You are on page 1of 12

ACI STRUCTURAL JOURNAL TECHNICAL PAPER

Title no. 110-S48

Strength Modeling of Concrete Deep Beams Reinforced


with Internal Fiber-Reinforced Polymer
by Matthias F. Andermatt and Adam S. Lubell

It is well-known that arch action in deep concrete members Lack of design provisions to acknowledge the additional
provides a beneficial increase in strength. However, while specific capacity resulting from arch action may result in uneco-
design code provisions are available for steel-reinforced deep nomical designs with excessively large cross sections for
members that account for this influence, current design code provi-
members without web reinforcement, especially in cases
sions for members containing internal fiber-reinforced polymer
(FRP) reinforcement do not consider the capacity obtained from where large concentrated loads are located close to support
arch action. Strut-and-tie-based modeling approaches for FRP- locations. Further, ignoring the contribution from arch action
reinforced members were developed through the adaptation of may suggest that unneeded strengthening is required when
existing models for steel-reinforced concrete deep members. evaluating the shear capacity of existing FRP-reinforced
When compared to test results for large-scale members with deep members.
small shear span-depth ratios (a/d), the recommended strut-and- A companion paper14 presented new test data for 12 large-
tie model (STM) is shown to be in better agreement than existing scale deep beams with internal FRP reinforcement that
shear provisions for FRP-reinforced members. The proposed STM
is also shown to accurately account for influences on capacity confirmed the presence of arch action. This paper reports
from the reinforcement ratio, reinforcement stiffness, and overall the development of appropriate modeling strategies for
member height. FRP-reinforced members with small a/d.15 A strut-and-tie
modeling approach is developed for use in predicting the
Keywords: deep beams; fiber-reinforced polymer reinforcement; reinforced capacity of FRP-reinforced deep beams through mechanics-
concrete; shear; shear span-depth ratio (a/d); strut and tie.
based modifications to existing STMs applicable to steel-
reinforced concrete construction. The proposed STM is vali-
INTRODUCTION
dated with the new test results for large-scale members14 and
The shear behavior of reinforced concrete members
other data available from moderate-scale tests reported in
without transverse reinforcement has been studied exten-
sively by many researchers over the last century.1 For steel- the literature.16,17 Comparisons are also made to capacity
reinforced concrete members, it has long been established predictions using existing sectional shear design models for
that there is a substantial change in the primary load-carrying FRP-reinforced slender members.
mechanisms as the shear span-depth ratio (a/d) approaches
2.5.2,3 Design codes and guidelines for steel-reinforced RESEARCH SIGNIFICANCE
construction acknowledge this difference in behavior and Internal FRP reinforcement has been used in concrete
specify different analytical capacity methods depending structures that are exposed to aggressive environments,
on the a/d.4-7 So-called “slender members” are commonly where use of corrosion-susceptible steel reinforcement may
defined as having an a/d greater than approximately 2.5 and introduce concerns regarding long-term durability. These
are assumed to have axial strain distributions that vary linearly structures may include bridges, parking garages, founda-
over the height of the cross section according to the well- tions, or those located in marine environments. Based on
known principle that plane sections remain plane. Design the member geometry or the loading conditions, members
code provisions typically use sectional models to determine within these structures may contain so-called disturbed
the shear and flexural capacities of slender members.1 When regions, where sectional design methods do not apply due
the a/d is less than approximately 2.5—herein termed “deep to the nonlinear distribution of strains. There is currently no
beams”—a nonlinear distribution of strains dominates the guidance for the special design of these disturbed regions
response and arch action becomes the primary force-transfer in members containing internal FRP reinforcement.4,11-13 By
mechanism following diagonal cracking.1 Strut-and-tie considering the case of so-called deep beams with small a/d,
models (STMs) represent one technique commonly adopted this paper develops a modeling approach based on the strut-
to design steel-reinforced deep beams.8-10 and-tie method. The model is validated for different FRP
While different shear models are used to analyze the reinforcement types and a range of common beam design
capacity of slender and deep steel-reinforced concrete configurations, with the results compared to strength predic-
members, current codes and guidelines for the design of tions using currently available sectional models.
fiber-reinforced polymer (FRP) reinforced concrete members
do not differentiate between these member types.4,11-13 Only
sectional shear design methods are provided and no sepa- ACI Structural Journal, V. 110, No. 4, July-August 2013.
rate analytical models are given to directly account for the MS No. S-2011-227.R1 received May 2, 2012, and reviewed under Institute
publication policies. Copyright © 2013, American Concrete Institute. All rights
arching action observed to occur in deep concrete members reserved, including the making of copies unless permission is obtained from the
internally reinforced with FRP.14 CAN/CSA S806-0211 even copyright proprietors. Pertinent discussion including author’s closure, if any, will be
published in the May-June 2014 ACI Structural Journal if the discussion is received
states that “analysis by strut and tie models is not permitted.” by January 1, 2014.

