ACI STRUCTURAL JOURNAL

TECHNICAL PAPER

Title no. 107-S41

Shear-Transfer Strength of Reinforced Concrete

by Khaldoun N. Rahal
A recently developed model for the calculation of shear strength in
reinforced concrete membrane elements subjected to in-plane
stresses and in beams subjected to shear and torsion is applied to
the shear-transfer problem. The modeling is different from the
commonly used shear-friction concept, and relates the strength at
a shear interface to the state of stress in a membrane element
along this interface. The shear strength is hence related not only to
the concrete strength and clamping steel, but also to the steel
parallel to the shear-transfer plane. The calculations of the simple
model are compared to the experimental results from 114 normalweight pushoff specimens and 15 composite beams available in the
literature and are found to be in very good agreement. The model
is also used to derive the empirically based coefficients of existing
methods that relate the shear-transfer strength to the square root of
the clamping stress.
Keywords: composite beams; reinforced concrete; shear-friction; shear
transfer; strength.

INTRODUCTION
Numerous design cases require the calculation of the
amount of reinforcement necessary to resist shear transfer
across an interface between two concrete members that can
slip relative to each other. The interface can be susceptible to
a potential crack or can be cracked due to previous conditions
such as external tension and shrinkage, and can be a cold
joint. The interface between a precast girder and a cast-inplace deck slab, and the bearing zones in precast girders,
corbels, and horizontal construction joints in walls, are
examples of shear-transfer cases. Refer to Fig. 1(a).
The design for shear transfer has been largely based on
empirical and semi-empirical methods that were developed
using the experimental results from pushoff specimens and
composite beam specimens. Figure 1(b) shows a typical
pushoff specimen similar to that used in the early tests by
Hofbeck et al.1 The applied compressive forces create
shearing stresses (v) along the critical plane, which could be
either precracked or uncracked. The shearing stresses at
ultimate conditions are typically assumed to be constant
along the interface plane, and an average shearing strength
along the plane is calculated. These shearing stresses act in
combination with compressive stresses (Fig. 1(b)).
There has been a considerable amount of experimental
tests on pushoff specimens,1-8 which led to the development
of numerous models.9-17 The well-known shear-friction
model of the ACI Code17 is based on the assumption that a
crack exists along the shear plane before the load is applied.
The failure occurs by sliding along the shear plane and the
opening of the crack around the aggregates. The clamping
steel is stressed to its yield strength and a friction force
proportional to the clamping yield force is activated. The
ACI nominal shear strength is given by
vACI = y fy y
ACI Structural Journal/July-August 2010

(1)

Fig. 1(a) Examples of shear transfer in reinforced concrete

structures; and (b) typical pushoff specimen and state of
stress along shear transfer plane.
but not greater than (0.2fc ) or 5.5 MPa (800 psi), where y is
the ratio of clamping reinforcement (in the y-direction), fy y
is the yield strength of clamping reinforcement (in the
ACI Structural Journal, V. 107, No. 4, July-August 2010.
MS No. S-2009-105.R3 received July 24, 2009, and reviewed under Institute
publication policies. Copyright 2010, American Concrete Institute. All rights reserved,
including the making of copies unless permission is obtained from the copyright proprietors.
Pertinent discussion including authors closure, if any, will be published in the May-June
2011 ACI Structural Journal if the discussion is received by January 1, 2011.

419

ACI member Khaldoun N. Rahal is a Professor of civil engineering at Kuwait

University, Kuwait City, Kuwait. He is Director and Past President of the ACI-Kuwait
Chapter and is a member of Joint ACI-ASCE Committee 445, Shear and Torsion.

Fig. 2Membrane element subjected to in-plane stresses

and summary of SMCS equations.
y-direction), fc is the compressive strength of the concrete,
and is the coefficient to account for friction.
Walraven et al.6 developed a model based on 88 test specimens
with concrete strength ranging from 21 to 68 MPa (3000 to
9900 psi). The model relates the shear strength to the
clamping reinforcement as well as fc , but does not place an
upper limit on the shear strength
v WFP = C 1 ( C 3 y f y y)
0.88(fc )0.406, C2 =
16.76(fc )0.406, C2

C2

(2)

0.167(fc )0.303, and C3 = 1 in

= 0.0371(fc )0.303, and C3 =

where C1 =
MPa (C1 =
0.007 in psi).
Hsu et al.13 adopted a more rational approach by considering
the concrete along the shear-transfer plane to be a membrane
element subjected to combined shearing and normal stresses.
They used the equations of the softened truss model to
calculate the shear strength and overall behavior. This
analysis differs in concept from the more commonly used
shear-friction models; however, the solution procedure is
computationally demanding and requires the use of a
computer. A simple semi-empirical equation was subsequently
proposed by Mau and Hsu.14
v MH = ( 0.66 f c ) y f y y 0.3 f c

(3)

Loov and Patnaik15 developed a similar equation for the

nominal shear strength
v LP = ( 0.6 f c ) 0.1 + y f y y 0.25 f c ( in MPa ) (4)
where is the factor to account for lightweight aggregates.
The factor 0.1 is replaced by 15 in psi, and has a negligible
effect at relatively large clamping steel. It was included to
avoid the discontinuities in the present codes at low
clamping stresses.15
Mattock16 proposed a trilinear model calibrated using the
results from 189 normalweight and lightweight test specimens
with fc ranging from 16 to 99 MPa (2300 to 14,350 psi)
v Mat = 2.25y f y y when y f y y K 1 1.45
420

