Magazine of Concrete Research, 1997, 49, No. 178, Mar.

, 55-65

Diagonal cracking and shear strength of

reinforced concrete beams
J.-F? Zhang*

Technical University of Denmark

The shear failure of non-shear-reinforcedconcrete beams with normalshear span ratios is observed to be
governed in general by the formation of acriticaldiagonal crack. Underthe hypothesisthat the cracking of
concrete introduces potential yield lines which may be more dangerous than the ones found by the usual plastic
theoty, a physical explanation isgiven for this phenomenon andaway to estimate the shearcapacity of
reinforced concrete beams, based on the theoty ofplasticity, is described. The theoretical calculations are shown
to be in fairly goodagreementwithtestresults jrom alarge number of experiments carried out by previous
investigators.

Notation TO shear strength

uncracked
of (microcracked)
concrete
length of shear span 5" shear
strength of cracked
(macrocracked)
con-
shear spanratio, h beingthe depth of thebeam crete
uniaxialcompressivestrength of concrete zc, shearstrengthdetermined by diagonalcracking
effective
compressive strength of uncracked load
(microcracked) concrete defined as f:o = v0f c @ mechanical degreeof longitudinalreinforcement,
effective
compressive
strength of cracked
@ = AsfY/bhfc
(macrocracked) concrete defined
as = v&
= vsv0.L
crackingmoment Introduction
diagonal cracking load
ultimateload or load carrying capacity The problem of shearbehaviourandshearstrength
startingpositionofcriticaldiagonal crack, i.e. of reinforced concrete beams without shear reinforce-
horizontaldistancefromthe end ofacritical ment has been perplexing researchers for many years.
diagonal crack to the edge of the support platen Ageneral review may be found.''2Recentdecades
inclinationofrelativedisplacementwithrespect haveseengreatprogress both in theoryandapplica-
toyieldline tion; however, it fact aisthat
fundamentally
a
inclination of yield line comprehensivesheartheory is still missing.
effectiveness
factor
uncracked
for (micro- Among themanyattempts,theplastictheoryfor
cracked) concrete non-shear-reinforcedconcrete beams andslabs was
sliding reductionfactor duetocracking developed in the 1970s by Nielsenand
effectivenessfactorfor cracked (macrocracked) It turnedoutthat in order togetagreementwith
concrete experimentalresults, an effectivenessfactor v had to
normal stress be introduced which besides the usual and easily
shear stress understandableparameterscontainedtheinfluenceof
theshearspan ratio. The physicalmeaning of this
* Department of StructuralEngineering, TechnicalUniversity of functionhasnever been reallyunderstood.
Denmark,Bygning 118, 2800 Lyngby, Denmark. In experiments it hasbeenobservedthattheshear
(MCR 541) Paperreceived 2 January 1996; accepted 14 February failureofnon-shear-reinforcedconcretebeamswith
I996 normal
a shear span ratioischaracterized by the
Delivered by ICEVirtualLibrary.com to: 55
IP: 160.94.45.157
On: Fri, 12 Apr 2013 11:53:430024-9831 0 1997 Thomas Telford Services Ltd
Zhang

formation critical
of
a diagonal crack. Since the and are collected together.13 The failure mechanism of
cracking reduces the
sliding
strength
of
concrete a beam (Fig. 2) may be briefly summarized as follows.
substantially, this phenomenon suggests
that
the Cracks will appear in the shear span when the load
crackingofconcrete may introducepotentialyield isincreased. Owing to the presenceofshearstresses,
lineswhichmightbemore dangerousthanthe ones they bend towardstheaxis of the beam. Diagonal
predicted by the usual plastic theory. cracks are thus formed. The development of
the
Under thishypothesis, the load carrying capacity diagonal cracks stops at a certain load level when the
of reinforced concretebeams may be calculated by last diagonal crack suddenly propagates through nearly
plastictheorywith an effectivenessfactor being a the total depthofthe beam. Othersecondarycracks
function of quite natural
parameters suchas the dueto
stress
redistribution may also appear. The
uniaxialcompressivestrength of concrete, the rein- specimeneithercollapsessimultaneously with the ap-
forcementratio and a size effectparameter,witha pearance of this diagonal crack or continues to sustain
reductionfactortakingtheinfluence of thecracking higher load until the concrete in the compression zone
into account. The hypothesis described here leads to a is crushed.
theory covering both
ordinary
reinforcedconcrete The last diagonal crack is referred to as the critical
beams andprestressedbeams. diagonal crack and the load at which it occurs is
termedthediagonalcracking load. The load at which
the specimen collapses is the ultimate load or the load
Shear failure mechanism of beams with carrying capacity. The cracking load PCrcorresponding
to the formation of a diagonal crack as shown in Fig.
diagonal cracks
3(a) is a function of the starting position of the crack
Tests simply
on supported non-shear-reinforced x. This function is qualitatively depicted as Curve 1 in
concretebeamsunder concentratedloadshave been Fig. 4.
reportedinnumerous papers.Fig. 1 showstypical Measurements of the relative displacements along a
crack patternsof such beams which have failed in diagonalcrackcarriedout ~ 5 ) show
by M ~ t t o n i '(Fig.
shear.7 Similarpictures may be found that at failure the relative displacements along
a

