Shear-friction truss model for reinforced concrete beams

S G Hong, Seoul National University, South Korea

T Ha*, Seoul National University, South Korea

26th Conference on OUR WORLD IN CONCRETE & STRUCTURES: 27 - 28 August 2001,

Singapore

Article Online Id: 100026034

The online version of this article can be found at:

http://cipremier.com/100026034

Thisarticleisbroughttoyouwiththesupportof

SingaporeConcreteInstitute

www.scinst.org.sg

AllRightsreservedforCIPremierPTELTD

YouarenotAllowedtoredistributeorresalethearticleinanyformatwithoutwrittenapprovalof
CIPremierPTELTD

VisitOurWebsiteformoreinformation

www.cipremier.com
 26 th Conference on Our World in Concrete & Structures: 27 - 28 August 2001, Singapore

Shear-friction truss model for reinforced concrete beams

S G Hong, Seoul National University, South Korea

T Ha*, Seoul National University, South Korea

Abstract

This paper presents a new model, called the "shear-friction truss model," for
slender reinforced concrete beams to derive a clear and simple design equation for
their ultimate shear strength. In this model, a portion of the shear strength is provided
by shear reinforcement as in the traditional truss model, and the remainder by the
shear-friction mechanism. Friction resistance is derived considering both geometrical
configuration of the rough crack surface and material properties. The inclined angle
of diagonal strut in the traditional truss model is modified to satisfy the state of
balanced failure, when both stirrups and longitudinal reinforcement yield
simultaneously. The vertical component of friction resistance is added to the modified
truss model to form the shear-friction truss model. Test results from published
literatures are used to find the effective coefficient of concrete strength in resisting
shear on inclined crack surfaces.

Keywords: shear-friction, truss model, reinforced concrete beam, aggregate interlock, shear crack

1. Introduction
The truss model adopted in current design codes is a good mechanical model to express the
behavior of reinforced concrete members after cracking. However, it cannot solely present the ultimate
strength of reinforced concrete beams, but must be accompanied by various other equations for
concrete contribution, which are usually derived from test results. For example, ACI 318-99 11.1.1
states that the nominal shear resistance Vn is:
(1 )
where Vc ' Vs are the shear components carried by concrete and shear reinforcement, respectively
[1].
While Vs can be calculated using truss analogy as described later, the prediction of Vc depends
only on the empirical equation which lumps the contribution of aggregate interlock, dowel action and
uncracked concrete together. For members subject to shear and flexure, the code specifies Vc as:

iKbwd
Vc = iJf:bwd (in SI unit) (2)

where bw is the web thickness and d is the effective depth of the member. Compared with the test
results on simply supported beams without web reinforcement, Eq.(2) generally underestimates the
shear capacity for beams with large longitudinal reinforcement ratios and overestimates the shear
strength for small steel percentages [2].
This paper presents a different approach to evaluating shear contribution by concrete in terms of
the shear-friction mechanism. It is assumed that rough crack surfaces playa major role in shear

309
capacity of cracked concrete and friction resistance is proportional to the strength of concrete, and
contact area. Based on these assumptions, a simple friction equation is derived and is added to the
truss model with modified crack angle.

2. Review of the conventional truss model

The response of uncracked reinforced concrete members to applied loads can be understood by
using elastic analysis. However, cracking of the concrete causes stress redistribution and the elastic
beam theory cannot be applied. Truss analogy, developed in the early twentieth century, visualizes the
load path in the cracked reinforced concrete beam as a statically determinate truss. It is an excellent
conceptual model to show the force flows in cracked concrete members.
As shown in Figs. 1 and 2, the truss model consists of the several vertical tension members and
diagonal compression members running parallel to the inclined shear crack. The stirrups crossing a
crack are collected into a vertical tension member, and all the concrete struts meeting at any vertical
section playa diagonal compression member. The traditional truss model assumes a 45 constant
angle of inclination as a conservative design approach. From the free body diagram of Fig. 1, the
shear force can be calculated as:
v= A/syjd
(3)
s
where At is the area of vertical reinforcement, fSY is the yield strength of the stirrups, jd is an
internal lever arm and s is the spacing of stirrup legs.
Although the analysis of reinforced concrete beams using the traditional truss model is clear and
simple, this model disregards the shear transferred by the aggregate interlock along the crack suriace,
the dowel action of longitudinal reinforcement and the uncracked compression zone [2]. The ACI code
uses Eq.(3) for Vs except that jd is replaced with d.