(5a)

v Mat = K 1 + 0.8y f y y when y f y y > K 1 1.45

(5b)

but not greater than 0.3fc or 16.6 MPa (2400 psi) for normalweight concrete and 0.2fc or 8.3 MPa (1200 psi) for sandlightweight concrete and all lightweight concrete. The factor
K1 is taken as 0.1fc but not greater than 5.5 MPa (800 psi).
To account for a normal stress y acting perpendicular to
the shear plane, the superposition of steel can be applied
and, consequently, the term y fy y is replaced by y fy y y,
where y is positive if tensile.
Equations (1) to (5) show that existing models are simple,
but are empirical or semi-empirical. More rational models,
such as those by Hsu et al.,13 have the advantage of being
applicable to other shear cases, but are iterative and, hence,
are not readily suitable for use in a design office. The challenge is
to develop a more rational model that shares the simplicity
and accuracy of empirical methods.
A recently developed model called the simplified model
for combined stress resultants (SMCS) is a simple, noniterative
model for the calculation of the shear strength and the mode
of failure of membrane elements subjected to in-plane
shearing and normal stresses.18 The model was generalized
to apply to reinforced and prestressed concrete beams
subjected to shear combined with flexure and axial forces,19
to pure torsion,20 and to torsion combined to flexure.21 This
paper extends the applicability of the model to solve the
shear-transfer problem.
RESEARCH SIGNIFICANCE
Most simple methods available to solve the shear-transfer
problem in shear-friction specimens and across cold joints in
composite beams are semi-empirical, and their application is
limited as they cannot be applied to other types of shear
problems such as shear and torsion in beams and shear in
membrane elements. This paper presents a simple, noniterative
model that is developed based on a rational theory and is
applicable to other types of shear problems. The proposed
model has a favorable combination of simplicity, generality,
and accuracy in comparison with existing models.
SMCS FOR PURE SHEAR
The SMCS model developed for pure shear was applied
without modification to the case of shear friction. A brief
background of the development of the SMCS is presented.
Full details of the model can be obtained from Rahal.18
Figure 2 shows a membrane element reinforced with
orthogonal steel subjected to in-plane shearing stresses and a
summary of the SMCS equations. The equations are developed
for the case of pure shear, and the effects of the normal
stresses are accounted for using the concept of superposition.
The model assumes that the main factors that affect the
pure shear strength of membrane elements are the amounts
and the strength of the orthogonal steel and the concrete
compressive strength. Other factors, such the maximum size
of the coarse aggregate and the spacing and diameter of the
reinforcement, have limited effects and are neglected in the
simplified model. The three main factors are efficiently
combined in the following reinforcement indexes
x fy x
y fy y
- ; y = ------------- x = -------------f c
f c

(6)

ACI Structural Journal/July-August 2010

where x,y are the reinforcement indexes in x- and y-directions,

respectively; x , y is the reinforcement ratios in x- and ydirections, respectively; and fy x , fy y is the yield strength
of reinforcement in x- and y-directions, respectively.
Several advanced models can be used to calculate the
strength and mode of failure of the membrane element
shown in Fig. 2. One such model is the modified compression
field theory (MCFT), which is capable of calculating the full
response of membrane elements subjected to in-plane
stresses.22 The equations of the MCFT were used to calculate
the relationship between the reinforcement indexes on one
side and the shear strength and mode of failure on the other
side. The spacing of the diagonal cracks was assumed to be
300 mm (12 in.) and the maximum coarse aggregate size was
taken as 19 mm (0.75 in.), and the analysis was run on 40 MPa
(5800 psi) concrete elements. Figure 3(a) shows the relationship
between the normalized strength vn/fc and the index y,
when x remained constant (= 0.2). For y values below that
corresponding to the point marked A, the y-reinforcement
yielded before concrete crushing at ultimate conditions. For
larger values, the strains remained below yield levels at ultimate.
The index corresponding to Point A is referred to as the
balanced reinforcement index in the y-direction. For the
results shown in Fig. 3(a), the amount of x-reinforcement (x =
0.2) was smaller than the balanced ratio in the x-direction,
and yielded before concrete crushing.
The analysis was repeated for different levels of x and the
curves shown in Fig. 3(b) were generated. Two balanced
reinforcement curves were obtained. For x smaller than the
balanced x-indexes, the x-reinforcement yields when ultimate
conditions are reached. Similarly, for y smaller than the balanced
y-indexes, the y-reinforcement yields when ultimate conditions are
reached. This gives rise to four classifications of reinforcement
corresponding to four different modes of failures: 1) fully
under-reinforced (UR), where both x- and y-reinforcement
yield before ultimate conditions are reached; 2) partially
under-reinforced (PUR), where only the x-reinforcement
yields at ultimate conditions; 3) partially under-reinforced
(PUR), where only the y-reinforcement yields at ultimate
conditions; and 4) fully over-reinforced (OR), where
concrete crushes before yielding in any steel.
For fully under-reinforced membranes, yielding in both the
x- and y-reinforcement prevents any restraint on the opening
in the diagonal crack and, hence, reduces the concrete
contribution provided mainly by aggregate interlock to nearly
zero. For this reason, the curves in Zone 1 (UR) can be calculated
using the plasticity theory,23 as follows.
v
----n- =
f c

x y

(7)

In the remaining three zones of Fig. 3(b), the concrete

contribution is considerable, and is built-in in the calculated
ultimate shear strength vn.
Figure 3(a) shows that Eq. (7) adequately calculates the
strength of fully under-reinforced elements, but overestimates the
strength of partially under-reinforced and fully over-reinforced
elements. It also shows that the additional strength obtained
by increasing the amount of steel beyond the balanced
level is limited. To maintain the simplicity of the model, the
strength provided by reinforcement in excess of the balanced
values is neglected and, hence, the amount of usable reinforcement is limited to this value. In Fig. 3(b), the upper limit on
ACI Structural Journal/July-August 2010

Fig. 3(a) Relationship between normalized shear capacity

and reinforcement index; and (b) shear strength and mode
of failure curves for membrane elements.
the reinforcement in one direction is shown to depend on
the amount of reinforcement in the other direction. This
is neglected in the simplified version of the model and a
single value is adopted for a given fc.
Equation (7) implies that the upper limit on the indexes
leads to an upper limit on the shear strength. Based on
experimental results from normal- and high-strength
concrete, the upper limit on the normalized shear strength vn/fc
was found to depend on fc and can be calculated18 using
1 f c
- ( in MPa )
= --- -------3 900

(8)

where is the upper limit on shear strength (and amount of

usable reinforcement index). The coefficient 900 is in MPa,
and is replaced by (130,500) if fc is in psi.
The effects of in-plane normal stresses were accurately
accounted for in SMCS 18,19,21 using the concept of
superposition of reinforcement. Using the concept of
superposition and introducing the upper limit on the
amount of usable reinforcement gives
x f y x x
y f y y y
- ; y = --------------------------
x = -------------------------f c
f c

(9)

421

where x , y are the normal stresses in the x- and y-directions

(taken as positive if tensile). The upper limit on x and y
leads to an automatic limit on the normalized shear stress.
v SMCS f c =

x y

v n = ( 0.55 f c ) y f y y 0.3f c

(11)

