the same load of 853 kN. Slab 5, with the linearity of the deflection profiles. Fig. 6 shows the
greatest circumferential spacing of the shear elements, deflections of slabs 1, 2, 6 and 9 from an early loading
seemed to have a different shape of failure surface. In a stage until the last load at which measurements were
plan section it appeared to be an octagon. made, while Fig. 7 presents load-versus-deflection
Slabs 6±9 had increasing numbers of layers of rad- curves for the innermost measured point for most of
ially arranged shear reinforcement. Slab 6, with the the slabs.
same total area of shear steel as slab 5, failed at The smallest final deflections were obviously those
1040 kN, while slabs 7 and 8 carried 1120 kN and of the slabs without shear reinforcement (slabs 1 and
1200 kN, respectively. Some concrete crushing on the 1A), while slab 9, with the most shear reinforcement,
bottom surface near the loaded area appeared in slab 9 had the highest final deflection and failure load. The
at failure, which occurred at 1227 kN (near the flexural final deflection and failure load appear to increase
failure load, 1250 kN, estimated by Johansen's yield together, but the deflections for equal loads were smal-
line theory). ler for the slabs which were more highly reinforced for
Magazine of Concrete Research, 1999, 51, No. 2 125
Downloaded by [ Imperial College London Library] on [24/04/17]. Copyright © ICE Publishing, all rights reserved.
Gomes and Regan
Shear reinforcement Failure
distribution Failure surface load: kN Slab 2 15.0 Deflection: mm
Without SR Slab 1
— 1 560 667 kN
640 kN10.0
Without SR
— 1 A 587 533 kN 480 kN

Loaded area Ase fy / layer 5 97.3 kN 5.0


CDL 2 693
373 kN 213 kN
5 129.3 kN Column
CDL 3 773
5 172.9 kN 1200 900 600 200 0 200 600 900 1200
CDL 4 853 Distance from centre: mm
5 270.2 kN
CDL 5 853
5 270.2 kN
Radial 6 1040 36.0 Deflection: mm
Slab 9
5 389.1 kN Slab 6 1173 kN
Radial 7 1120
5 389.1 kN
1120 kN 24.0 1013 kN
Radial 8 1200
5 404.2 kN .
640 kN 12 0 853 kN
‘Radial’ 9 1227
5 97.3 kN Column
373 kN
Radial 10 800
5 129.3 kN 1200 900 600 200 0 200 600 900 1200
Radial 11 907 Distance from centre: mm

Fig. 6. Deflection of slabs 1, 2, 6 and 9


Fig. 5. Sketches of the rupture surfaces (CDL, cross double
line; SR, shear reinforcement)
Load:
kN
Table 1. Ultimate loads and failure modes of the slabs 1200 Slab 8 Slab 9
P
Slab f cu : d: Asv f y =layer: Vu : Mode of Slab 7
N=mm2 mm kN kN failure 1000
Slab 6
1 50´3 159 Ð 560 Punching 4
5
1A 51´4 159 Ð 587 Punching 10
3
2 43´1 153 97´3 693 Within 2
3 49´0 158 129´3 773 Within/outside Slab 1 185
500
4 40´1 159 172´9 853 Outside
5 43´4 159 270´2 853 Outside
6 46´7 159 270´2 1040 Outside
7 42´3 159 389´1 1120 Outside
8 42´6 159 389´1 1200 Outside 100
9 50´0 159 404´2 1227 Outside
4 8 12 16 20 24 28 32 36
10 44´2 154 97´3 800 Within
11 43´2 154 129´3 907 Within Deflection: mm