ACI Structural Journal/July-August 2013 595


ACI member Matthias F. Andermatt is a Bridge Engineer at AECOM, Edmonton, effective crack spacing sze. The parameter sze accounts for
AB, Canada. He received his BSc in civil engineering and his MSc in structural engi- the reduction in aggregate interlock that occurs as the crack
neering from the University of Alberta, Edmonton, AB, Canada. His research interests spacing increases with member depth and as the aggregate
include large-scale testing of structural components and shear transfer in concrete.
size ag decreases. Because cracks in high-strength concrete
ACI member Adam S. Lubell is a Project Engineer at Read Jones Christoffersen tend to go through the aggregate rather than around, ag is
Ltd., Vancouver, BC, Canada, and an Associate Adjunct Professor of civil engineering linearly reduced to zero as fc′ increases from 60 to 70 MPa
at the University of Alberta. He received his PhD from the University of Toronto, (8700 to 10,150 psi).
Toronto, ON, Canada. He is Secretary of ACI Subcommittee 445A, Shear and Torsion-
Strut and Tie; and a member of ACI Committees 440, Fiber-Reinforced Polymer
Reinforcement; 544, Fiber-Reinforced Concrete; and Joint ACI-ASCE Committee Hoult et al.21
445, Shear and Torsion. His research interests include the design and rehabilitation The CSA S6-0618 and CSA A23.3-045 shear models for
of reinforced and prestressed concrete structures, and the development of structural steel-reinforced members use a so-called first-order (that
detailing guidelines to allow the use of high-performance materials. is, linear) approximation to the crack-width-versus-longitu-
dinal-strain relationship. This simplification is appropriate for
steel-reinforced concrete members due to the limited range
EXISTING SECTIONAL SHEAR MODELS of axial strains encountered for traditional reinforcing steel
As noted previously, there are no separate analytical strengths.20 To better account for the typically higher axial
models provided in current design standards4,11-13 that strains in FRP-reinforced members, Hoult et al.21 developed
allow direct consideration of the arch action observed in a so-called second-order (that is, nonlinear) approximation
FRP-reinforced members with small a/d.14,16,17 Instead, all to the diagonal crack width relationship. This model predicts
designs should be checked with sectional models for shear the shear capacity for members reinforced with only longitu-
and flexure. dinal reinforcement as
Five existing sectional shear design models for concrete
members reinforced with internal FRP are examined in this
0.30 1300
paper to understand their applicability for use in predicting Vc = ⋅ fc′bw dv (5)
the shear capacity of FRP-reinforced members with small 0.5 + (1000 e x + 0.15)
0.7
(1000 + sze )
a/d. The models are shown with the resistance factors taken
as unity and are provided in SI units (1 mm = 0.0394 in.;
1 N = 0.225 lb; 1 MPa = 145 psi). 31.5d
sze = ≥ 0.77d (6)
CSA S6-06 (Clause 16)4,18 16 + ag
The CSA S6-064,18 method for determining the shear
capacity of FRP-reinforced concrete beams is a mechanical The effective depth dv is taken as 0.9d and ex is given in
model based on simplifications to the Modified Compres- Eq. (2).
sion Field Theory (MCFT).19 This model is similar to the
technique used in the shear design of slender steel-reinforced CAN/CSA S806-0211
concrete members,4,5 as described elsewhere.20 The nominal For FRP-reinforced concrete members having an effective
shear capacity of FRP-reinforced slender beams without depth not exceeding 300 mm (11.8 in.), the shear capacity is
transverse reinforcement Vc is given as calculated as