(10)

Consequently, the equations of SMCS are reduced to Eq. (8)

to (10). Refer to Fig. 2. Comparison between the calculations
of the SMCS model and the results from 84 membrane
elements subjected to in-plane shearing and normal stresses
showed that its accuracy was very similar to those of the
MCFT22 and the plasticity theory,23,24 and provided more
favorable results than the plasticity theory in the identification
of the mode of failure. The model was also generalized to
apply to reinforced and prestressed concrete beams
subjected to shear combined with flexure and axial forces,19
to pure torsion,20 and to torsion combined to flexure.21
SMCS FOR SHEAR-TRANSFER SPECIMENS
Figure 1(b) shows an element along the shear transfer
plane of a pushoff specimen. This element is subjected to a
combination of a shearing stress along with a normal
compressive stress in the x-direction, x. The shear strength
of such elements can be calculated using the SMCS model if
x is available or assumed. The compressive stress is proportional
to the shearing stress and provides an increase over the pure
shear strength of the element. An accurate calculation of x
is not simple, especially because it varies considerably along
the shear plane.
To maintain the simplicity of the model while providing
conservative results, the stresses (x) are neglected. Neglecting
x does not have a significant effect on the results of the bulk
of the shear-friction specimens because the x-reinforcement
provided in most of these specimens was relatively large, and
the elements were over-reinforced in the x-direction.
Figure 3(b) shows that in membranes over-reinforced in the
x-direction, an increase in x (that can be caused by x) does
not have a significant influence on the shearing strength. In
the SMCS equations, this increase is neglected by placing
the limit . By neglecting the compressive stresses, the
strength of the shear-transfer specimens is reduced to the
calculation of the strength of the membranes subjected to
pure shear or shear combined with y.
It is to be noted that similar to theoretical models, many
experimental studies did not consider the importance of the
x-steel and, hence, did not report its full details. For
example, Mattock and Hawkins2 and Walraven and
Stroband7 did not report the bar size or fy x , and Mattock
et al.3 did not report fy x.
DERIVATION OF COEFFICIENTS
IN EMPIRICAL METHODS
Equation (10) is a general equation that is applicable to
various values of x and fc . For normal-strength members
over-reinforced in the x-direction, Eq. (1) can be reduced to
a format similar to those of existing methods that were
developed based on results from such members. For
example, all 33 specimens tested by Hofbeck et al.1 were
reinforced with x = 5.7% and were normal-strength
concrete (average fc = 25.9 MPa [3750 psi]). This large
amount of steel caused the specimens to be over-reinforced in
the x-direction and, consequently, the usable x is
limited to (equal to approximately 0.30 for such
422

concrete). Substituting this value of in Eq. (10) and

rounding the numbers gives

Equation (11) is very similar in format to Eq. (4) by Loov and

Patnaik15 and Eq. (3) by Mau and Hsu,14 but gives 8% and
16% smaller strength, respectively. The upper limits are also
within the same range. For normal-strength concrete where
fc is known, Eq. (11) can be further simplified by substituting
the value of the fc (the average of the tests) to give
vn = 2.80 y f y y MPa

(33.8 y f y y psi)

(12)

with an upper limit of 7.9 MPa (1140 psi). Equation (12) is

very similar in format to Eq. (13) proposed by Birkeland (as
reported in Reference 12) and Eq. (14) proposed by Raths.11
vn = 2.78 y f y y MPa

(33.5 y f y y psi)

(13)

vn = 3.11 y f y y MPa

(37.42 y f y y psi)

(14)

Equations (13) and (14) differ from Eq. (12) by only 1% and
10%, respectively. Consequently, the strength equations
proposed by Loov and Patnaik,15 Mau and Hsu,14 Birkeland
(as reported in Reference 12), and Raths11 can be seen as
special cases of the more general strength equation of the
SMCS method (Eq. (10)), and are most suitable for specimens
over-reinforced in the x-direction.
EXPERIMENTAL EVALUATION
OF PROPOSED MODEL
The calculations of the SMCS method and other methods
are compared with the experimental results from 114
normalweight precracked and uncracked concrete pushoff
specimens and 15 composite beam specimens. Table 1
summarizes the results of the comparison between the
observed strength and the calculations of six models. A more
detailed comparison is shown in Table 1A (Appendix A*) for
SMCS, the ACI code17 equations, and for the model
proposed by Mattock.16 The Mattock model is selected for
the detailed comparison because it was developed based on
a larger database, and because this database includes more
high-strength concrete specimens.
Hofbeck et al.1 pushoff specimens
Hofbeck et al.1 tested five series of normal-strength concrete
pushoff specimens to study the effects of precracking; concrete
strength; and the size, arrangement, and yield strength of
the clamping y-reinforcement. The 12 uncracked specimens of
Series 1 had an average fc of 28.4 MPa (4100 psi) and a
variable amount of y-steel achieved using 9.5 mm (No. 3)
bars at a variable spacing. The average provided x was
0.62, which was larger than the upper limit = 0.302. Hence,
the usable x was taken as 0.302. The comparison between
the observed and calculated nominal shearing strength is
shown in Fig. 4(a).
*The Appendix is available at www.concrete.org in PDF format as an addendum to
the published paper. It is also available in hard copy from ACI headquarters for a fee
equal to the cost of reproduction plus handling at the time of the request.

ACI Structural Journal/July-August 2010

Table 1Summary of correlation of calculations of six methods with experimental shear capacities
Average of observed-to-calculated shear strength (coefficient of variation, %)
Method
No.
Source of tests of tests fc , MPa

y fy y ,
MPa

x fy x ,
MPa

y , MPa

SMCS

Mattock16 Mau and Hsu14

Walraven
et al.6

Loov and
Patnaik15 ACI 318-0517

Hofbeck et al.1

33

16 to 31

0.34 to 10

17.7

1.17 (12.6) 1.18 (20.5)

1.04 (13.4)

1.05 (11.8)

1.18 (15.9)

1.69 (28.0)

Mattock and
Hawkins2

28 to 40

2.15 to 6.8

0 to 5.61

1.08 (8.86) 1.03 (14.2)

0.99 (19.1)

0.96 (10.1)

1.42 (21.1)

1.66 (13.7)

Mattock et al.3

26 to 29

3.65 to 5.6

16.6

2.76 to 0

1.05 (9.24) 1.05 (9.98)