7
Fig. I . Crack patterns of concrete beams after shear failure
56 Magazine qf Concr-ete
Research, 1997, 49, No. 178
Delivered by ICEVirtualLibrary.com to:
IP: 160.94.45.157
On: Fri, 12 Apr 2013 11:53:43
 Diagonal cracking and shear strength

failure criterion.Whenthecombination of stresses in

concrete exceeds the failure criterion, a sliding failure
(incontrasttoseparation failure) takesplacealonga
line, which is termedayield line, that is, a line of
discontinuity in displacements. The load found by the
VIx a x b plastic theory is the ultimate load or the load carrying
H 15
capacity.
Fig. 2. Schematic illustration of beam failing in shear Intheupperboundapproach of plastic theory,a
simple shear failure mechanism is assumed by setting
upayieldline asillustrated in Fig. 3(b)wherethe
relative displacement U is at an angle of a to the yield
I L ) o " a l I l , line.Theyield
boundsolution
line whichgivesthelowestupper
PO for beamsofconcretewithzero
tensile strength is the one running from the support to
theloadshown asadashedline in thefigure.
Forabeamwithdiagonalcracks,theultimateload
Fig. 3. (a) Diagonal crack pattern and (b) approximate yield corresponding to ayieldlinefollowingadiagonal
lineused in the plastic solution crackmaybelowerthan Po. For simplicity, theyield
lineandthediagonalcrackaredepictedas straight
lines in Fig. 3(b). The shear load P, corresponding to
'T such a yield line as a function of the
oftheyieldline
starting position
x is drawn as Curve 2 in Fig. 4,
curve 2
assumingtheupperendoftheyield line is fixed.
Shear faliure with the Curve 1 and Curve 2 cross at point S where PCr= P,.
formation of a critical How theshear failure happenscannowbeex-
plained.Whentheloadisincreased,diagonalcracks
appear in theshearspan.Owingprimarilytothe
x Curve 1 interlockaction of
aggregates, shear stresses are
Cracking load P,,
transferredalongthe
cracked sections.When the
development of the diagonal cracks reaches the point
where the sliding strength along the crack is exceeded,
thediagonalcrackistransformed into ayield line,
slidingoccursalongthe line, andthebeam fails in
Fig. 4. Crackingload Pc,. andultimateload P,, plotted shear.
against x
In Fig. 4 it canbeseen that whenthecrack is
located tothe right of point S, thecrackingloadis

I" lowerthantheshearcapacitydeterminedbyayield
line along this crack.Theyield
formed at this position,
dangerous.Whenthecrack
and
this
crack
line thuscannot be
is not
is located to the left of
point S, the shear capacity is lowerthanthecracking
load hencenoshearfailurecanoccurbecausea
critical diagonalcrackhasnotbeencreated at this
position.Only at point S doesthebeam fail in shear
as
consequence
a of
the
formation of
a critical
diagonalcrack.
Consequently, if the load at which the critical
diagonalcrack is created and
the
ultimateload
Fig. S. Measurement of displacements along a diagonal corresponding to yield a linefollowingthiscrack
crack at dflerent loading stages14 canbedetermined,boththepositionofthe critical
diagonal
crack
and
the
shear
capacitymay
be
diagonalcrackarenotperpendiculartothepathof calculated.
crack as in a pure separation failure but form an angle In experimentstheyield line doesnotalways fol-
smallerthan 90" withthecrack. In plastic theory low thecriticaldiagonalcrack all the way through.
terms,thismeans that slidingorsheardisplacements It mighthappen that theuppertipofthe critical
occuralongthecrack. diagonal crack ends in the compression zone under the
In plastictheory,concrete is assumed to bea rigid loadingplateor at theedgeoftheloading plate,
plastic material obeying the modified Mohr-Coulomb whereastheyield line normallyends at theedge of
Magazine qf Concrete Research, 1997, 49, No. 178
Delivered by ICEVirtualLibrary.com to:
57
IP: 160.94.45.157
On: Fri, 12 Apr 2013 11:53:43
 Zhang