1/3
0.40 1300 (1)  Vd 
Vc = 2.5 ⋅ f b d Vc = 0.035  fc′rEfrp bw d > 0.1 fc′bw d (7)
(1 + 1500e x ) (1000 + sze ) cr w v  M 

where the quantity Vd/M at the section under consideration


M dv + V is not to be taken as greater than 1.0; and Vc is not to exceed
ex = ≤ 0.003 (2) 0.2 fc′bw d .
2 E frp A frp
The shear capacity of members having an effective depth
exceeding 300 mm (11.8 in.) is calculated using Eq. (8),
which accounts for the observed decrease in shear strength
35dv with member depth.
sze = ≥ 0.85dv (3)
15 + ag
 130 
Vc =  fc′ bw d ≥ 0.08 fc′ bw d (8)
 1000 + d 
fcr = 0.4 fc′ ≤ 3.2 MPa (4)

ISIS Design Manual No. 3, Version 212


where dv is the effective shear depth taken as the greater of This shear model is similar to the CSA A23.3-9422 simplified
0.72h and 0.9d. The values for the moment M and shear V method of determining the shear capacity in steel-reinforced
in Eq. (2) are always taken as positive with M not taken as members but incorporates a reduction factor to account for
less than Vdv. The model assumes that the shear capacity can the stiffness of the longitudinal FRP reinforcement relative to
be predicted by the aggregate interlock mechanism for shear steel. The model predicts the shear capacity of FRP-reinforced
force transfer across a crack. The aggregate interlock mecha- members containing no transverse reinforcement and having
nism is dependent on the axial strain at midheight ex and the an effective depth less than 300 mm (11.8 in.) as

596 ACI Structural Journal/July-August 2013


Fig. 1—Beams in four-point bending.

equal to the member depth. Thus, in the case of a deep beam


E frp with an a/d of less than 2, the entire shear span will be a
Vc = 0.2 fc′ bw d (9)
Es D-region (refer to Fig. 1(a)). Strut-and-tie modeling is a
widely used technique to analyze D-regions and is described
For beams that have an effective depth greater than 300 mm in detail elsewhere.1,8-10,24,25 Note that STMs can also be used
(11.8 in.), Vc is calculated as to represent the so-called B-region further away from the
discontinuity where the axial strain distribution is assumed
to be linear and sectional shear models would also apply.1
 260  E frp The use of STMs allows for easy visualization of the
Vc =  fc′ bw d (10)
 1000 + d  Es flow of forces. In addition, these truss models represent all
internal force effects and do not require separate flexure and
shear models, as is the case for slender members analyzed
with sectional approaches. Strut-and-tie modeling is based
ACI 440.1R-0613
The ACI 440.1R-0613 shear model was developed23 such on the lower-bound theory of plasticity and the capacity of
that the shear capacity of both steel- and FRP-reinforced the model is always less than the true capacity if the truss is in
members could be predicted by a similar approach that equilibrium and has sufficient deformation capacity to allow
directly accounts for the reinforcement stiffness. The shear redistribution of forces into the assumed truss elements.10
capacity of concrete members is given as For the STM technique, the flow of forces in a simply
supported deep member after cracking is idealized using
a pin-jointed truss consisting of compression struts and
2 tension ties (Fig. 1(b)). The struts and ties intersect at nodes.
Vc = fc′ bwc (11)
5 Struts are used to represent the assumed compressive stress
fields in the concrete. Ties represent the primary tension
where c = kd is the cracked transformed section neutral axis reinforcement with the tie location made to correspond to
depth with k given by Eq. (12); and nf is the modular ratio of the centroid of the reinforcement. While the tie also consists
the FRP to the concrete. of the concrete surrounding the reinforcement, this concrete
is not directly considered in the design but will reduce the tie
elongation.6 It is assumed that no bond exists between the tie
( )
2
k = 2rn f + rn f − rn f (12) and surrounding concrete between nodal locations giving a
constant tie force between adjacent nodes. For the STM of a
deep beam under four-point bending shown in Fig. 1(b), this
allows use of a tied-arch analogy to represent the forces in
STRUT-AND-TIE MODELS the struts and main tension tie.
Nonlinear strain distributions in concrete members are The STM technique requires that all ties, struts, and
caused by changes in geometry or loading and are commonly nodal regions have sufficient strength relative to the forces
referred to as disturbed regions or D-regions.10 According to imposed on these elements by the selected model geometry.
ACI 318-086 and other codes of practice for steel-reinforced With reference to Fig. 2, the axial strength of a tie is given by
concrete members, the D-region will typically extend longi-
tudinally from the discontinuity for a distance approximately Ft = At ft (13)