r - - - - - - - - - - - - - - . .............
~~-_r-,_-r_~-_r~~=-I --~
~-ICI
I~""'" .,
i
. - - - - - - - - - - - - - - ................"." '1 I
V ~f ~fsy: I I,

. . . .=. . . . . . . . . .:.:.:.::i/1
: I
Tv
~~=======r.=+.:::-!
. . .~ . . ~fsy
/
.....T-:.y- -- !. - - !. - - ;:
................................................ ~~==~ . . .= ~~ .
v ............................;

Fig. 1 Internal forces at an inclined crack Fig. 2 Equivalent truss of a beam

Fig. 3 Crack patterns of slender reinforced concrete beam [3]

3. Shear-friction truss model

When a slender reinforced concrete beam is loaded, flexural cracks appear and extend vertically
up to the neutral axis of the beam. With increased loading, the flexural cracks develop to shear cracks
or new inclined shear cracks initiate in the uncracked region. This inclined shear cracks always occur
before shear failure and progressively cover the entire depth in the ultimate state (Fig. 3).
Additional components other than shear reinforcement (truss action) begin to resist shear loading
after inclined shear cracks appear, namely aggregate interlock along the crack surface, the dowel
action of longitudinal reinforcement and the uncracked compression zone. Although these components
have different mechanisms in resisting shear loading, their contribution can be explained commonly by

310
 (a) Frictional forces on the crack surface (b) Infinitesimal element on the crack surface
Fig. 4 Shear-friction mechanism
the shear transfer across rough crack surfaces by the friction mechanism. As the inclined cracks
develop into the uncracked region, the two cracked parts of the beam move relatively to each other as
shown in Fig. 4(a). This relative displacement along the rough crack surface is resisted mainly by
aggregate interlock, and the shear reinforcement prevents excessive opening of the shear crack to
make the interlocking mechanism possible. Both the uncracked compression zone and the dowel
action also provide additional confinements to the displacements. Shear failure occurs when crack
widths become so large that the friction mechanism no longer controls the relative movement of the
two cracked parts.
Fig. 4(a) shows the friction forces and two infinitesimal elements involving the crack surface.
These two elements are under different states of stress. Since uniaxial tension is applied to the lower-
left element, it cannot transmit friction forces across the crack. On the contrary, the upper-right
element is under uniaxial compressive stress which is in equilibrium on the crack surface .
. 2
(ja = (jc Sin a (4)
'l"a = (jc sinacosa (S)
where (jc is uniaxial stress along the average crack orientation, and (ja' 'l"a are the frictional
stresses normal and tangential to the crack surface, respectively.
Since shear-friction resistance is the sum of 'l"a in the direction of average crack path, its
magnitude depends on the COSa of the deviation angle of the individual contact surface. As shown in
Fig. S, crack orientation constantly changes from the average crack path due to the roughness of
shear crack. For the relative displacement of the two cracked bodies shown in Fig. 4, parts of the
crack surfaces which deviate in the counter-clockwise direction contribute to the shear-friction
mechanism. If the crack deviations across the entire crack surface are assumed to cancel out, only
half of the crack surface can be used to evaluate friction force, and shear-friction in the direction of
average crack path, 'l"f' is obtained by calculating the mean value of the counter-clockwise deviation
angle. Various probability density functions, called "contact density functions," have been used to
describe the distribution of local deviation angle. These functions p(O) have the following properties
(3):
a. Symmetry about 0 = O.
b. p(O) = 0 outside the region -tr/2::;; 0 ::;; +tr/2

c. [!!p(O)dO = 1
2

Li and Maekawa found that the cosine function satisfies all the required properties of the contact
density function and suggested taking p(O) = O.Scos(O) in the region -tr/2::;; 0::;; +tr/2. Their final
model was reported to produce excellent correlation with specimens having f;::;; SO MPa (3). However,

as(-)