the loading plate. For simplicity, we asssume that both furtherincreasing the load point B’ on the curveof
the yield line and the critical diagonal crack end at the thecracking load orpoint D’ on the curveof the
edge of the loadingplate. shear load might be reached. If point B’ is reached
Since the cracks are formed witha finite spacing, beforepoint D‘, the potential diagonal crack originat-
the theoreticalcriticaldiagonalcrackwillnot always ing from point B will be formed, and a sudden failure
be the one really observed in experiments. What may will take place where the load drops down from point
happenisillustrated in Fig. 6. Supposeadiagonal B’ to C’. If point D’ is reached before point B’ (in
crack originating from point A has been formed in the Fig. 6 D’ is depictedhigher than B’), the potential
beam.The theoreticalpositionoftheyieldlinelies diagonal crack from B cannot be developed, and a soft
between this crack and potential a diagonalcrack failure will takeplace alongtheexistingdiagonal
initiating from point B which is at a cracking distance crackinitiatedfrompoint A with the force slowly
from point A. The load requiredforproducingthe fallingwhendeformationsare increased.
existing crack at A can be read off the curve ofthe Hence the reason why the shearfailure is some-
cracking load which corresponds to point A‘. A shear timesverybrittle and sometimesmore soft is ex-
failurecannottake place in thebeam for this load plained. In both cases the real ultimate load will be
because theloadcausingashearfailure,correspond- higher than the theoretical one. Depending on whether
ing to point D’ on the curveof theshear load is point B’ or D’ is reachedfirst, the observed position
higher than the cracking load at point A‘. To produce of the yield line will be closer to or further from the
diagonal
a crack
originating from point B at the support, respectively. The failure line will follow the
bottom face requiresa load which can be read off at diagonal crack frompointA if point D‘ is reached
point B’ ofthe cracking load curve.The shear load first, andtheultimate load will be higher than the
corresponding to this diagonal crack can be read off at diagonalcracking load, otherwise the diagonal crack
C’. In the particular case considered in Fig. 6, C’ lies fromBwill be created andtheultimateload will be
higherthan A‘. When increasingthe load correspond- equal to the diagonalcrackingload.
ingtopoint A’, point C’ on thecurve of theshear
load is reached. Since no diagonalcrackexistsatthis
position, and since the shear capacity corresponding to
Cracking moment
theyieldline going throughuncrackedconcrete is
higherthantheshearloadrelatedtopoint C’, no Basically, the determination of the cracking moment
shear failure will take place for
this
load.
When is afracturemechanicsproblem.The load tostarta
crack and to develop it through the shear zone should
be calculatedusing the methodsdeveloped in fracture
mechanics.Onlya few attempts have been described
Shear load in theliterature.Thedifficulty lies in the fact that
corresponding
loadCracking to
corresponding to along
failure
shear even fora given starting point of the crack,the path
diagonalformation
a the crack of the crack is still unknown. It must be found by the
of a diagonal crack
calculation. One way of doing this is to use the crack
growthformuladeveloped in Reference 16 and solve
Theoretical it numerically. The crack tip, having reached a certain
failure load
position, will in the next step take the direction
Soft fallure determined by the maximum value of the stress
A’ Load
decreases slowly intensityfactorfor all directionsthrough that point.
- The stressintensityfactor may be calculatednumeri-
I Sudden
failure I
callyusingstandardfiniteelementprograms. To start
1 Load drops fromI
B to C’ I
I
I
thecrackgrowth, for instance from apointalong the
0 l
I X bottom side of the beam, a small notch may be
I I Options for real introduced.
I I critical diagonal
I I
However, in this paper we will use some very
crackn IP (Ti
simplemethods to calculatethecrackingmoment.
Accordingtostandardbending theory, themaximum
diagonal crack9’ crackingmoment at a flexurally crackedrectangular
Theoreticalcriticall /
Existing diagonal section may be determined by the elastic equivalent
crack
stressdistribution in Fig. 7(a) which gives
I r
0 A l M,, = iybh2ft (1)
Cracking distance
P
where y is aplasticityfactorofresistance, ft the
Fig. 6. Effect of crack spacing on the position of the criticul tensile strength of concrete, and h and b the depth and
diagonalcrackand on the ultimateload width of the beam respectively.
58 Magazine of Concrete Research, 1997, 49, No. 178
Delivered by ICEVirtualLibrary.com to:
IP: 160.94.45.157
On: Fri, 12 Apr 2013 11:53:43
 Diagonal cracking and shear strength

strength,isadopted.Expressing ft and f c in MPa,

k M 1.2, which is used in thefollowingcalculations.
Furtherreduction in tensilestrengthduetosize
effectsmight be takenintoaccount by a size effect
factor

where h is the depth of beam in metres. This formula

is basedonsplitting testsmade by Forsell.” The
Weibull root describing the size effect was found to be
0.27,which is hererounded upto 0.3. Thus,the
Fig. 7. (a) Elasticand (b) plasticequivalent stressdistri- tensile strength of concrete f t for a beam with depth h
bution in a frexurally cracked section at the maximum becomes
crackingmoment
213
f t = l.2($) s(h) (5)
M,, may alternatively be found by consideringa
fully plasticequivalentstressdistributionshown in
Fig.7(b),whichyields A modeltakingautomaticallyintoaccountthe size
effect is depicted in Fig. 9. Hereanequivalentplas-
M,, = ibh2f : (2)
tic tensilestress f t p isactingalonga‘processzone
Here, f: istheequivalentplastictensilestrength. By length’ h, of constantabsolutelength.Thisstress
comparing equations (1) and(2), it canbe seen that distribution might be imagined to be a good model of
*
.f, = f Y f t . the actual one at the maximum cracking moment MC,.
Analogously, thecrackingmomentforadiagonal When the deformationsare
increased beyond this
crackmight be determined by assuming similar a point,the h, zone will move up andthelengthunder
equivalentplasticstressdistribution as shown in Fig. the processzone will be the openvisiblecrack. By
8. Since thedirection of thetensilestress atcrack- simpleequilibriumequations, it is found that
ingisperpendiculartothe crack path, theinternal
moment along thecurvedcrackwithrespecttopoint
R equals that along thestraightchord AR with the
same tensilestressacting at right anglesto it. The By selectingsuitable values for ftpand h,, atensile
crackingmoment becomes strength as afunction of h similarto ( 5 ) may be
reproduced.
M,, = i.f:b~;, (3) In the followingsectionsthecrackingmomenthas
+
Here L A R = ,/(h2 x:) is the length of the chord AR. been calculatedusingequations (3)-(5).
For specimenswitharectangularcross-section of
standardlaboratorysize,thevalueof y is about 1.7,
thus f : zz 0.6ft.
Ultimate load according to plastic theory
For the tensilestrength ft. astandardexpression
ofthe type k(0.1f c ) 2 / 3 ,f c beingthecompressive Usingtheupperboundapproach of plastic theory,
shear
a failure
mechanismfor
non-shear-reinforced

