Professional Documents
Culture Documents
1992 Chana & Desai
1992 Chana & Desai
1992 Chana & Desai
Paper
Synopsis
This paper examines the design methodsf o r providing
resistance against punching shear, generally concerning the
strength requirementf o r transfer of the load from a flatslab
to the supporting column or from a column to its base. The
BSI committee B/525/2 (CSB39 as it was then called) had
expressed concern about the present design method given in
BS 81 10: Part 1: 1985 f o r providing resistance against
punching shear in reinforced concrete slabs. A s a result, an
examination was carried out, based on a test programme
funded jointly by the Department of Environment and the
British Cement Association (BCA). The conclusions of this
examination have assisted in reviewing the punching shear
clauses given in BS 8110: Part 1: 1985*. The results of this Flexural Triaxial Failure
exercise have also been used f o r recommending a revision to reinforcement stress state cone
@lo p
the clauses given in the Eurocode 'DD ENV 1992-1-1: 1992
Eurocode 2: Design of Concrete Structures Part I (General Fig 1. Punching shear failure mode
rules and rules f o r buildings)'. This BSI document is expected
to be available soon, and it will include a National
Application Document providing information and additional shearcapacityof the slab up to a limit corresponding to a specified
recommendations necessary f o r its use in the UK during its maximum value ofv. This procedure was continued progessively away from
prestandard (ENV) stage. the loaded area in steps of 0.75 h, until v was found to be less than v,. The
procedure is still followed in the current edition of BS 5400: Part 4: 19904,
Introduction the Code of Practice for concrete bridges.
The term 'punching shear' is associated withthe resistance of a slab when BS 81 10adopts asimilar approach, with the exception of usingd i n place
subjected to a concentratedload acting over an area,small in comparison of h for theshear perimeter which is ofa rectangular shapeinstead of one
with the dimensions of the slab itself. This is distinctly different from the with rounded corners. This is expressed in the equations given below, derived
shear force action on a linear member such as a beam, and thus requires from BS 8110 to give the resistance of the concrete cross-section (Vc).
a different conceptual model and a corresponding design method.
In the UK, design methodsfor flatslabs and forchecking punching shear
capacity have developed in parallel with the increasing popularity of flat
y f 2 [pp
p:Y?
slab construction. The benefits, such as the elimination of intrusions of V,= w p x c [ 400
F] :L ud
down-stand beams and increased usable space, are obvious and areattractive y m
in the design ofa variety of buildings, including offices, stores, and carparks. where
The resistance to punching shear isan important considerationin flat slab y, = 1.25 . . . .(la)
design and, recently, various prefabricated systems have been developed p < 3.0 . . . .(lb)
to provide shear reinforcement'. However, this paper is concerned solely f,,< 40N/mm2 . . . .(lc)
with the traditionalmethod of reinforcing slabs by providing links and the 400
corresponding design rules in BS 81 lo2.
- 2 1.0 . . . .(ld)
d
U is the shear perimeter
BS 8110 design method
It is helpful to relate the design methods to the failure modeshown in Fig The shear contribution of the reinforcement, V, is given by the sum of
1. CP 110: Part I: 19723gavedetaileddesign recommendationsfor all the links within the critical perimeter multiplied by the design stress in
punching shear. These included a concept of failure on the inclined face the links. Hence
of an inverted truncated cone or pyramid. This solid figure was formed V, = 'A,, fyv/Ym . . . .(2)
with the loaded area as onebase and the other on the far face of the slab, V, 2 {V - V,} ud . . . . (2a)
an area enclosed by a 'shear perimeter', at a distance of 1.5 h away from
the boundary of the loaded area, h being the overall depth of the slab. where
The shear stress resultingfrom theloaded area was assumedto be constant d 2 200mm
across the effective depth of slab (d) and along this shear perimeter (U). y, = 1.15
The distance of 1.5h determined the first critical section for checking the CA,, is the total area of vertical reinforcement links (in mm2) withinthe
capacity of the section, or the design shear stress (v) acting on the critical zone determined by the critical shear perimeter
perimeter (ud). This was chosen so that the same permissible design concrete Minimum CA,, to correspond to V, 2 0.4 ud . . . .(2b)
shear stress values (v,) as for beams could be adopted with an identical f,, is the characteristic strength of shear reinforcement (in N/mmz ,
partial safety factor (y,). The concrete shear resistance, v,, was a function 460N/mm2)
of the percentage of tension reinforcement ( p ) , the characteristic cube The maximum possible shear stress that can be supported is given by
strength (&,), and a factor related to the depth of the slab. If v exceeded v,=, < 0.8 K,)" or 5 N/mm2 on column face . . . .(2c)
v, andtheslab was at least 200 mm thick, shearreinforcementwas
necessary (with a certainminimum requirement) to enhance the punching There is one major difference, however, in the provision of links. CPZZO
required provisionof all the links on the critical perimeter and a similar
* The amendment to the punching shear clauses is included in the draftAmendment amount on a parallel perimeter at a distance of 0.75h. BS 81 10, on the
no. 3, document no. 92112045 DC, issued by BSI for public comment. other hand, requires the links to be distributed evenly in a zonewithin the
critical perimeter on at least two perimeters. This has been a subject of '& ;. . ' , , o.