Fig. S Crack orientations in a shear crack

311
since the crack profiles are less tortuous for higher strength, a different function is needed for high
strength concrete. This paper uses Li and Maekawa's contact density function to find the average
angle of local crack deviation, and depends on the effective coefficient to explain the effect of concrete
strength.

gacosada
aavg = = 32.7 (6)
gcosada
'!"f = CTc sin32.7cos 2 32.7 = 0.206CTc (7)
According to Eq.(7), shear-friction capacity of a rough surface is only about 20 percent of the
uniaxial compressive stress present in the direction of average crack angle. This uniaxial compression
is in the same direction as the diagonal compression strut in the traditional truss model but is resisted
by the cracked concrete element.
Based on the above arguments, the following assumptions are made to describe the overall
response of reinforced concrete beams to shear loading.
a. The shear-friction mechanism is responsible for the additional shear capacity to the truss
model for slender reinforced concrete beams with inclined shear cracks.
b. The dowel action of longitudinal reinforcement and the uncracked compression zone
contribute to the prevention of excessive opening of the inclined shear cracks, and are
included in the shear-friction mechanism.
c. Frictional resistance is uniformly distributed along the effective contact area of the crack
surface and depends on the strength of the concrete.
The first assumption can be expressed as:
V=~+v, (8)
where ~, V, are the shear components carried by truss action and friction mechanism, respectively.
According to the second assumption, Vf comprises the resistance by aggregate interlock, dowel
action and the uncracked compression region. When combining Eq.(7) with the concept of effective
contact area (Ae ), the magnitude of friction can be calculated as:
F= '!"fAa
= 0.206CT bw (0.5d) (9)
c sin(}

= 0.1 03v,f; ~wd

sm(}
where CTc is replaced by v,f;, the effective strength of concrete in resisting shear on an inclined
crack surface.
Eq.(9) is different from the usual friction equation, which has been adopted by many codes and
previous researches and is given by:
F=J1N (10)
where J1 is the coefficient of friction and N is the normal compressive force on the contact surface.
Because this normal force on the contact surface is indispensable for the friction mechanism to be
available, there have been various attempts to determine the source of normal forces. These forces
include shear-resisting forces of shear reinforcement and tensile forces of longitudinal reinforcement.
However, the shear-friction mechanism on the inclined shear cracks of reinforced concrete members
is different from the usual friction theory in that it involves force transfer across the crack surface, and
thus depends more on the concrete strength than the normal forces on the contact surface.
The truss component ~ is calculated using the same approach as that of the traditional truss
model except that the inclined angle of the diagonal strut, i.e. the crack angle, is determined for the
case of balanced failure of the stirrup, crack surface and longitudinal reinforcement. For the free body
diagram of Fig. 6, equilibrium requirements for the left free body yield T' = 0.5~ cot(} + F cos () .
Summing the moments of forces applied to the right free body about the point at the upper right end
gives:

A/yd - T'd +FCOS(}.~ -(~ + FSin(})( a- ~ cot(}) = 0 (11 )

312
 elf
d

T'
~==========='--'T -
T'-

Fig. 6 Free body diagram of cracked reinforced concrete

Because this quadratic equation for colO involves an unknown variable, v" an iteration process is
needed to calculate both (J and v,. The crack angle has a maximum value if only the longitudinal
reinforcement yields and has a minimum value when only stirrups crossing the inclined crack yield.
Therefore, the crack angle gets larger as the amount of stirrup increases. When (J is determined, ~
is given as:
v. = A/syd
(12)
t stan(J
Considering that V, is the vertical component of the friction force F, Eq.(8) can be rewritten as:

V = A/sy jd + 0.1 03v,f;bwd (13)

stan(J
Inclined crack angle (J and effective coefficient v, are obtained by the following procedure:
1. Assume a linear relationship of f; and v,: v, = A . f; + B
2. Calculate (J and V using Eq.(11) and Eq.(13).
3. Find new coefficients A, B using linear regression of f; and v,.
4. Repeat steps 2 and 3 until satisfactory convergence is reached.