 `Within' and `outside' mean failure surface within or outside the


Fig. 7. Load±deflection curves for the innermost measured
shear reinforcement region, respectively. point (deflection relative to slab edge)

shear. It also can be seen in Fig. 7 that beyond a certain most all the shear steel yielded, while in slab 5 only
limit only a small gain of load is achieved for a consid- one element in the third layer reached a yield strain and
erable increase of deflection. The failure load of slab 9 none in the fourth layer yielded.
(1227 kN) was 2´2 times that of slab 1, while the Slab 7, with five layers, had only one shear rein-
deflection was increased 4´5 times, and the failure load forcement line instrumented, and the innermost two
enhancement from slab 8 to slab 9 was only 2%, while elements reached yield. The third element had a strain
the deflection was increased by around 19%. near yield. Slab 8, with six layers and the same yield
force per layer as slab 7, developed yield in the third
Shear reinforcement strains and fourth layers. The added layer improved the failure
All the shear reinforcement elements of slabs 2 and load by 7% and made more elements yield.
3 yielded before failure. Twenty-one of all 24 elements Slabs 10 and 11, with five layers of shear reinforce-
in slab 4 also yielded. Slabs 5 and 6 had four layers ment, had almost all shear elements up to the third
and the same areas but different distributions of shear layer yield before failure. Fig. 8 shows a sketch of the
reinforcement. In slab 6, with a radial distribution, al- positions of the yielded shear elements for slabs 5±11.
126 Magazine of Concrete Research, 1999, 51, No. 2
Downloaded by [ Imperial College London Library] on [24/04/17]. Copyright © ICE Publishing, all rights reserved.
Punching strength of slabs reinforced with steel I-section beams

Slab 5 1
1
1
1
Slab 6
4
2
4 elements of the three innermost layers yielded before
4
3
3
3
4
4
4
4
4 4
4 failure of the slabs, as did all the shear elements of
11 34 42 22
4
4
slabs 2 and 3.
23 44 44 44
It can be seen that if only 75% of the potential con-
12 42 44 33 4 4
4
4 crete contribution were assumed, slabs 2 and 3 would
4 4
4
3
4
4 4
4 4 4 require more than two layers of shear steel and slabs 10
4
3
1
4
1 4 and 11 at least four shear reinforcement layers to reach
their actual failure loads.
If a full concrete contribution term calculated ac-
Slab 8 1 cording to the British code is considered it seems, for
Slab 10 Slab 10
2
4
1
3 slabs 2 and 3, that both layers of shear reinforcement,
4
4 1
4
4 3
2 spaced at 0´5d, can be taken into account. The results
3
Slab 9
4 4
4
4
4
4 for slabs 10 and 11, with five layers each, suggest that
4 4
three layers cooperate, which corresponds to a 20±258
1 22 4 23 3 4 4 4 4 3 21 11 344 4 4321 inclination of the failure surface.
4 4 4 4 44
4 Slab 7 4 4
24
4
3
3 4
3 1
3
4
4
4
4
1
Codes
2 3 3 1
4 2
4
Slab 11 1 Slab 11 The radial arrangement of reinforcement used in
4
4 4 Slab 9 some of the tests does not comply fully with the recom-
4 2
1 mendations of BS 8110. Therefore two perimeters were
considered in treating failures outside the reinforced
Fig. 8. Sketch of positions of yielded shear elements for slabs
zone, the suggested square and a perimeter defined by
5±11: 1, less than 0´6 3 yield strain; 2, between 0´6 and 0´8
3 yield strain; 3, between 0´8 and 1´0 3 yield strain; 4, the centre lines of the outermost studs plus 0´75d.
yielded Furthermore, the third layer of the shear steel in the
authors' tests was positioned at the limit of the second
and third zones defined by the British code, and hence
Discussion the failure load was calculated for two cases, assuming
that the failure surface crosses either two or three layers
Shear steel and concrete contributions of shear steel. The limit on f cu of 40 N=mm2 and the
An analysis of the contributions from the shear steel limit on the circumferential spacing of the shear rein-
and the concrete to the resistances of slabs 2, 10 and forcement were ignored. The strength limit dictated by
11, which failed by punching at failure surfaces within the nominal shear at the periphery of the column and
the shear reinforcement region, and slab 3, with its load has also been disregarded. If it were taken into
failure on the borderline between being within and account, half the slabs would be governed by it and
being outside the shear reinforcement region, is shown ratios of experimental to calculated strengths of up to
in Table 2. The concrete contribution is taken as equal 1´45 would be obtained. Table 3 presents a comparison
to the characteristic punching resistance of a slab with- of the loads and modes of failure in the authors' tests
12
out shear reinforcement, calculated according to BS with BS 8110.
12
8110 (Vck ). The concrete contributions (failure loads) The best results obtained with the British code for
of slabs 1 and 1A are included for comparison. In slabs the authors' tests occur when three layers of studs
10 and 11 only four lines of shear steel were instrumen- instead of two layers are regarded as crossed by the
ted, and with the exception of one in the third layer, all failure surface, and when the suggested British square
perimeter placed 0´75d from the last layer is used
rather than the chamfered perimeter. The ratios of ex-
perimental and calculated strengths are satisfactory but
the failure mode is incorrectly predicted in several
Table 2. Shear steel and concrete contributions
cases. In part this may be due to some of the circumfer-
Vu ÿ 0:75Vck Vu ÿ Vck 8nAsv f y ‡ Vck ential spacings of the shear elements having been great-
Slab Vck : er than 1´5d. It may also be that the second part of
8Asv f y 8Asv f y Vu
kN
equation (3) is overconservative.
1 575 Ð Ð 1´03 Slabs 10 and 11, with five layers of shear steel,
1A 579 Ð Ð 0´99
failed at a surface within the shear reinforcement
2 516 3´15 1´82 1´03
3 564 2´71 1´62 1´06 region. The actual failure loads were respectively 12%
10 526 4´17 2´82 1´02 and 17% higher than the predicted loads calculated on
11 523 3´98 2´97 1´00 the assumption that only two layers of shear reinforce-
 n ˆ 2 for slabs 2 and 3, and 3 for slabs 10 and 11. ment worked together. When three layers are consid-
Vu ˆ test failure load. ered to work together, the differences between the
Vck is obtained from equation (2) using u ˆ 4c ‡ 12d. actual and predicted failure loads drop to 2% on the
Magazine of Concrete Research, 1999, 51, No. 2 127
Downloaded by [ Imperial College London Library] on [24/04/17]. Copyright © ICE Publishing, all rights reserved.
Gomes and Regan