l x
Fig. 8. Plasticequivalentstressdistribution
x
I
in adiagonally
I

Fig. 9. Modeltakingautomaticallyintoaccountthesize
crackedsectionatthemaximumcrackingmoment eflect
Magazine of Concrete Research, 1997, 49, No. 178 59
Delivered by ICEVirtualLibrary.com to:
IP: 160.94.45.157
On: Fri, 12 Apr 2013 11:53:43
Zhang

concrete beams is set up by assuming a straight yield Effectiveness factors and effective
line
followingapproximately the
critical
diagonal compressive strengths
crack(Fig. 10). Here, a is theangle of therelative
displacement U to the yield line. In the over-reinforced In the plastic theory of concrete, concrete is treated
case, U isperpendiculartothelongitudinalreinforce- as arigidplasticmaterial which isfar different from
ment,thus a = 90" - p. Neglecting the
tensile the realmaterialproperties.Therefore, as mentioned
strengthofconcrete,ayieldlinefollowingthedotted already, an effectivenessfactor v and an effective
line in Fig. 10, that is, approximately the one observed compressivestrength f,*= v,fc are introduced in the
in tests, will givethe same load carrying capacity. theory.
For normalbeams,thewidth is smallcompared TheCoulomb criterion usually applied in plastic
with the depth andthespan,thusaplanestressstate theory as theslidingfailurecriterionhasthe form
may be assumed. Under the modified Coulomb failure /TI = c -,m = c - otancp (12)
criterion with zero tension cut-off, the internal work of
thisfailuremechanismisI5 where 0 and T are thenormalstressandshearstress
respectively. The constant c is termed the cohesion, p
thecoefficientof friction and cp the friction angle.
The normalstress o ispositiveas tension.
Here f,*= v f c is the effective compressive strength, v Expressed in principalstresses, the sliding criterion
being the effectiveness factor. Since a = 90" - /3 and becomes
cot@ = ( a - x ) / h , wehave kol - 03 = fc (13)
where f c is the uniaxialcompressivestrength, 01 and
0 3 are the principal stresses, assuming 01 > 0 2 > 0 3 ,

The externalwork is

and
The work equation W , = W E yieldstheupperbound 2c cos cp
solution
fc 1 - sin cp

In the following
we
must
distinguish between
uncrackedconcreteandcrackedconcrete.The term
uncracked concrete
means
concrete without visible
xu being theaverageshearstressatfailure. cracks. However, it must be borne in mind that even if
In the
usual
plastic theory, withouttaking
the novisiblecracksare formed, the strength ofconcrete
diagonal cracks into consideration, the minimum upper might be reduced because of microcracking. Thus the
bound solution for beams with zero tensile strength of strength of uncrackedconcrete might be morecor-
concrete corresponds toayieldlinefromsupportto rectly termedthestrengthofmicrocrackedconcrete.
load whichgives In auniaxialcompressiontest,concrete is found to
be highlymicrocracked in the direction of theload.
These cracks are initiated because of the difference in
theelasticconstants between the cement paste and
aggregates. In theshearspanofa beam microcracks
areformed in thecompressivezone in the same way.
This
solution
referred
is to
as the usual plastic
Since the directions of the principal compressive
solution in the following.
stresses change
because
of the redistribution of
stresses during cracking in the tensile zone,arather
I f complicatedstate of microcracking will be the result.
At present, the strength reduction dueto this phe-
nomenon can only be taken into account in a very
rough way by an effectivenessfactor vo. The reduced
strengthofmicrocrackedconcreteis taken intoac-
I
count by areductionof the cohesion c of the virgin
Lo " a "Lo material assuming that the friction angle is unchanged.
The reducedcohesionisdesignated by c0 and the
Fig. 10. Shear failure mechanism by a yield line following reduced compressive strength ,f,*O = v,&, where v0 =
the critical diagonal crack C O / C , since from equation (13) it is seen that the
60 Muguzine qf Concrete Research, 1997, 49, No. 178
Delivered by ICEVirtualLibrary.com to:
IP: 160.94.45.157
On: Fri, 12 Apr 2013 11:53:43
 Diagonal cracking and shear strength

effectivecompressivestrengthisproportionaltothe v, M 0.67. For shear/bending which we aredealing

cohesionif q is constant.Thus with here, the value of v, is chosen as a constant 0.5.