0.88 (9.18)

0.94 (10.1)

0.95 (9.59)

1.40 (18.8)

Mattock4

41

1.6 to 13.3

1.09 (8.40) 1.06 (18.0)

0.93 (12.4)

0.91 (8.57)

1.06 (16.5)

1.97 (22.8)

Mattock et al.5

28

1.5 to 7.7

24.3

1.00 (8.04) 0.99 (4.28)

0.89 (10.8)

0.91 (9.77)

1.02 (15.6)

1.38 (10.2)

Walraven
et al.6

31

17 to 48

1.1 to 15.2

9.45

1.17 (15.3) 1.12 (11.5)

1.03 (11.5)

0.99 (11.8)

1.18 (13.4)

1.70 (18.5)

Walraven and
Stroband7

99

3.3 to 14.9

0.92 (12.6) 1.02 (10.5)

0.66 (12.6)

0.53 (6.13)

0.72 (12.9)

2.43 (31.4)

Nagle and
Kuchma8

18

93 to 121 0.73 to 6.3

5.84

1.18 (19.8) 0.91 (21.4)

0.42 (19.9)

0.43 (15.5)

0.45 (21.0)

1.48 (21.3)

Loov and
Patnaik15

15

19 to 48

||

1.19 (9.76) 1.56 (35.3)

0.98 (9.85)

1.03 (9.38)

1.04 (8.96)

1.64 (14.1)

All tests

129

16 to 121 0.34 to 15.2

2.76 to 5.61 1.14 (14.8) 1.14 (26.7)

0.90 (26.5)

0.90 (26.8)

1.02 (29.2)

1.67 (25.7)

0.4 to 7.7

Longitudinal reinforcement relatively large, upper limit (=) is assumed.

Based on assumed value of f
y x = 460 MPa (69 ksi) (results do not change for any fy x > 240 MPa [35 ksi]).

Calculated based on assumed value of fy x = 460 MPa [69 ksi].

Composite beam specimens.
||Near flexural compression zone, upper limit (=) is assumed.
Note: 1 MPa = 145 psi.

The 16 specimens of Series 2, 3, and 4 were cracked, and

their average fc was 26.6 MPa (3850 psi). The variable
amount of y-reinforcement was achieved by changing the
number of the 9.5 mm (No. 3) bars in Series 2 and by
changing the size of the bars while maintaining their number
constant in Series 3. Series 4 was similar to Series 2 but the
clamping reinforcement had a larger yield strength. The
average provided x was 0.66 and, hence, the usable value
was limited to the upper limit = 0.302. The results of the
comparison are shown in Fig. 4(b). The specimens of Series
5 had a relatively lower concrete strength, with an average fc
of 17.5 MPa (2540 psi), and the amount of y-reinforcement
was varied. The usable x was limited to = 0.314. The
results of the comparison are shown in Fig. 4(c).
The comparisons in Fig. 4 and Table 1 show that the SMCS
captures the trends observed in the tests and that the results were,
in general, similar to those of the other methods. On the other
hand, the ACI calculations were more conservative but less
accurate than the other methods.
Importance of reinforcement parallel
to transfer plane
Nagle and Kuchma8 tested 18 large-scale high-strength
concrete specimens with the shear transfer steel inclined at
angles of 25 and 35 degrees to the shear transfer plane. The
precracked specimens modeled the shear transfer along diagonal
cracks in large-scale bridge girders. The concrete compressive
strength ranged from 93 to 121 MPa (13,500 to 17,550 psi)
and the average was 104 MPa (15,080 psi).
In all of these tests, the ratio of x-reinforcement x was
approximately 0.0123, which was only 22% of the x =
0.057 used in the Hofbeck et al.1 tests. The x-direction
reinforcement index x was, on average, 0.0567 and, hence,
was not limited to the upper limit of = 0.218 calculated
using an average fc of 104 MPa (15,080 psi). These experimental
results offer a test of the importance of considering the steel
ACI Structural Journal/July-August 2010

parallel to the shear transfer interface, and a test of the

adequacy of the SMCS and the other methods.
Substituting the average x and in Eq. (10) reduces it to
the following
v n = ( 0.24 f c ) y f y y 0.22f c

(15a)

Substituting the value of the average fc reduces it further to

vn = 2.43 y f y y MPa

(29.3 y f y y psi)

(15b)

with an upper limit of 22.6 MPa (3280 psi). Equation (15)

gives only 36% of the strength calculated using Mau and
Hsu's equation (Eq. (3)), approximately 40% of that by the
Loov and Patnaik equation (Eq. (4)), approximately 87% of
that by Birkelands equation (Eq. (13)), and 78% of that by
Raths (Eq. (14)).
Figure 5 compares the experimental with the calculated
strength. All methods described in the Introduction section
are included in the comparison to test their adequacy to this
case of low x index. The reported values of the clamping
stress and the experimental shear strength were adjusted by
Nagle and Kuchma8 for the inclination of the clamping steel.
The results show an inadequacy of the methods proposed by
Mau and Hsu,14 Walraven et al.,6 and Loov and Patnaik.15
The calculations of all 18 specimens were severely
unconservative, with calculated strength in some cases
exceeding three times the observed strength. The Walraven
et al.6 method was developed based on specimens with a
concrete compressive strength below 70 MPa (10,150 psi).
Even if such a conservative value is taken as a limit for the
applicability of the method, the results remained severely
unconservative, as shown in Fig. 5. The same is true for the
Mau and Hsu14 and for the Loov and Patnaik15 methods. The
423

Fig. 5Shear transfer in large-scale high-strength concrete

pushoff specimens with relatively low level of x-reinforcement.8

Fig. 6Effect of stress in direction perpendicular to transfer

plane2,3 on shear strength.
the steel parallel to the shear plane to avoid unconservative
designs in empirically based methods.
Fig. 4Shear strength of normal-strength reinforced concrete
pushoff specimens.1
results of the SMCS method were generally adequate,
whereas those by the ACI Code were more conservative.
Mattocks method does not take into consideration the
x-reinforcement, and overestimated the strength of a
specimen by 85%. However, it yielded adequate overall
results for this case because Eq. (5a) underestimates the
strength of specimens with relatively small clamping
stresses. This is shown in Specimen 3 (Hofbeck et al.1; refer
to Table A1) and is also shown in a following section using
the Loov and Patnaik tests.15 In conclusion, the Nagle and
Kuchma8 tests clearly show the importance of accounting for
424