Table 3. Comparisons of failure loads and modes of failure in authors' tests with BS 8110 
12

Vtest Vtest Vtest Vtest Vtest


Slab Vtest : Actual mode of failure
kN V (1)
k V (2)
k V (3)
k V (4)
k V (5)
k
Predicted Actual
1 560 Ð Ð 0´97 Ð 0´97 Punching Punching
1A 587 Ð Ð 1´01 Ð 1´01 Punching Punching
2 693 0´93 0´93 1´18 1´46 1´18 Outside Within
3 773 0´94 0´94 1´21 1´51 1´21 Outside Within/Outside
4 853 0´99 0´93 1´18 1´50 1´18 Outside Outside
5 853 0´91 0´84 0´98 1´27 0´98 Outside Outside
6 1040 1´09 1´00 1´17 1´50 1´17 Outside Outside
7 1120 1´12 1´03 1´13 1´45 1´13 Outside Outside
8 1200 1´20 1´11 1´07 1´38 1´11 Within Outside
9 1227 1´17 1´07 0´77 0´99 1´07 Within Outside
10 800 1´12 0´98 0´82 1´05 0´98 Within Within
11 907 1´17 1´06 0´94 1´21 1´06 Within Within
 V (1) : using equation (3) and assuming that the failure surface crosses not more than two layers of studs.
k
V (2)
k : using equation (3) and assuming that the failure surface crosses not more than three layers of studs.
V (3)
k ˆ vc ud, considering a square perimeter 0´75d from the last layer of studs, or with u ˆ 4c ‡ 12d for slabs 1 and 1A.
V (4)
k ˆ vc ud, considering a chamfered perimeter defined by the centre lines of the outermost studs plus 0´75d.
V (5) (2)
k ˆ lesser of V k and V k .
(3)
2
Limit of f cu . 40 N=mm and limit on circumferential spacing of shear reinforcement disregarded.

unsafe side for slab 10 and 6% on the safe side for 1´16. The predicted failure load of slab 5, using the
slab 11. limit of 3´0d for the circumferential spacing, is almost
13
The CEB model code with ãm for concrete taken the same (850 kN) as the actual failure load (853 kN).
as 1´33 seems to give satisfactory results (Table 4). The This slab had four layers of shear reinforcement distrib-
predicted modes of failure for all slabs agree with the uted in double cross-lines around the column. The
actual ones. The results are on the safe side (slabs 7 circumferential spacing of the fourth layer was 3´0d.
and 8 with Vtest =Vcalc ˆ 1:29 and 1´43, respectively). If If ãm for concrete were taken as 1´5, the results most
the circumferential spacing of the outermost shear steel affected would be those for the slabs without shear
is limited to 3´0d, instead of 2´0d as used by the code, reinforcement, for which Vtest =Vk would be reduced to
the ratios Vtest =Vcalc of slabs 7 and 8 reduce to 1´19 and 0´90 and 0´94.

