Mechanism of shear transfer in a reinforced

concrete beam
Akin A. Olonisakin and Scott D.B. Alexander
Abstract: This paper presents an analysis of the results of five tests conducted on four reinforced concrete beams. The
tests were performed principally to investigate the mechanics of internal shear transfer in a transversely loaded concrete
beam with no shear reinforcement. Test specimens consisted of simply supported wide beams with steel flexural
reinforcement. The reinforcement for two of the beams was epoxy coated. The shear span to depth ratios were 2.93,
3.32, and 3.81. Measured strains on the reinforcement were used to divide the total shear into its beam and arching
action components. In all tests, beam and arching action shear transfer mechanisms were found to coexist. Apart from
that with the longest span, all tests ended with rupture of the concrete along a diagonal failure surface. It is concluded
that shear failure may be caused by a shift in the internal mechanics of shear transfer from beam action to arching
action. Because this shift may be initiated by the yielding of reinforcement, it can be associated with the formation of
a plastic hinge. There was no observed effect on the mechanics of shear transfer that could be attributed to epoxy
coating of the reinforcement.
Key words: arching action, beam action, one-way shear, shear transfer, reinforced concrete beam, bond forces, bar force
gradient.
Rsum : Cet article prsente une analyse des rsultats de cinq tests conduits sur quatre poutres en bton arm. Les
test taient principalement effectus pour tudier la mcanique du transfert de cisaillement interne dans une poutre en
bton charge transversalement sans renforcement de cisaillement. Les spcimens de test consistaient de larges poutres
supportes simplement avec un renforcement en flexion en acier. Le renforcement de deux des poutres tait enduit
d'poxyde. Les rapports trave de cisaillement/profondeur taient 2,93, 3,22 et 3,81. Les contraintes sur le renforcement
mesures ont t utilises pour diviser le cisaillement total en composantes d'action de poutre et d'action de cintrage.
Dans tous les tests, il a t trouv que les mcanismes de transfert de cisaillement de l'action de poutre et de l'action
de courbure coexistaient. Hormis celui avec la trave la plus longue, tous les tests se sont termins par une rupture du
bton le long d'une surface d'effondrement diagonale. Il est conclu que l'effondrement de cisaillement pourrait tre
caus par un changement de la mcanique interne du transfert de cisaillement d'une action de poutre une action de
cintrage. Parce que ce changement pourrait tre initi par le flchissement du renforcement, il peut tre associ la
formation d'une charnire plastique. Aucun effet, qui pourrait tre attribu l'enduit en poxyde, n'a t observ sur la
mcanique du transfert de cisaillement.
Mots cls : action de cintrage, action de poutre, cisaillement sens unique, transfert de cisaillement, poutre en bton
arm, forces de liaison, inclinaison de force de barre.
[Traduit par la Rdaction] O lonisakin and Alexander 817
Introduction
In a reinforced concrete one-way flexural member, ne-
glecting tension in concrete, the bending moment M is ex-
pressed as the product of the tensile steel force T and the
effective moment arm jd. One-way shear V is the gradient
of bending moment along the length of the member. That is,
[1] V
M
x
Tjd
x
jd
T
x
T
jd
x
= = = +
d
d
d
d
d
d
d
d
( ) ( )
The first component of eq. [1] is the shear resulting from
a gradient in steel tensile force on constant lever arm and is
carried by beam action. The beam action shear requires
bond forces between the concrete and the steel reinforce-
ment and is conveniently modeled by a critical nominal
shear stress. Beam action is characteristic of slender flexural
members (B-regions) and may be limited by yielding of the
reinforcement or by bond failure. The second component of
eq. [1] is the shear resulting from a constant steel tensile
force acting on a varying lever arm and is carried by internal
arching action. It requires only remote anchorage of the
reinforcement. Shear transfer by arching action predomi-
Can. J. Civ. Eng. 26: 810817 (1999) 1999 NRC Canada
810
Received December 14, 1998.
Revised manuscript accepted June 16, 1999.
A.A. Olonisakin and S.D.B. Alexander.
1
Department of
Civil & Environmental Engineering, University of Alberta,
Edmonton, AB T6G 2G7, Canada.
Written discussion of this article is welcomed and will be
received by the Editor until April 30, 2000 (address inside
front cover).
1
Author to whom all correspondence should be addressed
(e-mail: [email protected]).
nates in deep beams and regions adjacent to discontinuities
or disturbances in either the loading or the geometry of the
member (D-regions).