External clamping stresses

The effects of an external compressive or tensile stress
perpendicular to the transfer plane can be accounted for by
the superposition of reinforcement. Mattock et al.3 tested two
series of specimens subjected to tensile stresses. Series E and F
had an average y f y y of 3.73 and 5.52 MPa (540 and
800 psi), respectively. The yield strength of the x-reinforcement
was not reported, but the results using the SMCS method
would not change for any fy x larger than 240 MPa (35 ksi)
for this series of specimens. A value of 460 MPa (69 ksi) was
used to report the x fy x shown in Tables 1 and A1.
The effects of a compressive clamping stress were investigated
by Mattock and Hawkins2 in their series of 10 specimens tested
using a modified pushoff test setup. The x-reinforcement
was not reported, and a value of x equal to is assumed.
ACI Structural Journal/July-August 2010

Figure 6 shows the experimental results from the three

series of specimens (E, F, and 10) subjected to shear and a
clamping stress, and compares them with the strength calculated
using SMCS. A very good agreement is obtained. The
average and coefficients of variation values reported in
Table 1 show that good accuracy was obtained by using
Mattocks method.16 The tables also show that the ACI
equations were conservative, especially in combined shear
and compression. The upper limit of 5.5 MPa (800 psi) on
the strength was critical and did not allow benefit from the
compressive clamping stresses.
Tables 1 and A1 include comparisons with experimental
results from other pushoff specimens.4-7
Composite beam specimens
Loov and Patnaik15 tested the shear transfer between the
webs and the flanges of 16 composite beams. The webs were
cast first and the top surfaces were left as-cast with the coarse
aggregates left protruding without efforts to produce rough
surfaces. Stirrups extended through the cold joints and
provided the clamping forces after the flanges were cast. The
joint in one beam was smooth and was not included in the
evaluation. The beams were tested in a three-point loading
set up and, hence, the web-flange interface was located near
the flexural compression zone. This proximity provided a
state of considerable level of compression in the x-direction
(parallel to the shear interface), and, hence, for the SMCS
calculations, x is considered to be large enough to be
limited by the upper value .
The ACI Code17 provides special design provisions for the
horizontal interface between precast girders and cast-inplace slabs when the shearing stress is smaller than 3.5 MPa
(500 psi).
v ACI = ( 1.8 + 0.6y f y y ) 3.5 MPa

(16a)

v ACI = ( 260 + 0.6y f y y ) 500 psi

(16b)

For clean and intentionally roughened joints, and for clean

but not intentionally roughened joints with minimum transverse
steel as per ACI 318-05, Section 11.5.6, a 0.55 MPa (80 psi)
stress is allowed.
In 12 of the beams, fc was approximately 35 MPa (5075 psi)
and the clamping force was variable. Figure 7 shows the
experimental shear strengths and compares them with the
calculations of the three methods, in addition to the results of
Loov and Patnaiks15 own method, which was calibrated
using these test results. The ACI equations provided very
conservative calculations, whereas Mattocks16 equations
gave very conservative results at low levels of clamping
forces, showing the disadvantages of a linear relationship
between the strength and the clamping stress. The SMCS
provided an accurate lower-bound of the results and was
more suitable than the empirically developed methods.
Overall correlation
Table 1 shows the average and the coefficient of variation
for the ratios of the experimental to the calculated shear
strength for the 129 test results included in this study. A
comparison between the results of the six methods shows that
the proposed SMCS model provided the most accurate results.
ACI Structural Journal/July-August 2010

Fig. 7Shear transfer across cold joint between webs and

flanges of composite beams.15
CONCLUSIONS
This paper showed that the SMCS method can be applied
to the shear-transfer problem. The calculations of this
method and five existing semi-empirical methods were
compared to the experimental results from 15 composite
beams and 114 normalweight concrete pushoff specimens,
and the proposed method was shown to provide the most
accurate results.
The better accuracy of the proposed method was mainly
due to its ability to account for the effects of the reinforcement
in the direction parallel to the shear-transfer plane. The five
other models that were discussed in this paper do not account
for this factor, and were calibrated using test specimens
where this reinforcement was relatively very large. Four of
the methods were severely unconservative when used to
calculate the shear strength of recently tested pushoff specimens
with moderate (and practical) levels of such reinforcement.
Meanwhile, the ACI code results were adequate for these tests
because of the markedly large margin of conservatism built-in
in the ACI shear-friction model.
It was also shown that the semi-empirical equations that
relate the shear strength to the square root of the clamping
stress can be seen as special cases of the more general
strength equation of the proposed SMCS method. The semiempirical equations can be derived from the SMCS method
for the case where the specimens are over-reinforced in the
direction parallel to the shear-transfer plane.
The noniterative SMCS model is a generalized model
applicable also to membrane elements subjected to in-plane
stresses and to beams subjected to combined shear and
flexure and to combined torsion and flexure. Hence, it has a
favorable combination of generality, accuracy, and simplicity.
NOTATION
vACI
vexp
vLP
vMat
vMH
vn
vSMCS
vWFP

=
=
=
=
=
=
=
=
=

ultimate shear strength calculated using ACI equations

experimentally observed ultimate shear strength
ultimate shear strength calculated using model by Loov and Patnaik15
ultimate shear strength calculated using model by Mattock16
ultimate shear strength calculated using model by Mau and Hsu14
nominal shear strength
ultimate shear strength calculated using proposed SMCS model
ultimate shear strength calculated using model by Walraven et al.6
angle of inclination of clamping reinforcement with respect to
shear transfer plane