Load carrying capacity

When a crack is formed, the tensile strength normal By momentequilibriumfor onepart ofthebeam
tothecrackpath becomes very smalland may be shown in Fig. 10, thediagonalcrackingloadisfound
neglected,butconsiderablecohesionstill exists be- by
cause of theaggregateinterlockaction.Thesliding
criterion along a cracked surface may be expressed as
V($LO + U) = Mcr (25)

JzJ= c’ - p’a = c’ - a t a n q ’ (1 7) where V = Pc,, LO beingthelengthoftheloading

platens.
Here c’ and cp’ arethecohesionandfriction angle Insertingequation (3) intoequation (25) yields
along thecrack respectively.
No substantial change in the friction
angle
of
normalgravelconcreteisobserved from uncrackedto
cracked concrete, but the cohesion of cracked concrete Here, z, istheaverageshearstressat the diagonal
is
reducedsignificantly.
Here we use reduction
a crackingload.
factor v, =c‘/co toshowthereductionofthe co- As demonstratedabovefor beams failing in shear
hesion due to cracking. Hencetheeffectivecompres- with a critical diagonal crack, the ultimate load equals
sive strength of cracked concrete may be expressed as the diagonal cracking load. Then, by putting tu = re,,
thestartingposition of thecriticaldiagonalcrack x
f,*= (18) is determined and
consequently the loadcarrying
We may summarize theexpressionsfortheeffective capacity can be calculated by equation (10) or(26).
compressivestrengths of uncrackedandcracked con-
crete respectively in the following way:

uncrackedconcrete:
Experimental verification
Only over-reinforced beams failing in shear
are
treated. The criterionforover-reinforcement is, ac-
crackedconcrete: cordingtothe usual plastic s0lution,3’~~
@2 jV,VO (27)
S,* = vsf:O = vsvo.fc,
v = vsvo (20)
where @ = A,j+/bhh. To facilitatethecomparison
Regarding V O , we will partly adhere totheempiri- with
test
results
the
shearstress z = V/bh is
calformulaeobtained in the usual plastic solution; normalized by being divided by vg fc. For beams with
putting concentratedloads,thevalueof I in equation (21) is
foundto be around 1.6. If the value of x found by
V0 = nflCrc)f2(h)f3(P) (21)
solving the equation tu = z, is negative,thenthe
Here, /z is aconstant depending on the loading type, value zero is assignedto x.
and f l , f 2 , f 3 are functionsof one variable as in- In Fig. 11 theresults of 280beams from 8 test
series9- I 1.18-22
dicated. The expressionsforthesefunctionsare: are compared with the calculated shear
strength. For comparisonthe usual plasticsolution
J’Icfc) =-
3.5
dfc
( 5 < f c < 60 MPa) cfc in MPa) (1 1) with f,*= fzo = v0fc isplottedas well. From
the figure it can be seen that thefollowingis true.
(22)
For beamswithshearspan ratio a / h largerthan
f i ( h ) = 0.27(1 + l/Jh) (0.08 < h < 0.7 m) approximately2, the test resultsare in rathergood
agreementwiththecalculation. As observed in ex-
( h in m) (23)
periments, in thisrangeofshearspans,theshear
+
f 3 ( p ) = 0 . 1 5 ~ 0.58 ( p < 4.5%) ( p in %) (24) failure is always accompanied by theformationofa
criticaldiagonalcrack.
The cohesion along crack a dependson
many When a / h 0.75,
visible
no critical
diagonal
parameters. Tests have shownthat when the initial cracks
initiating
from
bending are observed in
crack width is within a limited range, the value of c’ experiments. Therefore, the test results lie higher than
does not vary substantially.
Previous
analysis4
has the presentcalculation.Besides,whentheshear span
shown that for the specimens of the push-off type, v0 gets smaller and the width of the beam compared with
is about 0.7 and V,VO about0.45whichresults in theshearspanbecomes larger, theeffectofplane
Magazine of Concrete Research, 1997, 49, No. 178 by ICEVirtualLibrary.com to:
Delivered 61
IP: 160.94.45.157
On: Fri, 12 Apr 2013 11:53:43
Zhang

Range of critlcal
diagonal cracks
R
I

A Testsby Krefeld
0 Tests by Leonhardt
Y Testsby Cossio
*- X Tests by Kani
z + Tests by Mathey
0 Testsby Moody
l Testsby Morrow

1.2h (alh = 2)
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 X x
a/ h 0.98h (alh = 1.5)
0.81h (alh = 1)
Fig. 11. Comparison of theoretical shear strength as func-
Fig. 13. Possible positions q j critical diagonal crack for
tion of a / h and test results f o r beams with concentrated
beams with a l h < 2
loads