This paper reports tests on four wide-beam elements. The
beams were companion specimens to two slab-column con-
nection specimens, reported in Alexander et al. (1995). The
goal of all of these tests was to examine the mechanisms of
shear transfer by experimentally measuring some of the
components of eq. [1]. In addition to applied loads and reac-
tions, forces in the flexural reinforcement were estimated on
the basis of strain measurements and coupon test results. As
one of the shear mechanisms identified in eq. [1] requires
force gradient in the reinforcement, it was thought that fac-
tors affecting bond might also affect shear. Therefore, two of
the beam specimens were built with epoxy-coated reinforce-
ment while the remaining two were built with uncoated rein-
forcement.
Test details
The test specimens consisted of four concrete beams,
155 mm in thickness, 750 mm in width, and 1.4 m in length.
All beams were simply supported and reinforced with a bot-
tom mat of grade 400 No.15M bars spaced at 150 mm each
way. The main (longitudinal) bars were placed at an effec-
tive depth of 128 mm. The transverse reinforcement, present
in a prototype structure as shrinkage and temperature rein-
forcement, functioned in the test specimens mainly as crack
initiators. Figure 1 shows the geometry and reinforcement
details for a typical test specimen. All specimens had a rein-
forcing ratio of 1.04%. Test information and results are sum-
marized in Table 1.
The steel reinforcing bars for all the test specimens came
from the same batch with measured yield strength of
425 MPa. Reinforcement for two beams, CB1 and CB2, was
sent out to the industry to be epoxy-coated. The remaining
specimens, RB1 and RB2, were built with uncoated rein-
forcement. Despite the fact that all the reinforcement was
taken from the same lot, the epoxy-coated bars had a slightly
higher measured yield strength of 439 MPa after treatment.
The compressive cylinder strength of the concrete used for
all the test specimens was 32.5 MPa.
Tensile strains in the main bars were measured by three
rows of gauges located at a distance of 75, 225, and 375 mm
(sections 1, 2, and 3, respectively) on either side of the beam
centre-line. A total of 15 electrical resistance strain gauges
with nominal resistance of 120 and gauge length of 5 mm
were used for each test specimen. To minimize the effects of
tension stiffening, strain gauges were positioned at the
points of intersection of the longitudinal and transverse bars
where it was deemed most likely that the beam would crack.
The layout of the strain gauges is as shown in Fig. 1.
The loading and span dimensions for all the test beams
are as shown in Fig. 2. A central vertical load was applied
by means of a hydraulic jack operating between the beam
surface and a reaction frame. The load from the jack point
was measured with a commercial load cell. Two tests,
(a) and (b), were performed on specimen CB1. The remain-
ing three specimens were each tested once to failure. In
CB1(a), the jack load was transferred centrally to the beam
through a 38 75 HSS. In the rest of the tests, the jack load
was distributed to two central points on the beams by means
of a 200 mm steel channel section. All the beam specimens
had between 175 to 225 mm overhang on each side of the
simple supports to provide anchorage for the main bars.
Beam deflections were measured with a linear variable dif-
1999 NRC Canada
O lonisakin and Alexander 811
Mark
Span
Depth
f
y
(MPa)
Type of
reinforcement
V
max
(kN) Failure mode
CB1(a) 3.81 439 Epoxy coated 113 Bending
CB1(b) 3.32 439 Epoxy coated 129 Shear/compression
CB2 2.93 439 Epoxy coated 130 Shear/compression
RB1 3.32 425 Uncoated 123 Shear/compression
RB2 2.93 425 Uncoated 128 Shear/compression
Table 1. Description of tests.
interval
750
4 @
150
128
155
700 700
Line of
symmetry
7 @ 150
Symm.
75
225
375
1st
interval
2nd
Longitudinal gauges
Section
1
Section
2
Section
3
Fig. 1. Description of test specimens.
ferential transformer (LVDT). The slippage between the
steel and concrete was monitored during the tests with hori-
zontal LVDTs attached to the ends of the main bars.
At a reinforcing ratio of 1.04% and shear span to depth
ratios ranging from 2.93 to 3.81, the beams were expected to
behave as slender flexural members. Results of tests on
beams without stirrups, summarized in MacGregor (1997),
suggest for predicting shear failure, the simplified method of
shear design in Standard A23.3-94 (CSA 1994) works best
for reinforcing ratios of about 1%. At lower reinforcing ra-
tios, the simplified method produces scattered and generally
unconservative predictions of strength, while at higher rein-
forcing ratios the method tends to be a lower bound of the
test data.