425

REFERENCES
1. Hofbeck, J. A.; Ibrahim, I. O.; and Mattock, A. H., Shear Transfer in
Reinforced Concrete, ACI JOURNAL, Proceedings V. 66, No. 2, Feb. 1969,
pp. 119-128.
2. Mattock, A. H., and Hawkins, N. M., Shear Transfer in Reinforced
ConcreteRecent Research, Journal of the Prestressed Concrete
Institute, V. 17, No. 2, 1972, pp. 55-75.
3. Mattock, A. H.; Johal, L.; and Chow, H. C., Shear Transfer in Reinforced
Concrete with Moment or Tension Acting across the Shear Plane, Journal
of the Prestressed Concrete Institute, V. 20, No. 4, 1975, pp. 76-93.
4. Mattock, A. H., Shear Transfer under Monotonic Loading across an
Interface between Concretes Cast at Different Times, University of Washington
Report SM 76-3, Sept. 1976, 66 pp.
5. Mattock, A. H.; Li, W. K.; and Wang, T. C., Shear Transfer in Light-Weight
Reinforced Concrete, Journal of the Prestressed Concrete Institute, V. 21,
No. 1, 1976, pp. 20-39.
6. Walraven, J. C.; Frenay, J.; and Pruijssers, A., Influence of Concrete
Strength and Load History on the Shear Friction Capacity of Concrete
Members, Journal of the Prestressed Concrete Institute, V. 32, No. 1,
1987, pp. 66-84.
7. Walraven, J. C., and Stroband, J., Shear Friction in High-Strength
Concrete, High-Performance Concrete, High-Performance Concrete,
SP-149, V. M. Malhotra, ed., American Concrete Institute, Farmington
Hills, MI, 1994, pp. 311-330.
8. Nagle, T. J., and Kuchma, D. A., Shear Transfer Resistance in HighStrength Concrete Girders, Magazine of Concrete Research, V. 59, No. 8,
2007, pp. 611-620.
9. Walraven, J. C., Fundamental Analysis of Aggregate Interlock,
Journal of the Structural Division, ASCE, V. 107, No. 11, Nov. 1981,
pp. 2245-2270.
10. Birkeland, P. W., and Birkeland, H. W., Connections in Precast
Concrete Construction, ACI JOURNAL, Proceedings V. 63, No. 3, Mar.
1966, pp. 345-368.
11. Raths, C. H., discussion of the paper, Design Proposals for
Reinforced Concrete Corbels, by A. H. Mattock, PCI Journal, V. 22,
No. 2, 1977, pp. 93-98.

426

12. Shaikh, A. F., Proposed Revisions to Shear-Friction Provisions,

Journal of the Prestressed Concrete Institute, V. 23, No. 2, 1978, pp. 12-21.
13. Hsu, T. T. C.; Mau, S. T.; and Chen, B., Theory of Shear Transfer of
Reinforced Concrete, ACI Structural Journal, V. 84, No. 2, Mar.-Apr.
1987, pp. 149-160.
14. Mau, S. T., and Hsu, T. T. C., Readers Comments on Influence of
Concrete Strength and Load History on the Shear Friction Capacity of
Concrete Members, Journal of the Prestressed Concrete Institute, V. 33,
No. 1, 1988, pp. 166-168.
15. Loov, R. E., and Patnaik, A. K., Horizontal Shear Strength of
Composite Beams with a Rough Interface, Journal of the Prestressed
Concrete Institute, V. 39, No. 1, 1994, pp. 48-69.
16. Mattock, A. L., Shear-Friction and High-Strength Concrete, ACI
Structural Journal, V. 98, No. 1, Jan.-Feb. 2001, pp. 50-59.
17. ACI 318-05, Building Code Requirements for Structural Concrete
(ACI 318-05) and Commentary (318R-05), American Concrete Institute,
Farmington Hills, MI, 2005, 430 pp.
18. Rahal, K. N., Simplified Design and Capacity Calculation of Shear
Strength in Reinforced Concrete Membrane Elements, Engineering
Structures, V. 30, No. 10, 2008, pp. 2782-2791.
19. Rahal, K. N., Shear Strength of Reinforced Concrete, Part II: Beams
Subjected to Shear, Bending Moment and Axial Loads, ACI Structural
Journal, V. 97, No. 2, Mar.-Apr. 2000, pp. 219-224.
20. Rahal, K. N., Torsional Strength of Reinforced Concrete Beams,
Canadian Journal of Civil Engineering, V. 27, No. 3, June 2000, pp. 445-453.
21. Rahal, K. N., Combined Torsion and Bending in Reinforced and
Prestressed Concrete beams Using SMCS, ACI Structural Journal, V. 104,
No. 4, July-Aug. 2007, pp. 402-411.
22. Vecchio, F. J., and Collins, M. P., Modified Compression Field
Theory for Reinforced Concrete Elements Subjected to Shear, ACI
JOURNAL, Proceedings V. 83, No. 2, Mar.-Apr. 1986, pp. 219-231.
23. Brstrup, M. W., Plastic Analysis of Shear in Reinforced
Concrete, Magazine of Concrete Research, V. 26, 1974, pp. 221-228.
24. Brstrup, M. W., discussion of Shear Strength of Reinforced
Concrete, Part I: Membrane Elements Subjected To Pure Shear, by K. N.
Rahal and closure by author, ACI Structural Journal, V. 97, No. 6, Nov.Dec. 2000, pp. 910-913.

ACI Structural Journal/July-August 2010

Appendix A
This Appendix provides detailed listing of the analyzed specimens and the results from three of the models
(SMCS, ACI code and Mattock16 Method).
Table A1: Properties of Test Specimens and Results of SMCS, ACI and Mattock16 Methods

Push-off 3

Pushoff 2

Push-off 1

Type
of
test

ID

fc

y fy-y

x fy-x

vexp

vSMCS

vMat

vACI

(MPa)
1.54
1.46
3.08
2.91
4.62
4.37
6.15
5.83
7.69
7.28
9.23
8.74
1.54
3.08
4.62
6.15
7.69
9.23
0.34
1.56
3.08
5.21
7.25
2.01
4.01
6.02
8.02
10.03
1.54
3.08
4.62
6.15
7.69
6.63
6.79
2.15
3.74
3.77
3.81
3.65
3.78
3.68
5.43
5.56
5.60

(MPa)

(MPa)

1.1A
1.1B
1.2A
1.2B
1.3A
1.3B
1.4A
1.4B
1.5A
1.5B
1.6A
1.6B
2.1
2.2
2.3
2.4
2.5
2.6
3.1
3.2
3.3
3.4
3.5
4.1
4.2
4.3
4.4
4.5
5.1
5.2
5.3
5.4
5.5
10.7
10.8
10.1
E1C
E2C
E3C
E4C
E5C
E6C
F1C
F4C
F6C

(MPa)
27.0
29.9
26.5
28.8
26.5
27.0
31.1
26.6
31.1
28.0
29.7
27.9
21.4
21.4
26.9
26.9
28.8
28.8
27.9
27.7
21.4
27.9
27.9
28.1
28.1
29.9
29.9
23.4
16.9
18.1
16.4
17.8
18.1
27.7
27.7
40.0
26.6
29.1
27.3
26.3
27.7
27.5
29.1
26.8
28.6