0.:'. : . ::*g
controversy, as the provisions in BS 8110 can result in half the number
of links given by C P l 10. The BS 81 10 design method has been criticised
on the grounds that the rules are optimistic and that no justification for
the change from CP110 has ever been published. It should be noted that
BS 5400: Part 4: 1990 gives design rules for punching that are similar to
CP110.
The final draft of Eurocode ENV EC2: Part 1 (Generalrules and rules
for buildings)' adopts a similar approach to BS 81 10, except for variations
such as the shape and limitation on the length of critical perimeter and Main steel Extra T12 hanger bars
the limitation on the maximum punching shear capacity of a slab to 1.6 Fig 2. Shear reinforcement in slabs
times the capacity of the concrete section alone. (This limitation is replaced
by the one corresponding to eqn. (2c) in the UK National Application
Document which will accompany the ENV EC2for its use inthe UK.) Also,
ENV EC2 rules generally give results for v, very similar to those obtained
from theBS 8110 formulae with percentages of tension steel ( p ) around 1TO.
The BCA/DoE test programme was undertaken to verify whether there
is any needto modify the BS 8110 approach. Althoughthere is no evidence
of any failure ofa building attributable to the shortcomings ofthe punching
shear capacity of a flat slab, it was considered necessary to examine the
level of safety implied bythe present designmethod, adoptingan approach
without any preconceptions, and to take a fresh look at the fundamental
principles.
t
Fig 4. Conventional punching shear specimen
Reaction block
Hydraulic jack
can be mobilised, and consider it to be something ‘in hand’. Perhaps this Anchor
approach would indirectly maskother deficiencies, whichis not desirable.
The presenttesting programme hasbeencarried out using the
conventional punching shear test specimens describedlater, which do not
account for mobilisation of membrane action. An additional DoE/BCA
test programmeto investigate effects of membrane action on full-scale slabs
is being carriedout separately. The results are expected to be available later
this
year, and it is hoped that thiswill
lead to further design
recommendations for slabs where membrane action can be relied on.
7 9 9
Discussion
Fig 11 shows the test data plotted as V/V, against V,/V,
where
Fig 7. Failure mode
V is the failure load (normalised for f, = 40 N/mm2)
V, = v, U d, i.e. design concrete shear resistance to BS 8110
V, = C A,, O.87fp, i.e. design resistance due to all links within a 1.5
d perimeter, calculated according to BS 8110
The BS 8110 approach for design is to consider
v, = v, + v,
which can be rearranged to give
%=l+- VL .. . .(3)
VC VC
where VDis the design shear resistance.
The currentBS 8110 design values which can beexpressed as a straight
line are also shown on the figure. .Clearly, BS 81 10 is unsafe for values
of VL/Vcgreater than 0.8.For VL/Vcless than 0.8, the BS 8110 approach
is satisfactory. It should be noted that, for V,/V, = 0 (i.e. no links), the
average value of V,/V, from many test results is approximately 1.25.
The rearrangement of links within the first failure zone makes no
significant difference to the failure load, and thereis no indication that
Fig 8. Failure mode vlewea from lop OJ Slao any arrangement is markedly superior (compare results for specimens 5
to 7). Similarly, improvements in the anchorage of links to pass around
aN the tension reinforcement in specimen 3 leads onlyto amarginal increase
Test arrangement and method in failure load (compare the results for specimen 2 with specimen 3).
The first arrangement is shown inFig 5 and a slab under test in Fig 6. The Thetest results indicatethattheconcreteterm, V,, andthe link
slab specimens were supported on a 300 or 400 mm square block in the contribution, V,, cannot be considered as being simply additive. As the
centre representing a column. Load was applied by means of hydraulic jacks applied shear increases ((V,/V,) increasing), eitherorboth of these
at eight locations on thecircumference of a circle of diameter 2.4 m, acting contributions reduce.
through loadcells and prestressing cables. The hydraulic jacks were linked In general terms, the design shear resistance can be expressed as
to a common supply so that the force on each cable was the same. V, =K , V, + K2 V, .. . .(3)
The load was generally applied in 10equal increments. At each load stage
where K , and K2 are factors
the crack pattern was recorded, along with the deflections.