Test results
The beam deflections and average strain readings in the
longitudinal bars recorded during the tests indicated reason-
able symmetry about the midspan of the beams. As ex-
pected, the transverse bars were effective in initiating cracks
at the gauge locations.
Figure 3 shows the central jack load (P) versus the central
deflection for all the test specimens. Apart from test CB1(a),
all tests resulted in the rupture of the beam along a diagonal
failure surface. Anchorage failure was ruled out because no
slippage between the steel and concrete was observed in the
displacements monitored by the horizontal LVDTs. The fail-
ures were judged to be shear/compression failures, although
in the case of specimen RB1 rupture took place shortly after
the formation of a folding mechanism. Figure 4 shows pho-
tos of the failed specimens. Test CB1(a) was stopped when
it became evident that the beam was developing a folding
failure. The loading geometry was adjusted and the speci-
men re-tested to failure as CB1(b).
Analysis of test results
Strain gauge readings were converted to bar forces using
the results of coupon tests of the reinforcement. Making use
of symmetry, there are five strain gauges located at each of
the three sections defined in Fig. 1. At every load step, the
bar forces determined for each set of five gauges were aver-
aged to produce experimental measures of the bar forces, T
1
,
T
2
, and T
3
, at the corresponding sections. The subscripts 1,
2, and 3 indicate the section number in Fig. 1.
From the average bar force values at sections 1, 2, and 3,
the bar force gradient was computed over the first and sec-
ond intervals at every load step. Figure 5 shows bar force
gradients over the second interval versus the central deflec-
tion for the test beams. CB1(b) with epoxy-coated reinforce-
ment and RB1 with uncoated reinforcement had similar
loading geometry. Both reached a limiting bar force gradient
of about 205 N/mm, after which the bar force gradient de-
clined. CB2 with epoxy-coated reinforcement and RB2 with
uncoated bars both reached maximum force gradients of
about 255 N/mm. The maximum force gradient measured
for CB1(a) was 228 N/mm.
In all cases, a decline in force gradient over a particular
interval is associated with the onset of yielding at the more
highly strained section defining that interval. However, well
in advance of any yielding, the force gradient over the inter-
val softens. This softening may be the result of a gradual de-
terioration of the bond between the steel bars and the
surrounding concrete. The authors were, however, surprised
that there was no obvious correlation between the magnitude
of the force gradient and the presence or absence of epoxy
coating. One possibility is that adhesion bond was not signif-
icant in these specimens. Alternatively, it may be that there
are factors other than bond that have a more significant ef-
fect on force gradient.
The values of force gradient measured in the beams com-
pare closely with those values for two-way plates of
200 N/mm reported by Alexander et al. (1995). This sug-
1999 NRC Canada
812 C an.J.C iv.Eng.Vol.26,1999
175 525 525 175
38 x 75 HSS
175 525 525 175
200 mm channel
225 475 475 225
200 mm channel
(a)
(b)
(c)
Fig. 2. Span dimensions and loading: (a) CB1(a);
(b) CB1(b) and RB1; and (c) CB2 and RB2.
0
50
100
150
200
250
300
0 2 4 6 8 10 12 14
Deflection (mm)
T
o
t
a
l
l
o
a
d
(
k
N
)
CB1(a)
CB1(b)
RB1
CB2
RB2
Fig. 3. Load versus deflection.
gests that there ought to be a fundamental link between one-
and two-way shear.
To examine the mechanics of shear transfer more closely,
it is first necessary to recast eq. [1] in terms that can be mea-
sured experimentally. Conditions at successive sections
along the beam are used to approximate the gradient terms
in eq. [1]. The bending moments at each section, M
1
, M
2
,
and M
3
, are calculated at every load step using load cell
readings and the known geometry of the simply supported
beam. For an interval between sections i and i + 1, eq. [1]
can be rewritten as
[2] V
M M
s
Tjd Tjd
s
i i i i
=

=

+ + 1 1
( ) ( )
At each section, the tensile force T is calculated from
strain gauge measurements. The effective internal moment
arm, jd, at each gauge location is estimated as follows:
[3] j d
M
T
1
1
1
= ; j d
M
T
2
2
2
= ; j d
M
T
3
3
3
=
A tacit assumption behind eqs. [2] and [3] is that flexural
tension in the concrete is insignificant. Ideally, this requires
that the beam be fully cracked and that the gauged sections
coincide with crack locations. Because transverse reinforc-
ing bars were placed at the gauged sections, cracks tended to
occur at the gauged sections. To establish that a section is
fully cracked is more difficult. Where flexural tension in the
concrete is significant, eq. [2] will produce an estimate of j
that is well in excess of unity. Figure 6 shows the effective
moment arm factor j obtained from eq. [3] plotted against
the central deflection for tests CB1(a), CB2, RB1, and RB2.