(MPa)
5.17
5.82
6.90
6.76
7.59
7.38
9.38
8.83
9.66
9.54
9.88
9.79
4.07
4.69
5.79
6.90
8.97
9.55
1.66
3.59
4.69
7.09
7.94
4.86
6.76
8.14
9.66
9.10
3.52
4.83
5.59
5.48
6.97
9.97
7.69
9.72
6.08
6.41
4.92
4.64
3.63
2.54
6.81
5.79
5.54

(MPa)
3.55
3.62
4.98
5.03
6.09
5.99
7.56
6.86
8.45
7.85
8.93
8.44
3.19
4.51
6.14
7.09
8.17
8.69
1.70
3.62
4.51
6.62
7.82
4.12
5.83
7.35
8.49
7.19
2.86
4.17
4.89
5.58
5.66
8.39
7.55
9.47
5.50
5.19
4.71
4.26
3.79
2.76
6.90
5.83
4.95

(MPa)
3.46
3.28
5.11
5.21
6.34
6.20
8.03
7.32
9.26
8.41
8.92
8.38
3.37
4.60
6.38
7.61
8.65
8.65
0.78
3.52
4.60
6.95
8.36
4.41
6.02
7.81
8.98
7.01
2.92
4.27
4.93
5.34
5.42
8.32
8.21
10.21
5.65
5.37
4.88
4.45
3.85
2.06
7.25
6.03
5.14

(MPa)
2.15
2.04
4.31
4.08
5.30
5.41
5.50
5.32
5.50
5.50
5.50
5.50
2.15
4.28
5.38
5.38
5.50
5.50
0.48
2.19
4.28
5.50
5.50
2.81
5.50
5.50
5.50
4.68
2.15
3.61
3.29
3.56
3.61
5.50
5.50
5.50
5.24
4.31
3.76
3.18
2.39
1.28
5.50
5.37
3.98

17.7

16.6#

2.67
0.00
5.61
0.00
-0.69
-1.12
-1.38
-2.07
-2.76
0.00
-1.38
-2.76

vexp

vexp

vexp

vSMCS

vMat

v ACI

1.46
1.61
1.39
1.34
1.24
1.23
1.24
1.29
1.14
1.22
1.11
1.16
1.28
1.04
0.94
0.97
1.10
1.10
0.97
0.99
1.04
1.07
1.02
1.18
1.16
1.11
1.14
1.27
1.23
1.16
1.14
0.98
1.23
1.19
1.02
1.03
1.10
1.23
1.05
1.09
0.96
0.92
0.99
0.99
1.12

1.49
1.78
1.35
1.30
1.20
1.19
1.17
1.21
1.04
1.13
1.11
1.17
1.21
1.02
0.91
0.91
1.04
1.10
2.13
1.02
1.02
1.02
0.95
1.10
1.12
1.04
1.08
1.30
1.20
1.13
1.13
1.03
1.28
1.20
0.94
0.95
1.07
1.19
1.01
1.04
0.94
1.23
0.94
0.96
1.08

2.40
2.85
1.60
1.66
1.43
1.36
1.71
1.66
1.76
1.74
1.80
1.78
1.89
1.10
1.08
1.28
1.63
1.74
3.43
1.64
1.10
1.29
1.44
1.73
1.23
1.48
1.76
1.95
1.63
1.34
1.70
1.54
1.93
1.81
1.40
1.77
1.16
1.49
1.31
1.46
1.52
1.98
1.24
1.08
1.39

Push-off 4
Push-off 5
Push-off 6
Push-off 7
Push-off 8

A1
A2
A3
A4
A5
A6
A6A
A7
N1
N2
N3
N4
N5
N6
110208t
110208
110208g
110408
110608
110808h
110808hg
110706
210204
210608
210216
210316
210808
120208
120408
120608
120808
120706
120216
230208
230408
230608
230808
240208
240408
240608
240808
250208
250408
250608
250808
10
11
12
13
14
15
glsh_4_13_25
glsl_2_13_25
glsh_4_13_35
glsl_2_13_35
g2sh_4_16_25

41.5
41.5
40.1
40.6
42.2
40.7
41.2
41.2
28.8
26.9
27.6
28.6
27.1
28.4
30.5
26.1
25.0
26.1
26.1
25.0
25.0
26.9
31.1
31.1
31.1
31.1
21.4
25.1
25.1
25.1
25.1
24.8
24.8
47.7
47.7
47.7
47.7
16.9
16.9
16.9
16.9
32.5
32.5
32.5
32.5
99.0
99.0
99.0
99.0
99.0
99.0
92.7
92.7
92.7
92.7
97.6

1.57
3.13
5.05
6.73
7.78
10.59
10.59
13.30
1.54
3.20
4.76
6.18
7.72
7.72
2.43
2.43
2.43
4.86
7.29
9.72
9.72
5.58
1.06
7.29
10.12
15.17
9.72
2.43
4.86
7.29
9.72
5.58
10.12
2.43
4.87
7.29
9.72
2.43
4.86
7.29
9.72
2.43
4.86
7.29
9.72
3.33
6.67
10.00
13.27
7.50
14.94
2.24
1.12
2.03
1.01
4.28

24.3
24.3
24.3
24.3
24.3
24.3

9.45##

5.84

5.24
5.52
7.93
9.79
10.34
12.14
12.83
13.38
3.17
5.38
6.62
7.93
8.10
8.21
5.08
5.50
5.08
6.44
7.39
8.39
8.58
7.19
3.22
9.72
9.25
10.11
7.97
5.36
6.53
6.78
7.31
6.92
6.53
6.72
10.83
12.56
14.19
4.65
6.04
6.55
6.29
6.83
8.69
9.65
9.94
6.30
11.20
15.00
18.10
11.00
17.61
4.21
2.01
3.70
2.53
5.62

4.32
6.11
7.65
8.87
9.70
11.14
11.20
11.84
3.66
5.11
6.30
7.30
7.97
8.57
4.71
4.39
4.31
6.21
7.61
7.64
7.64
6.75
3.14
8.23
9.29
9.29
6.63
4.31
6.10
7.47
7.66
6.51
7.59
4.79
6.78
8.30
9.58
3.60
5.09
5.32
5.32
4.79
6.78
8.30
9.55
8.58
12.14
14.87
17.13
12.88
18.17
3.44
2.44
3.12
2.20
4.76