strain fields starts to

influencethe load carrying line RB is a yield line following a crack at a position
capacity substantially. The restraints imposed by plane wheretheshearstrengthequals that of the usual
strain fields in such beams will raise the ultimate load plasticsolution TO correspondingto a yield line in
considerably compared with that determined under the uncracked concrete from support to load. The possible
assumption of a plane stress field. This may be one of positionsof a yield linecoinciding with adiagonal
the reasons why the test results for beams with a / h 6 crack lie between RA and RB. If a diagonal crack can
0.75 lie even abovethe usual plasticsolution. be created between RA and RB, the yield line will
For beams with a / h between 0.75 and 2, the test follow thiscrack and the load carryingcapacity is
results almost fill the space between the usualplastic somewherebetweenthe usual plasticsolutionandthe
solution and the present calculation. In the calculation, presentsolution.
Otherwise,the yield line either
x is negativefor a / h < 2, so the diagonalcracking follows the existing crack closestto the support or
load T,, is lower than the ultimate load tu for a yield runs from supportto load, which yields an ultimate
linefollowing a diagonalcrack(Fig. 12). Thus the load closetothe usual plastic solution.Thedistance
critical
diagonal
crack
is
bound to
start
from the AB for a / h < 2 is shown in Fig. 13 for some typical
support(lineRA in Fig. 13) andtheshearstrength values of a / h . AB is found to be of the same order of
determined by thepresentcalculationcorrespondsto magnitude as the cracking distance, the mean cracking
point So in Fig. 12. Since cracksare formed with distance being
about 0.3-0.5h and the maximum
finite spacing, if thedistance between an existing distance being abouttwice as large. Hence any point
crack nearthesupport and the supportisnot long between A and B will be a possible origin of a crack
enoughto make possible a newcrack,the yield line andconsequentlythe load carryingcapacity maybe
will eitherfollowtheexistingcrack orgo through found anywhere between the usual plastic solution and
uncracked concrete fromsupport to load. In Fig. 13 thepresentsolution.
Othercracks may be createdas well in this range,
the most likely onebeing a splittingcrackrunning
from supporttoload.Theformationand the effect of
=t suchacrack will not be treated here in detail.
The comparison between the shear strength obtained
by the presentcalculation, real, and that obtained in
tests, xteSt,from thereferencesmentioned above with
a / h 2 2 is shown in Fig. 14. The mean value of the
ratio rtest/r,,l is 1.1 0, the standard deviation is 0.14
and the coefficientofvariation is 0.13.
If the reinforcement in a beam is totally unbonded,
then
the beam acts like a tied arch from the
beginning, and the failure is caused by the crushing
ofconcrete in thecompressivechord. No diagonal
cracks occurbeforefailure,andthe load carrying
capacity is higher than that with
bond.This is
Fig. 12. Diagonalcracking load and ultimate load as confirmed by t e ~ t s . ”In ~ theplastic theory, the force
function of x l h for different values of a / h in thelongitudinalreinforcement is assumedconstant,
62 Magazine of ConcreteResearch, 1997, 49, No. 178
Delivered by ICEVirtualLibrary.com to:
IP: 160.94.45.157
On: Fri, 12 Apr 2013 11:53:43
 Diagonal cracking and shear strength
5.0

4.0

3.0
5
2.0
n Testsby Krefeld
3 TestsbyLeondardt
x Testsby Cossio 1.o
X TestsbyKani
+ Testsby Mathey
0 TestsbyMoody 0.0
_7 TestsbyMorrow 0.0 1.0 2.0 3.0 4.0 5.0 6.0
T?
! TestsbyVan denBerg alh
O . O , , , ~ , , , , , , , ~ , , , , , , , , , , , , ,
0.0 0.5 1 .o 13 2.0 2.5 Fig. 15. Comparison o theposition of thediagonalcrack,
I$
r,,: MPa x l h , withtestresults

Fig. 14. Comparison of shearstrengthobtainedbytheore-

tical calctrlation and by tests f o r beamswithconcentrated rcr =
loads and with a f h > 2
I * (1 + [iao - X>/hI2)Q/h
(a/h + L o / h ) ( a o / h + t L o / h-) t ( a o / h + Lo/h)2
meaning no bond is takenintoconsideration. There-
(30)
fore, what is estimated by the usual plastic solution is
close to the load carrying capacity of beams with little By putting tu = rcr thestartingposition x of the
or no bond.13 criticaldiagonalcrack is calculatedandconsequently
In Fig. 15 the calculatedstartingposition ofthe the
load carrying capacity may be determined by
critical diagonalcrack is compared with test results. 18 eitherequation (29) orequation (30).
Themean crackingdistance L,, is calculatedaccord- I"
In the equations f ,* = v,vo f c and f are defined as
ing
to
Reference 23 and the maximum cracking beforeand
distance L,,,,, = 2Lc,.
yo , = 0.5
= 1 . 2 f l m f 2 ( h ) f 3 ( ~ )v, (31)
The startingpositionofthecriticaldiagonalcrack
for a / h 3 2 may be approximated by a0 is determined using an upper bound solution of the
usual plastic t h e ~ r y ,that
~ is,
X
- h 0.74(: - 2)

The shear strength of deep beams and corbels may be

determined by the same principles.
In this case x is found to be negative when a / h < 4.
Fig. 17 showsthecomparison of theoreticalcalcula-
tion and test resultsfor a / h 2 4.9.18 Themean value
Shear strength of beams subjected to of the ratio rtest/rcalis 1.1 1, thestandarddeviationis
0.1 1 and the coefficient of variation is 0.10.
uniform load
Shear failurepatternssimilarto those observed in
beamswithconcentratedloadsare observed in beams
subjected
to
uniform 1 0 a d . ~ . ' Under
~ the sameas- Shear strength of prestressed beams
sumptions as above,theshearcapacityfor an over- Criticaldiagonal cracks arealsoobserved in shear
reinforcednon-shear-reinforcedbeam such astheone failures of non-shear-reinforced prestressed beams.6
shown in Fig. 16 may be determined by