4. Effective coefficient of concrete strength

Many researches on the shear-friction mechanism were carried out by direct shear tests or
precracked pushoff tests. Test results show that shear-friction resistance increases with the concrete
strength and the confinement effect. However, since the general configuration of the specimens is
quite different from that of a beam under transverse loading, most of the test results are not directly
applicable in evaluating the shear capacity of a beam. The shear transfer mechanism of the well
known shear-friction specimens is developed only by the aggregate interlock along a precracked
straight surface confined by tension ties. It does not include the effect of dowel action or uncracked
compressive region.
Table 1. Comparison with test results by Mphonde [4, 5]
Specimen
No.
b d aid f'c Av s fy Vexp (Jpre Vpre Vexp/Vpre
B50-3-3 150 298 3.16 22.1 16 86 302.7 76.1 22.4 78 0.977
B50-7-3 150 298 3.16 39.8 16 86 302.7 93.9 25.2 93 1.013
B50-11-3 150 298 3.16 59.7 16 86 302.7 97.9 27.0 102 0.958
B50-15-3 150 298 3.16 83 16 86 302.7 111.2 27.1 102 1.086
B100-3-3 150 298 3.16 27.9 36 86 265.8 95.2 29.1 104 0.915
B100-7-3 150 298 3.16 47.1 36 86 265.8 120.5 31.4 117 1.028
B100-11-3 150 298 3.16 68.6 36 86 265.8 151.7 32.4 123 1.229
B100-15-3 150 298 3.16 81.9 36 86 265.8 115.6 32.2 122 0.944
B150-3-3 150 298 3.16 28.7 52 86 284.2 139 33.7 122 1.135
B150-7-3 150 298 3.16 46.6 52 86 284.2 133.4 35.5 134 0.993
B150-11-3 150 298 3.16 69.5 52 86 284.2 161.5 36.4 141 1.148
B150-15-3 150 298 3.16 82.7 52 86 284.2 150 36.2 140 1.075
Note: The units are mm for length, mm 2 for area, kN for force and MPa for stress

313
 ~
118" or 3/16" or
118" + 3/16" stirrups at
3.5" spacing
. /

I 11111111111111111111111111 I I ~rmLlI
l' 84 in
f 3 No. : \ " cover on A.

>
90 in
Fig. 7 Specimen detail [4]
Table 1 and Fig. 7 give the specimen details and test results of twelve simply supported reinforced
concrete beams loaded at midspan [4]. These beams have an identical sectional property and span
ratio, but differ in concrete strength, which ranges from 22 MPa to 83 MPa, and the amount and yield
strength of stirrups. The effect of concrete strength and shear reinforcement on the shear-friction
mechanism is shown in Fig. 8(a) and (b). Shear-friction components of the measured shear strength
are obtained by subtracting truss components from the test results. As shown in the figure, shear-
friction resistance of a beam increases for higher concrete strength, but shows only marginal increase
as the amount of shear reinforcement increases. This relationship between concrete strength and
friction component confirms the third assumption, and the effective coefficient of concrete strength in
resisting shear-friction is developed by following the steps explained in the previous section as shown
in Fig. 9.
v, = -0.003f; + 0.431 (f; in MPa) (14)

120 Vexp -Vs (kN) 120 V-V

exp s
(kN)

90 90


60 60
30 30
f; (kN) A/y (kN)
o 20 40 60 80 100 o 5 10 15 20

(a) (b)

Fig. 8 Effect of (a) concrete strength and (b) amount of shear reinforcement on shear-friction

V f 0.5

0.4

0.3

0.2

0.1
0 20 40 60 80 100
f'c
Fig. 9 Effective coefficient of concrete strength in resisting shear-friction

314
 (a) Normal strength concrete (b) High strength concrete
Fig. 10 Crack profiles