3.52
6.66
8.05
9.44
10.45
12.21
12.35
12.35
3.48
5.25
6.56
7.81
8.14
8.52
5.00
4.55
4.44
6.50
7.83
7.50
7.50
7.16
2.39
8.94
9.33
9.33
6.43
4.45
6.40
7.52
7.52
6.95
7.45
5.47
8.66
10.60
12.54
3.64
5.07
5.07
5.07
5.19
7.14
9.08
9.74
7.49
10.84
13.50
16.12
11.50
16.55
4.57
2.28
3.74
1.87
8.60

2.19
4.38
5.50
5.50
5.50
5.50
5.50
5.50
2.16
4.48
5.50
5.50
5.43
5.50
3.40
3.40
3.40
5.22
5.22
5.00
5.00
5.39
1.48
5.50
5.50
5.50
4.28
3.40
5.02
5.02
5.02
4.96
4.96
3.40
5.50
5.50
5.50
3.38
3.38
3.38
3.38
3.40
5.50
5.50
5.50
4.66
5.50
5.50
5.50
5.50
5.50
2.84
1.42
2.33
1.16
5.43

1.21
0.90
1.04
1.10
1.07
1.09
1.15
1.13
0.87
1.05
1.05
1.09
1.02
0.96
1.08
1.25
1.18
1.04
0.97
1.10
1.12
1.06
1.03
1.18
1.00
1.09
1.20
1.24
1.07
0.91
0.95
1.06
0.86
1.40
1.60
1.51
1.48
1.29
1.19
1.23
1.18
1.43
1.28
1.16
1.04
0.73
0.92
1.01
1.06
0.85
0.97
1.22
0.82
1.19
1.15
1.18

1.49
0.83
0.98
1.04
0.99
0.99
1.04
1.08
0.91
1.02
1.01
1.02
1.00
0.96
1.02
1.21
1.14
0.99
0.94
1.12
1.14
1.00
1.35
1.09
0.99
1.08
1.24
1.20
1.02
0.90
0.97
1.00
0.88
1.23
1.25
1.18
1.13
1.28
1.19
1.29
1.24
1.32
1.22
1.06
1.02
0.84
1.03
1.11
1.12
0.96
1.06
0.92
0.88
0.99
1.35
0.65

2.39
1.26
1.44
1.78
1.88
2.21
2.33
2.43
1.47
1.20
1.20
1.44
1.49
1.49
1.49
1.62
1.49
1.23
1.42
1.68
1.72
1.33
2.17
1.77
1.68
1.84
1.86
1.58
1.30
1.35
1.46
1.39
1.32
1.98
1.97
2.28
2.58
1.37
1.79
1.94
1.86
2.01
1.58
1.75
1.81
1.35
2.04
2.73
3.29
2.00
3.20
1.48
1.41
1.59
2.17
1.04

Composite Beam 15

97.6
97.6
97.6
114.4
114.4
114.4
114.4
114.4
114.4
120.6
120.6
93.0
93.0
37.4
34.9
30.5
34.7
34.8
37.1
35.8
35.6
37.1
37.6
32.7
34.6
19.2
44.0
48.3

2.77
3.86
2.50
3.26
2.17
2.95
1.97
6.28
5.79
0.81
0.73
3.20
2.82
4.36
1.66
2.73
6.03
1.63
1.62
6.06
0.77
1.62
0.77
0.40
7.72
0.82
0.80
0.80

3.96
3.85
3.24
5.52
4.36
4.92
3.96
8.88
7.62
1.80
1.57
6.37
4.90
7.76
4.27
6.82
8.10
5.54
5.25
9.25
3.12
4.64
3.46
2.57
9.20
2.92
3.94
4.01

3.83
4.30
3.46
4.15
3.39
3.76
3.07
5.77
5.26
2.07
1.87
4.11
3.67
6.90
4.13
4.99
7.85
4.09
4.19
7.98
2.84
4.19
2.91
1.97
8.88
2.22
3.16
3.29

5.64
7.11
4.61
6.65
4.43
5.44
3.62
10.05
9.29
1.65
1.35
6.52
5.20
7.23
3.74
5.23
8.29
3.67
3.65
8.43
1.73
3.65
1.73
0.90
9.64
1.85
1.80
1.80

Longitudinal reinforcement relatively large, upper limit (=) is assumed.

g2sl_2_16_25
g2sh_4_16_35
g2sl_2_16_35
g3sh_4_13_25
g3sl_4_13_25
g3sh_4_13_35
g3sl_4_13_35
g4sh_4_16_25
g4sl_4_16_35
g5sh_2_10_25
g5sh_2_10_35
g6sh_4_16_25
g6sh_4_16_35
1
2
3
4
5
6
7
8
9
10
11
12
13
15
16

3.51
4.42
2.87
4.14
2.75
3.38
2.25
5.50
5.50
1.02
0.84
4.06
3.23
4.36
2.80
3.44
5.50
2.78
2.77
5.50
2.26
2.77
2.26
2.04
5.50
2.29
2.28
2.28

1.03
0.90
0.94
1.33
1.29
1.31
1.29
1.54
1.45
0.87
0.84
1.55
1.33
1.13
1.03
1.37
1.03
1.36
1.25
1.16
1.10
1.11
1.19
1.30
1.04
1.32
1.25
1.22

0.70
0.54
0.70
0.83
0.99
0.90
1.09
0.88
0.82
1.09
1.16
0.98
0.94
1.07
1.14
1.30
0.98
1.51
1.44
1.10
1.80
1.27
2.00
2.86
0.95
1.58
2.19
2.23

Near flexural compression zone, upper limit (=) is assumed.

based on assumed value of fy-x=460 MPa (69 ksi) (results do not change for any fy-x >240 MPa or 35 ksi)

##

calculated based on assumed value of fy-x=460 MPa (69 ksi)

Note: 1 MPa = 145 psi

1.13
0.87
1.13
1.34
1.58
1.45
1.76
1.61
1.39
1.76
1.87
1.57
1.52
1.78
1.53
1.98
1.47
1.99
1.89
1.68
1.38
1.67
1.53
1.26
1.67
1.27
1.73
1.76