Table 4. Comparisons of failure loads and modes of failure in authors' tests with CEB Model Code 
13

Vtest
Slab V (1)
k : V (2)
k : V (3)
k : Vk Vtest Mode of failure
Vk
kN kN kN kN kN
Predicted Actual
1 Ð Ð 549 549 560 1´02 Punching Punching
1A Ð Ð 553 553 587 1´06 Punching Punching
2 1092 565 649 565 693 1´23 Within Within
3 1253 794 708 708 773 1´09 Outside Within/Outside
4 1068 901 712 712 853 1´20 Outside Outside
5 1148 1203 731 731 853 1´17 Outside Outside
6 1214 1213 871 871 1040 1´19 Outside Outside
7 1115 1556 842 842 1120 1´29 Outside Outside
8 1124 1557 844 842 1200 1´43 Outside Outside
9 1282 1624 1140 1140 1227 1´08 Outside Outside
10 1123 669 809 669 800 1´20 Within Within
11 1102 763 803 763 907 1´19 Within Within
 V (1) : failure load calculated for the zone immediately adjacent to the column or loaded area.
k
V (2)
k : failure load calculated for the zone in which the shear steel is placed.
V (3)
k : failure load calculated for the zone outside the shear steel.
All these failure loads were calculated considering the actual strengths of the materials. The coefficients 0´09 in equation (5) and 0´12 in
equation (6) were changed to 0´12 and 0´16, respectively, to correspond to the characteristic strength ÿãm ˆ 1:33 for concrete. The limit of 2´0d
on the circumferential spacing of shear steel has been observed in the calculation of V (3)
k Ðsee Fig. 1.

128 Magazine of Concrete Research, 1999, 51, No. 2


Downloaded by [ Imperial College London Library] on [24/04/17]. Copyright © ICE Publishing, all rights reserved.
Punching strength of slabs reinforced with steel I-section beams