ACI Structural Journal/July-August 2013 597


Fig. 2—Nodal region of STM shown in Fig. 1.

where At is the cross-sectional area of reinforcement; and ft


e1 = e s + ( e s + 0.002) cot 2 q (16)
is the limiting stress in the reinforcement. The strength of a
strut is given by
where the parameter q represents the angle between the strut
Fs = Acs fc (14) axis and the axis of the adjoining tie; and es is the tensile
strain in the tie crossing the strut at the strut location. From
Eq. (15) and (16), it is observed that as the strut angle
where Acs is the applicable cross-sectional area of concrete decreases or as the tensile reinforcement strain increases, the
taken perpendicular to the strut axis; and fc is the maximum compressive load-carrying capacity of the strut is reduced.
permitted stress in the concrete strut. Note that the appli- Thus, the member capacity can decrease with increases in
cable cross-sectional area and the compression stress limit the a/d if the member capacity is governed by the strength of
may both be different at each end of the strut. The strength the diagonal strut in Fig. 1(b). Further, the member capacity
of nodal regions must be such that the stress is kept below a can be influenced by the axial stiffness of the reinforcement.
fraction of fc′ where the fraction is dependent on the number For steel-reinforced concrete members, the tie strength
of ties crossing a nodal zone, as described in a later section. in the CSA A23.3-045 STM is evaluated based on the yield
The nodal stresses are calculated using the area of the nodal strength of the steel. In adapting these provisions to FRP-
zone face that is perpendicular to the force acting on the face. reinforced members, it is necessary to directly consider the
While the basic strut-and-tie modeling principles are linear elastic stress-strain response of the FRP to evaluate
similar in each design code for steel-reinforced concrete efrp. Note that efrp is the strain in the FRP reinforcement away
members,4-7 they can differ significantly with respect to the from the nodal zone (that is, to the right of Location d in
evaluation of the limiting capacities of each element. Three Fig. 2(a)).
different STMs were evaluated in this study. Two of the For the STM shown in Fig. 1(b), the midspan tie strain (that
models were based on the CSA A23.3-045 provisions, while is, at Location a in Fig. 1(b)) is taken as efrp and calculated
the third model was based on the STM provisions given directly from the corresponding tie force and reinforcement
in ACI 318-08,6 Appendix A. These models were selected properties, as described in a later section. For simplicity (and
because they can highlight inherent model assumptions that referred to herein as “CSA Approach 1”), the tie strain es in
may impact their applicability for FRP-reinforced members, Eq. (16) is set equal to efrp. Shear capacities calculated with
as FRP mechanical properties can be significantly different CSA Approach 1 are given the notation Vcsa1. According to
than conventional steel reinforcing bars. CSA A23.3-04,5 Commentary Clause N11.4.2.3, the value
of es can be taken as the tie strain magnitude at the center-
CSA A23.3-045 STM provisions line of the strut in cases where the strain varies across the
The CSA A23.3-045 STM does not impose a limitation strut width. Thus, for the case in Fig. 2(a) where the tie force
on the angle between adjoining struts and ties. Instead, a changes from zero at Location b to the maximum tie force
MCFT-based approach is used to establish the maximum at Location d, the tie strain at the intersection with the strut
compressive stresses permitted in struts by considering centerline (Location c) is approximated as es = 0.5efrp. This
the influence of cracking caused by coexisting transverse is referred to as “CSA Approach 2” and the corresponding
tensile strains.19 It is noted that the STM provisions in capacity predictions are given the notation Vcsa2.
CSA S6-064 and AASHTO LRFD7 use a similar MCFT- CSA A23.3-045 also requires that the stresses in the nodal
based approach. The limiting compressive stress in a strut zones not exceed 0.85fc′ for CCC nodes (compression only),
crossed by a tie is given by 0.75fc′ for CCT nodes (compression with one tie), and
0.65fc′ for CTT nodes (compression with ties in more than
fc′ one direction). For new construction, an orthogonal grid of
fc = ≤ 0.85 fc′ (15) crack control reinforcement is required near each face of the
0.8 + 170 e1
member. The ratio of the reinforcement area to the gross
concrete area is to be at least 0.002 in each direction and the
where e1 is the transverse tensile strain calculated as maximum bar spacing is not to exceed 300 mm (11.8 in.).