(G1) which are functions of V,/V,.
Results The concrete contribution K , V, is assumed to be as follows:
All the specimens appear to have failed by punching, with the failureplane
passing through the links in the slab.Figs 7 and 8 show typical modes of K , V, = V, when V, = V, and reduces linearly to
failure for the specimens.
K , V, = P V, when V, = 2Vc
The sequence of cracking on all the specimens wassimilar. Radial cracks
formed in the middle of the slab, extending gradually to the edges. Some Hence for V, < V, 2V,,
circumferential cracks thendeveloped prior to a punching shear failure.
Typical development of cracking on the top faceof the slab is shown in
K , V, = (2-P) - (1 - P ) VD/VcI VC ... .(4)
Fig 9. The link contribution is expressed as K, V,. These can be considered
2.2 - /
1.0 -
-
V
VC
l.6t /L
.o
''O r /
VL l VC
Fig 10. Section through slab after failure Fig 11. Test results compared with BS 8110 design
to perform in two ways. They provideresistance through the development This equation indicates a constant concrete contribution and reduced
a link
of tensile stress across the shear crack and they provide confinement of contributionas VD/V, increases. It islikely thatthe 'pure' concrete
the concrete. The link resistance should increase with quantity of links and contribution reduces as V,/V, becomes greaterthan 1 owing to the
hence is directly proportional to V, (aVJ.However, the test data show progressive cracking of concrete making the concrete cross-section less
that the effectiveness is steadily reduced with increasing shear force, V,. effective in resisting the shear stresses. However, this is compensated by
This reduction can be considered as being a function of the ratio VD/V,. the confinement action of the link contribution and, for simplicity, the
Further research is needed to understand and verify these factors. The total concrete contribution can considered
be as remaining constant equal
contribution of links can be summed up as V, [a - B VD/V,l,where 01 to v,.
and B are non-dimensional coefficients. The term [3 - 1.25 VD/V,l can be considered as a 'link effectiveness
m e ultimate shear capacity of a slab reinforced with vertical links then
is factor', comparing the contribution of the links with the BS 8110 formula.
given by the equation The value of this factor is tabulated below in relation to VD/V,:
Link
. . . .(5) effectiveness
'D/ C'
factor
. . . .(6) 1 1.75
1.2 1.5
1.4 1.25
From the test data inFig 13, a reasonablelower bound is given by values 1.6 1.o
of V,/V, of 0,0.6 and 2.0 corresponding to VD/V,of 1.0, 1.6, and 2.0, 1.8 0.75
respectively. Inserting these values into eqn. (6) gives 2.0 0.5
a = 3 (2-4
P = 1.25 ( 2 - P ) In practice, minimumlinks will be required for VD/V, less than around
1.4. For VD/V,equal to 1.6, the effectiveness of links is identical with
and the design equation becomes
BS 8110. For VD/Vcgreater than 1.6, the link effectiveness reduces
VD = VC + V, [3 - 1.25 VD/Vcl . . ..(7) linearly and at VD/V,equal to 2, the links are only half as effective as
NOTES:
(1) f,, is the average value of cubes cast and cured alongside the specimen.
(2) Failure load includes self-weight of the specimen.
(3) Failure load for 40 N/mm2 is calculated by multiplying the failure load by (M/&.)".
(4) BS 8110 design strength includes the 7, factors.
assumed in BS 81 10, i.e. twice the number of links are required. Fig 12 (NOTE: for vertical links sincr equals unity, being the angle between the
compares the calculated amount of shear reinforcement for themodified shear reinforcement and theplane of the slab). It is also recommendedthat
formula and the current BS 81 10 approach. the maximum shear capacity is limited to 2 V,.
;l;i
l an acceptable margin of safety.It is shownthat thedesign rules inBS 8110
18
for shearforcesless than 1.6 V, aresatisfactory.Also,no changeis
proposed to the design of slabs without shear reinforcement where a vast
l
1.6 amount of test data shows excellent correlation withthe BS 8110 equation.
l It should be noted that the effects of membrane action are ignored in
the design rules. Significant increases in shear capacity have been observed
1.4 Modified equation in slabs when membrane action can be mobilised. Research is currently
v, = v, +V, (3-1.25 VD )
- under way at BCA to quantify this effect and propose design guidance,
VC
taking into account the enhancement due to membrane action.
1.2
l
I References
-
VL
Fig 13. Comparison of test results with modified equation and proposed revision to BS 8110