Prior to cracking, the contribution of concrete tension to the
moment of resistance of the beam results in a value of j that
is not reasonable, exceeding unity. With significant flexural
cracking, the magnitude of j drops to a value between 0.5
and 0.9.
Not surprisingly, Fig. 6 shows that section 1 cracks first,
with cracking at section 2 following closely. In all cases,
section 2 appears to be fully cracked at a deflection of 3 to
4 mm. Cracking at section 3 is delayed until well after
1999 NRC Canada
O lonisakin and Alexander 813
Fig. 4. Photographs of failed specimens: (a) CB1 and CB2; and (b) RB1 and RB2.
0
50
100
150
200
250
300
0 2 4 6 8 10 12 14
Deflection (mm)
B
a
r
f
o
r
c
e
g
r
a
d
i
e
n
t
(
N
/
m
m
)
CB1(a)
CB1(b)
RB1
RB2 CB2
Fig. 5. Measured bar force gradients.
cracking at section 2. Section 3 cannot be considered fully
cracked until the final 2 or 3 mm of deflection.
The significance of the preceding discussion is to limit the
application of eqs. [2] and [3]. To apply eqs. [2] and [3] to
the first interval requires full cracking at sections 1 and 2.
This means that the analysis at the first interval is valid at
deflections greater than about 3 to 4 mm. For the second in-
terval the analysis is valid only near the end of each test,
where full cracking at section 3 is a reasonable assumption.
As was the case for eq. [1], eq. [2] can be broken down
into two components, namely beam and arching action.
[4] V V V = +
a b
Expanding the right hand side of eq. [2] leads to the fol-
lowing expressions for beam and arching action shear.
[5] V
j j d T T n
s
i i i
b
=
+

+ +
( ) ( )
1 1 1
2
[6] V
j j n T T dn
s
i i i
a
=
+

+ +
( ) ( )
1 1 1
2
The term V is the total vertical shear; V
a
and V
b
are the
components of the total vertical shear transferred by arching
action and beam action, respectively; n is the total number of
bars at each of the gauge location (n = 5 in all tests); and s is
the length of beam in shear between the two sections.
It should be noted that the values of j, T, and M are not in-
dependent of each other. They are constrained to be stati-
cally consistent. While eqs. [5] and [6] provide a criterion
for dividing the applied shear into its two fundamental com-
ponents on the basis of strain measurements, those two com-
ponents, namely V
a
and V
b
, will always sum to the applied
vertical shear V.
Figures 7 to 11 show plots of total vertical shear (P / 2),
arching shear (V
a
) and beam shear (V
b
) against the central de-
flection for all the test beams. In all graphs, the point of first
measured yielding of at least one bar at the more highly
1999 NRC Canada
814 C an.J.C iv.Eng.Vol.26,1999
0
0.5
1
1.5
2
0 2 4 6 8
Deflection (mm)
j
(c)
0
0.5
1
1.5
2
0 2 4 6 8
Deflection (mm)
j
(d)
0
0.5
1
1.5
2
0 5 10 15
Deflection (mm)
j
Section 1
Section 2
Section 3
(a)
0
0.5
1
1.5
2
0 2 4 6 8 10
Deflection (mm)
j
(b)
Fig. 6. Effective moment arm factor: (a) CB1(a); (b) RB1; (c) CB2; and (d) RB2.
strained section is indicated with an asterisk. Figure 8 shows
results for test CB1(b) over only the second interval because
the strain gauges at section 1 did not survive the unloading
sequence at the end of test CB1(a).
Comparing the (a) and (b) parts of Figs. 7, 9, 10, and 11,
it can be seen that the contribution of arching action is more
significant in the first interval than in the second interval.
This is attributed to two factors. First, the first interval is in
closer proximity to a point of load application than is the
second interval. This would tend to favour arching action
over beam action. Second, strains are generally greater in the
first interval with yielding of the reinforcement typical.
Since yielding of reinforcement limits force gradient, the
spread of yielding is parasitic on beam action shear. Arching
action must pick up shear that is shed from the beam action
mechanism.