I *
Tu = y.fc
-_-
h h
It can be seen from Fig. 16 that ao determines the end
positionofthe yield lineatthetop face. a
X
By moment equilibriumforonepart of thebeam,
thediagonalcrackingload is found to be Fig. 16. Shear span of abeamsubjected to uniformload
Magazine of Concrete Research, 1997, 49, No. 178 63
Delivered by ICEVirtualLibrary.com to:
IP: 160.94.45.157
On: Fri, 12 Apr 2013 11:53:43
Zhang

Thehnctions f1, f2, f3 aredefined by equations

(22)-(24) and

The function f4 takes into account the enhancement of

the effectiveness factor by a prestressing force. Fig. 19
shows a com arison of the theoretical calculation with
test results. 23F4 Themeanvalue of the ratio Ztest/rc,I is
0.94,the standard deviation is 0.13 and the coefficient
of variation is 0.14.

0.0 0.5 1 .o 2.0

2.5 1.5
MPa Conclusions
Fig. 17. Comparison of theoreticalcalculationwithtest In this paper a theory based on the plastic theory is
results for beams with uniform load and with a l h 3 49'1s proposedtodescribethe failure mechanismand to
estimate thestrength of reinforcedconcretebeams
Basedonsimilarassumptions to thoseabove,the failing in shearowing to theformationofa critical
shear capacity for an over-reinforced prestressed beam diagonalcrack.From thecomparisonswith test re-
suchastheoneshown in Fig. 18 is sults, the followingconclusions may bedrawn.

( a ) The theory presented here, based on the hypothesis

that cracking in concrete introduces potential yield
lineswhichmightbemoredangerousthanthe
ones found in the usual plastic solution, leads to a
The
diagonal
cracking
momentdetermined
is as rational explanation of the shear failure mechanism
before by moment equilibrium which gives of non-shear-reinforced concrete beams.
( b ) Under the hypothesis that the ultimate load equals
thediagonalcrackingload for beams failing in
shear owing to the formation of a critical diagonal
crack,theposition ofthe critical diagonalcrack
Here use = Fse/bh istheaverageprestressingstress maybefoundandtheloadcarryingcapacitycan
over the whole section. consequently be determined.
This expression is found by taking into account the ( c ) Theprinciplesofthetheorymaybeapplied to
effectoftheprestressingforce in equation (25) as if beams under any load combinations.
thebeamweresubjected to anexternalnormalforce (4 The theory explains the difference in load carrying
equal in sizeandposition to theprestressing force. capacity
between beamswith
bond
and
those
As above, the starting
position
of
the critical
diagonalcrack is found by putting z, = tCT, andthe
loadcarryingcapacity is determinedconsequently.
In the equations f,"= v,vo f c and ,fr
are defined as
beforeand

& 0.0 0.5 1 .o 1.5 2.0 2.5

b. a b T& MPa

Fig. 18. Shear span of a prestressed beam subjected to Fig. 19. Comparison of theoretical calculation and test
concentrated loads results for prestressed beams with concentrated loads24'25
64 Magazine of Concrete Research, 1997, 49, No. 178
Delivered by ICEVirtualLibrary.com to:
IP: 160.94.45.157
On: Fri, 12 Apr 2013 11:53:43
 Diagonal cracking and shear strength

withoutbond. It also givesarationalexplanation andreinforced concrete. Technical University of Denmark,