The effective coefficient is found to decrease linearly as the concrete strength increases, which is
in agreement with other experimental results showing that the shear-friction mechanism is less
effective in high strength concrete than in normal strength concrete [3, 7]. The theoretical basis for this
phenomenon is strongly related to the properties of the constituents of concrete. As stated earlier, the
shear-friction mechanism involves force transfer across crack surfaces. For normal strength concrete,
these crack surfaces are formed along the interface of the aggregate and the cement matrix due to the
higher stiffness of the aggregate than the cement matrix (Fig. 10(a. Failure of the interface to
transmit shearing force results from loss of contact when crushing of the cement matrix occurs. In the
case of high strength concrete, cracks usually penetrate into individual aggregate particles to form
much smoother profiles as shown in Fig. 10(b).
The test results are compared with the predicted values by the shear-friction truss model (Eq.(13
and the ACI code (Eq.(1 in Fig. 11. While the predicted values according to the ACI code are very
conservative compared with the measured values, those of the shear-friction truss model are in
accordance with them for all ranges of concrete strength. The mean value and standard deviation of
the measured Ipredicted values are 1.042 and 0.091, respectively. Fig. 11 also shows that Eq.(13)
underestimates the capacity of beams with higher stirrup ratio compared with beams with low stirrup
ratio. This tendency can be explained if the equation for the effective coefficient (Eq.(14 is improved
to reflect the marginal increase of shear-friction resistance for higher shear reinforcement amount as
shown in Fig. 8. Shear-friction contribution is about 53 percent of overall shear capacity and has
higher values for members with higher concrete strength and lower shear reinforcement as shown in
Fig. 12. This again proves the strong correlation of the concrete strength with the friction resistance.
o AC1318-99
Shear-friction prediction V,/V
pred
v.,xp
-1.6 (%)
Vpred
~
1.1
l
1.4
I
!

12

! 1
I
I .,
I I
0.$ 0 1 2 3 4 5 6 7 * 9 10 11 12 13
Specimen No. Specimen No.

Fig. 11 Comparison with test results Fig. 12 Shear-friction contribution

5. Conclusion
1. Existing truss models are generally conservative and incomplete in predicting the ultimate
shear strength of reinforced concrete beams.
2. The shear-friction truss model is developed to accommodate the contribution by the dowel
action of longitudinal reinforcement, aggregate interlock, and uncracked compression zone. A
shear-friction component is added to the traditional truss model which is modified to satisfy

315
 the balanced failure of stirrups and longitudinal reinforcement. Shear-friction resistance is
proportional to concrete strength and contact area. The effective coefficient is introduced to
describe the influence of concrete strength on the shear-friction mechanism. The shear-
friction model is well in agreement with test results.
3. Shear-friction resistance assumes about 50 percent of total shear strength, and is more
effective in the case of normal strength concrete than high strength concrete.

Acknowledgements
This work was supported by the Brain Korea 21 Project. Opinions, findings, and conclusions in
this paper are those of the authors and do not necessarily represent those of the sponsors.

References:
[1] ACI Committee 318, "Bui/ding Code Requirements for Reinforced Concrete (ACI 318-99}",
American Concrete Institute, 1999.
[2] MacGregor, James G., "Reinforced Concrete, Mechanics and Design, 3rd Edition", Prentice Hall,
1997.
[3] Ali, Mohamed A., White, Richard N., "Enhanced Contact Model for Shear Friction of Normal and
High-Strength Concrete", ACI Structural Journal, American Concrete Institute, 1999, 348-360.
[4] Chen, Simon A., "A Shear-Friction Truss Model for Reinforced Concrete Beams Subjected to
Shear", Ph.D. Thesis, University of Alberta, 1993.
[5] Mphonde, Andrew G., "Use of Stirrup Effectiveness in Shear Design of Concrete Beams", ACI
structural Journal, American Concrete Institute, 1989, 541-545.
[6] Peng, Liying, "Shear strength of Beams by Shear-Friction", Master Thesis, University of Calgary,
1999.
[7] Walraven, Joost, Frenay, Jerome, and Pruijssers, Arjan, "Influence of Concrete Strength and Load
History on the Shear Friction Capacity of Concrete Members", PCI Journal, Prestressed Concrete
Institute, 1987, 66-84.

316