Over the second interval in tests CB2, RB1, and RB2, the
shear carried by beam action reaches a fairly stable value
until reduced by the onset of yielding. This value ranged
from a low of 80 kN in RB1 to a high of 95 kN in CB2. Test
CB1(b) did not display the same level of stability in its beam
action shear component although this may have been a result
of its loading history. The maximum value of beam action
shear measured for CB1(b) was 78 kN, which is comparable
to the other three specimens. It is noted that the higher val-
ues of beam action shear were not associated with uncoated
reinforcement, as was expected.
At failure over the second interval of specimens CB1(b),
CB2, RB1, and RB2, beam action accounted for between
one and two thirds of the total shear. Nevertheless, the total
shear/compression failure shears are relatively consistent,
ranging from a low of 123 kN to a high of 130 kN. These
1999 NRC Canada
O lonisakin and Alexander 815
0
25
50
75
100
125
0 2 4 6 8 10 12 14
Deflection (mm)
V
e
r
t
i
c
a
l
s
h
e
a
r
(
k
N
)
(b)
0
25
50
75
100
125
0 2 4 6 8 10 12 14
Deflection (mm)
V
e
r
t
i
c
a
l
s
h
e
a
r
(
k
N
)
Beam action
Arching action
Total shear
(a)
Fig. 7. Components of shear for CB1(a): (a) first interval; and
(b) second interval.
0
25
50
75
100
125
150
0 2 4 6 8
Deflection (mm)
V
e
r
t
i
c
a
l
s
h
e
a
r
(
k
N
)
Beam action
Arching action
Total shear
Fig. 8. Components of shear for CB1(b).
0
25
50
75
100
125
0 2 4 6 8 10
Deflection (mm)
V
e
r
t
i
c
a
l
s
h
e
a
r
(
k
N
)
Beam action
Arching action
Total shear
(a)
0
25
50
75
100
125
0 2 4 6 8 10
Deflection (mm)
V
e
r
t
i
c
a
l
s
h
e
a
r
(
k
N
)
(b)
Fig. 9. Components of shear for RB1: (a) first interval; and
(b) second interval.
failure shears compare reasonably with an unfactored resis-
tance of 109 kN as given by the simplified method for shear
design in Standard A23.3 (CSA 1994).
It is apparent from Figs. 8 to 11 that at shear/compression
failure, the beam action mechanism is shedding load to the
arching action mechanism. In all cases except that of the
second interval of specimen CB2, the loss of beam action
shear is a consequence of yielding of the reinforcement. In
the case of the second interval of CB2, the drop in beam ac-
tion shear at failure is attributed to a loss of bond.
Discussion
While far from conclusive, these tests raise questions
about the nature of shear failure. There is the question of
what shear force V is appropriate to use for the computation
of the nominal shear stress V bd / in a slender beam. The
concept of a limiting nominal shear stress is more consistent
with the beam action mechanism than with the arching ac-
tion mechanism. Given that both arching action shear and
beam action shear contribute to the total shear at or near the
ultimate load, it is conceptually incorrect to assign all of the
load to beam action. This means that even though the simpli-
fied method for shear design given in Standard A23.3-94
(CSA 1994) reasonably predicts the failure strength of the
beams, its underlying mechanics are not consistent with the
behaviour of the beams. For the tests reported here, beam
action accounted for only about one to two thirds of the total
failure shear. Alternatively, if one considers the maximum
values of beam action shear, then the appropriate unfactored
resisting shear stress would be about 78% of the critical
shear stress given in the standard.
It is worth noting that for all tests ending in a shear/com-
pression failure, shear was being shed from the beam action
mechanism to the arching action mechanism. In most cases,
this internal redistribution of shear was the result of yielding
of the reinforcement, which reduced the difference in bar
force and hence the force gradient between successive sec-
tions of the beam. However, the failures themselves may be
the result of some critical stress state that is associated with
the shifting of load from the beam action mechanism to the
arching action mechanism.
Failure associated with shifting shear from a beam action
mechanism to an arching action mechanism has been ob-
served before. Kani (1964) noted a significant increase in the
shear capacity of deep beams with reduced bond of rein-
forcement provided that the reinforcement was anchored at
its ends. Kani reduced both the bond strength and stiffness
by casting soft concrete locally around the reinforcement.