Conclusions in square zones around the column used by the British


code may be easier to achieve in a flat slab but may
The punching failure load of a reinforced concrete
not make optimum use of the reinforcement. The maxi-
slab may be at least doubled with the use of I-beam
mum value of the concrete cube strength allowed by
offcut steel sections as shear reinforcement. In the tests
the code (40 N=mm2 ) may be increased up to at least
many shear elements yielded and did not present any
50 N=mm2 , and the strength limit imposed by the shear
problem with anchorage. If the shear reinforcement
at the periphery of the column or load could be disre-
force per layer is kept constant for all layers, sensible
garded or replaced by one with a higher stress as in the
spacing criteria are adopted and the same concrete 13
CEB Model Code. With these modifications the
strength is used throughout the slab, there are basically 12
strength predictions of BS 8110 are satisfactory al-
two types of mechanism of failure: (a) with the failure
though the failure mode is incorrectly predicted in
surface reaching the compressed face of the slab at the
some cases.
column and (b) with the failure surface outside the 13
In general, the CEB recommendations give reason-
shear reinforcement region.
able results. The predicted failure mechanisms agreed
The mechanism of failure with the failure surface at
with those observed and the predicted failure loads
the column but not crossing any shear reinforcement
were generally only slightly conservative. The limit of
can be avoided by the use of an appropriate limit for
2d for the maximum spacing between the outermost
the spacing between the innermost reinforcement layer
shear reinforcement elements used to determines the
and the column face. Placing the innermost shear re-
effective perimeter could be increased.
inforcement not more than 0´5d from the column face
has been shown to be satisfactory to avoid failure in
this region for loads up to twice the resistance of a
similar slab without shear reinforcement. References
The resistance to failure with a failure surface at the 1. TALBOT A. N. Reinforced Concrete Wall Footings and Column
column and crossing the shear reinforcement is a func- Footings. University of Illinois, Engineering Experiment Station,
tion of the concrete contribution plus the steel contribu- Urbana, 1913, Bulletin 67.
2. ELSTNER R. and HOGNESTAD E. Shearing strength of reinforced
tion. The steel contribution comes only from the shear concrete slabs. ACI Journal, 1956, 28, No. 1, 29±58.
elements crossed by the failure surface. As this failure 3. KINNUNEN S. and NYLANDER H. Punching of Concrete Slabs With-
surface has a minimum inclination of about 258, only out Shear Reinforcement. Institutionen foÈr Byggnadsstatik Kun-
the shear steel placed up to (d ÿ x)cot 258 from the gliga Tekniska HoÈgskolan, Stockholm, 1960, Meddelande 38.
column face and with enough embedment may be ex- 4. KINNUNEN S. Punching of Concrete Slabs with Two-way Rein-
forcement with Special Reference to Dowel Effect and Deviation
pected to contribute. of Reinforcement from Polar Symmetry. Institutionen foÈr Byg-
The mechanism of failure with the failure surface gnaddsstatik Kungliga Tekniska HoÈgskolan, Stockholm, 1963,
outside the shear reinforcement region is a function of Meddelande 41.
the distribution of the shear reinforcement. A slab with 5. REGAN P. E. and BRAESTRUP M. W. Punching Shear in Reinforced
an inadequate shear steel distribution may fail much Concrete: a State of Art Report. Comite Euro International du
BeÂton, Lausanne, 1985, Bulletin d'Information 168.
earlier than a slab with less but better distributed shear 6. GOMES R. B. Punching Resistance of Reinforced Concrete Flat
reinforcement. A radial distribution of I-beam offcuts Slabs with Shear Reinforcement. PhD thesis, University of West-
presented better results than a cross-line. minster, London, 1991.
For the shear reinforcement elements in a region 7. GRAF O. Tests of reinforced concrete slabs under concentrated
around the column to work together, their radial and load applied near one support (Versuche uÈber die widerstands-
faÈhigkeit von eisenbetonplatten unter konsentrierter last nahe ei-
circumferential spacings should be limited. The limits nem Auflager). Deutscher Ausschuss fuÈr Eisenbeton, 1933, 73,
remain to be researched more accurately, but they can 1±16.
be taken, on the safe side, as 0´5d for radial spacings 8. WHEELER W. H. Thin flat-slab floors prove rigid under test.
and 2´5d for the circumferential spacing of the outer- Engineering News-Record, 1936, 116, No. 2, 49±50.
most layer. 9. AMERICAN CONCRETE INSTITUTE COMMITTEE 318. Building Code
Requirements for Reinforced Concrete. American Concrete Insti-
In general, the pattern of flexural cracks of the slabs tute, Detroit, 1971, ACI-318.7.
did not change with either the mechanism of failure or 10. DILGER W. and GHALI A. Shear reinforcement for concrete slabs.
the failure load. Radial cracks appeared, spreading from Proceedings of the ASCE, 1981, 107, No. ST12, 2403±2430.
the slab centre, and circumferential cracks enveloping 11. REGAN P. E. Shear combs, reinforcement against punching. The
the column (or loaded area) occurred on the top side of Structural Engineer, 1985, 63B, No. 4, 76±84.
12. BRITISH STANDARDS INSTITUTION. Structural Use of Concrete. BSI,
the slab, Milton Keynes, 1985, amended 1993, BS 8110.
The distribution of deflection along the slab radius is 13. CEB-FIP. CEB-FIP Model Code 1990. CEB, Lausanne, 1991,
in general almost linear during all loading stages, and Bulletin d'Information 203±205.
the slab rotates about the column region. Apparently
there is no interference of the failure mechanism with
the linearity of the deflection profile. Discussion contributions on this paper should reach the editor by
The distribution of the shear reinforcement elements 29 October 1999

Magazine of Concrete Research, 1999, 51, No. 2 129


Downloaded by [ Imperial College London Library] on [24/04/17]. Copyright © ICE Publishing, all rights reserved.

You might also like