598 ACI Structural Journal/July-August 2013


However, this reinforcement need not be directly consid-
ered within the truss model calculations and the potential
impact of omitting this reinforcement on the overall capacity
is ignored for the members without web reinforcement under
consideration in this paper.

ACI 318-086 STM provisions


Unlike the CSA A23.3-045 STM, the ACI 318-086 STM
limits the minimum angle q between the axis of any strut
and tie entering a node to 25 degrees. This limit is to avoid
incompatibilities due to the shortening of the strut and
lengthening of the ties. The limiting compressive stress in
a strut containing no longitudinal reinforcement and regard-
less of whether it is crossed by a tension tie is given by

fc = 0.85b s fc′ (17)

where bs = 1.0 for struts of uniform cross-sectional area


over its length, 0.75 for bottle-shaped struts with distrib-
uted reinforcement, and 0.6 for bottle-shaped struts without
distributed reinforcement. The limiting strut stress is inde-
pendent of the strut inclination (a/d) and the strain in the Fig. 3—Flowchart to solve capacity of FRP-reinforced deep
tension ties crossing a strut. However, Eq. (17), which was beam using STMs. (Note: V = P/2.)
empirically derived for steel-reinforced deep beams, does
account for the reduced strut capacity if the strut is located
such that the compression forces can spread (that is, a with the centroid of the FRP reinforcement. The height of the
bottle-shaped strut), as is the case with the diagonal struts bottom node ha is defined by the location of the reinforce-
in Fig. 1(b). ment and is taken as twice the distance from the bottom of
Similar to the CSA A23.3-045 STM for steel-reinforced the beam to the centroid of the reinforcement. The heights
concrete members, the tie strength in the ACI 318-086 STM of the top nodes/compression zone da and the corresponding
is evaluated based on the yield strength of the steel. In inclination angle q of the diagonal struts were defined during
adapting these provisions to FRP-reinforced members, it is the calculations. In all instances, the shear capacity was
necessary to consider the linear elastic stress-strain response determined iteratively by incrementing the total applied load
of the FRP when determining the capacity of the STM. P. A flowchart defining the calculation procedure is shown
ACI 318-086 requires that the stresses in the nodal zones in Fig. 3 for the case of the CSA A23.3-045 model and is
not exceed 0.85bnfc′, where bn is 1.0 for CCC nodes, briefly described herein.
0.80 for CCT nodes, and 0.60 for CTT nodes. Contrary to the For a given applied load P, the force in each STM element
CSA A23.3-045 STM, the ACI 318-086 STM does not require (that is, the struts, ties, and nodes) was compared to the
distributed reinforcement if the appropriate bs factor is used limiting capacities. This required that the truss geometry be
in Eq. (17). If bs = 0.75 is used in Eq. (17), the axis of the established. For the selected P, the moment at midspan was
strut must be crossed by reinforcement proportioned to resist calculated. By assuming that the stress in the top horizontal