In one specimen, the reinforcement was debonded com-
1999 NRC Canada
816 C an.J.C iv.Eng.Vol.26,1999
0
25
50
75
100
125
150
0 2 4 6 8
Deflection (mm)
V
e
r
t
i
c
a
l
s
h
e
a
r
(
k
N
)
(b)
0
25
50
75
100
125
150
0 2 4 6 8
Deflection (mm)
V
e
r
t
i
c
a
l
s
h
e
a
r
(
k
N
)
Beam action
Arching action
Total shear
(a)
Fig. 10. Components of shear for CB2: (a) first interval; and
(b) second interval.
0
25
50
75
100
125
150
0 2 4 6 8
Deflection (mm)
V
e
r
t
i
c
a
l
s
h
e
a
r
(
k
N
)
Beam action
Arching action
Total shear
(a)
0
25
50
75
100
125
150
0 2 4 6 8
Deflection (mm)
V
e
r
t
i
c
a
l
s
h
e
a
r
(
k
N
)
(b)
Fig. 11. Components of shear for RB2: (a) first interval; and
(b) second interval.
1999 NRC Canada
O lonisakin and Alexander 817
pletely. Beams with compromised or no bond could not de-
velop beam action shear to any significant degree and were
obliged to carry virtually all shear by arching action. For
these beams there was no shift in the shear transfer mecha-
nism since all the shear was carried by arching action from
the outset. The beams with competent bond failed in shear at
lower loads than their counterparts with compromised bond.
The authors description of the specimens as having failed
in shear/compression may be questioned. Consider specimen
RB1. From its appearance in Fig. 4b one might conclude
that it failed in shear. However, the loaddeflection curve
shown in Fig. 3 clearly shows that RB1 failed after the for-
mation of a folding mechanism. Strain gauge readings indi-
cate full yielding at section 1 and partial yielding at
section 2 just prior to rupture. Figure 9 shows that yielding
of the reinforcement was forcing a change in the mechanics
of shear transfer within the beam. If the changing shear
transfer mechanism created a critical stress state in the con-
crete, one might argue that RB1 failed in shear. However,
because the change in internal shear mechanics was initiated
by the spread of yielding associated with the development of
a plastic hinge, one might argue that RB1 failed in flexure.
The authors hold that shear failure is more appropriate.
Yielding of the reinforcement and the formation of a plastic
hinge created a critical state for the gradient of the bending
moment and not its magnitude.
Conclusions
Five tests were performed on four wide beams. Two of the
beams were reinforced with epoxy coated bars whereas the
remaining two had uncoated reinforcement. Strain measure-
ments of the reinforcement were used to estimate the aver-
age force gradients in the reinforcing bars and to divide the
total shear into beam and arching action components.
1. Beam action shear did not account for all or even a con-
sistent fraction of the total shear at failure. Although the
unfactored shear resistance given by the simplified
method for shear design accurately predicts the shear
failure loads for the beams, it is not consistent with their
behaviour. As it is based on an assumed critical shear
stress, the simplified method is consistent with beam ac-
tion and not arching action.
2. Based on these results and those of Kani (1964), it is
concluded that shear/compression failure may be associ-
ated with a shift in the internal mechanics of shear
transfer from beam action to arching action.
3. Yielding of reinforcement within the shear span forces a
shift from beam action to arching action. As a conse-
quence, the formation of a plastic hinge may precipitate
a shear failure.
4. Epoxy coating of flexural reinforcement did not affect
the magnitude of measured force gradients on the rein-
forcement.
References
Alexander, S.D.B., Lu, X., and Simmonds, S.H. 1995. Mechanism
of shear transfer in a column-slab connection. Proceedings of
the Second Engineering Materials Symposium, Annual Confer-
ence of the Canadian Society for Civil Engineering, June 13,
Ottawa, Ont., Vol. 2, pp. 207216.
Canadian Standards Association (CSA). 1994. Design of concrete
structures. Standard A23.3-94, Canadian Standards Association,
Rexdale, Ont.
Kani, G.N.J. 1964. The riddle of shear failure and its solution. ACI
Journal, 61(4): 441467.
MacGregor, J.G. 1997. Reinforced concrete: mechanics and
design 3rd ed. Prentice-Hall, Inc. N.J.
List of symbols
a: shear span of beam
b: width of beam
d: effective depth of beam
f
y
: yield strength of steel reinforcement
j: effective lever arm factor
M: bending moment
P: applied central vertical load
T: tensile bar force
V: vertical shear force
V
a
: arching action component of vertical shear
V
b
: beam action component of vertical shear