of the effect of prestressing. Institute of Building Design, Report No. 123, 1977.
5. JENSENJ. F.,BRESTRUP M. W., BACHF. andNIELSENM. P.
Compared with the existing empirical methods, this Nogle Plasticitetsteoretiske Bj~lkelmninger.Technical Univer-
theory
offers
considerableadvantages. Firstly, by sity of Denmark, Structural Research Laboratory, Report R No.
101, 1978.
means ofthis theory, the shearcapacityofbeams
6. ROIKJERM,, NIELSENM. I?, BRESTRUPM. W. and BACHF.
withdiagonal crackscan be calculated more accu- Forskydningsfors0g med
Spmzdbetonbjcelker uden
Forskyd-
rately, and eventhepositionofthecriticaldiagonal ningsarmering. Technical University of Denmark, Structural
crack may be approximately determined. Secondly, Research Laboratory, Internal Report No. 157, 1977.
this theory leads to a more thorough understanding of 7. KANI G. N. J. How safe areour large reinforced concrete
the shear failure
mechanism. Thirdly, the
theory beams. 1 ACI, 1967, March, 128-141.
8. TAYLORR. Some shear tests on reinforced concrete beams
unifies the shear
analysis of reinforcedconcrete without shear reinforcement. Mag. Concr. Res., 1960, 12, No.
beams, deep beams,corbelsandprestressedbeams. 36, Nov.
It must be admittedthat some simplifications made 9. LEONHARDT F. and WALTHER R. Schubversuche an einfeldrigen
hereareratherrough.Moreresearch is needed in the Stahlbeton-balken mit und ohne Schubbewehrung. Deutscher
followingaspects. Ausschuss fur Stahlbeton. Heft 151, 1962.
IO. MATHEYR. G. and WATSTEIND. Shear strength of beams
without web reinforcement containing deformed bars of
The effectiveness factor for uncracked concrete YO
different yield strengths. 1 ACI, 1963, Feb., 183-207.
and the sliding reduction factor due to cracking v,. 11. K A N I G. N. J. The riddle of shear failure and its solution.
A more rational way of distinguishingthem and I ACI, 1964, Apr., 441-466.
determining theirvaluesmay lead to
better
a 12. TAYLORH. I? J. Shear stresses in reinforcedconcrete beams
understanding of the failure mechanism. without shear reinforcement. Cement and Concrete Association
Report TRA 407, 1968, Feb.
The crackingmoment and thediagonalcracking 13. ZHANGJ.-I? Strengthofcracked concrete. Part I - shear
load. By strictly following the methods of fracture strength of conventional reinforced concrete beams. deep beams,
mechanics, the cracking moment corresponding corbelsandprestressed reinforcedconcretebeamswithout
to theformation of adiagonalcrackmight be shear reinforcement. Technical University of Denmark, Depart-
calculated more precisely, which may yield a better ment of Structural Engineering, Series R, No. 3 1 I , 1994.
14. MUTTONI A. Die Anwendbarkeit derPlastizitatstheorie in der
estimate of theposition of thecriticaldiagonal
BemessungvonStahlbeton. Institut fur Baustatik und Kon-
crack and the load carrying capacity. struktion, ETH Zurich, Bericht Nr. 176. 1990, June.
The ultimate load. The ultimate load according to 15. NIELSEN M.P. Limitanalysisandconcrete plasticity. Prentice-
the plastic theory is determined here by assuming a Hall, Englewood Cliffs, 1984.
straight yield line which is seldom found in reality. 16. NIELSENM. P. An energy balance crack growth formula.
Bygningsstatiske Meddelelser, Nr. 3-4, Sept. 1990, 71 - 125.
More accurate results might be obtained by using a 17. FORSELLC. Schubfestigkeit und Schubbewehrung der Beton-
curved yield line. balken. Trans.RoyalInstitute of Technology, 78, Stockholm,
1954.
18. KREFELD W.J. and THURSTON C. W. Studies of the shear and
diagonal tension strength of simply supported reinforced
Acknowledgements concrete beams. 1 ACI, 1966, Apr., 451-475.
The work reported in this paperwas carried outat 19. COSSIOR. D.D. andSIESS C. P. Behaviour and strength in
shear of beams and frames without web reinforcement. 1 ACI,
theDepartment of StructuralEngineering, Technical 1960, Feb., 695-735.
University of Denmark, under
the
supervision of 20. MOODYK. G., VIESTI. M,, ELSTNER R. C. and HOGNESTAD E.
Professor Dr Techn. M. P. Nielsen. The author wishes Shear strength of reinforced concrete beams. Part 1 - tests of
to expresssincere appreciationtoProfessorNielsen simple beams. 1 ACI, 1954, Dec., 317-332.
for his inspiringadviceandencouragement.Financial 21. MORROWJ. and VIEST 1. M. Shear strength of reinforced
concrete frame members without web reinforcement. I ACI,
supportfromthe Danish Council forScientific and 1957, March, 833-868.
Technical Research (STVF) gratefully
is acknowl- 22. VAN DEN BERG F. J. Shear strength of reinforced concrete
edged. beams without web reinforcement. Part 2 - factors affecting
load at diagonal cracking. I ACI, 1962, Nov., 1587-1600.
23. FEDDERSEN B. andNIELSEN M. P. Revneteorier for Enaksede
Spcendingstilstande. Technical University of Denmark, Structur-
References al Research Laboratory, Report R No. 162, 1983.
24. SOZENM. A., ZWOYER E. M. and SIESSC. P. Strength in shear
1. BRESLER B. andMACGREGORJ. G. Review of concrete beams
in beamswithout web reinforcement. University of Illinois,
failing in shear. I ASCE, 1967, STI, Feb., 343-372.
Engineering Experiment Station Bulletin No.452, 1959.
2. RECANF? E. Research on shear: a benefit to humanity or a
25. K A R J. N. Shear strength of prestressed concrete beams without
waste of time? Struct.Engr, 1993, 71, No. 19, Oct., 337-347.
web reinforcement. Mag. Concr. Res., 1969, 21, No. 68, Sept.,
3. NIELSENM. P, BRESTRUP M. W., JENSEN B. C. and BACHF.
159-170.
Concrete plasticity beam shear - shear in joints - punching
shear. Special Publication, Danish Society for Structural
Science and Engineering, Structural Research Laboratory,
Technical University of Denmark, Lyngby, 1978. Discussion contributions on this paper should reach the editor by
4. JENSEN B. C. Some upplications c?f plastic analysistoplain 26 September 1997

Magazine of Concrete Research, 1997, 49, No. 178 by ICEVirtualLibrary.com to:

Delivered
65
IP: 160.94.45.157
On: Fri, 12 Apr 2013 11:53:43