the tensile forces caused by the spreading of the compres- strut was at its maximum permitted value of 0.85fc′, the top
sion forces. Alternatively, if fc′ does not exceed 41 MPa strut height da and the force in the FRP tie Ft were calculated
(6000 psi), the reinforcement ratio of the strut needs to be simultaneously. With the dimension of the top strut known,
at least 0.3%, as calculated with ACI 318-08,6 Clause A3.3. the diagonal strut angle q was calculated based on the
However, the specimens under consideration in this paper centerline geometry of the strut. The forces in the diagonal
contained no distributed web reinforcement, which was struts were determined from equilibrium. The FRP strain
accounted for by using bs = 1.0 or 0.6, depending on the at midspan efrp was calculated by dividing the previously
strut type. determined FRP tensile force by the axial stiffness EfrpAfrp.
Equation (16) was used to calculate the appropriate trans-
Applying STMs to FRP-reinforced members verse tensile strain across the diagonal strut. Subsequently,
The first step in strut-and-tie modeling involves the selec- the limiting compressive strut stress fc was calculated using
tion of a suitable truss model to transfer the loads to the Eq. (15) at the bottom of the diagonal strut and compared to
supports. In design, the selection of an appropriate model the applied strut stress at this location. A similar calculation
typically involves some iteration or trial and error of the at the top of the diagonal strut was completed. These steps
truss geometry to ensure that the stresses in all struts, ties, were repeated as the load was incremented, with the strain
and nodal regions are within the permitted stress limits, as in the FRP being recalculated each time due to its linear
defined in the following sections. For use in analysis, an elastic response. The maximum predicted load was defined
iterative solution technique is typically required to solve the by reaching one of the stress limits. The maximum predicted
maximum capacity of a given truss model. capacity (Vcsa1 or Vcsa2) for all specimens examined in this
A simple truss model with direct diagonal compression study was limited by the diagonal strut capacity near the
struts between the load and support locations was adopted lower support and not by tie failure or nodal failure.
for predicting the capacities of the members examined in this For capacity predictions using the ACI 318-086 STM
study (Fig. 1(b)). The centerline location of the tie coincided approach, a similar iterative approach was required to estab-

ACI Structural Journal/July-August 2013 599


Table 1—Capacity predictions of specimens in current study using STMs and sectional models
STMs Sectional models

CAN/
CSA A23.3-045 ACI 318- ACI ISIS CSA CSA Hoult
STM 086 STM 440.1R-0613 Canada12 S806-0211 S6-064,18 et al.21
Spec- h, d, bw, fc′, Efrp*, Vexp,
imen mm mm mm MPa a/d r, % GPa kN Vexp/Vcsa1 Vexp/Vcsa2 Vexp/Vaci Vexp/Vc Vexp/Vc Vexp/Vc Vexp/Vc Vexp/Vc
A1N 306 257 310 40.2 1.07 1.49 41.1 407 1.39 1.09 1.00 11.1 8.9 5.0 6.1 5.2
A2N 310 261 310 45.4 1.44 1.47 41.1 236 1.15 0.90 0.66 6.1 4.8 2.8 3.4 3.0
A3N 310 261 310 41.3 2.02 1.47 41.1 122 1.11 0.88 0.55 3.2 2.6 1.5 1.9 1.7
A4H 310 261 310 64.6 2.02 1.47 41.1 96 0.67 0.52 0.28 2.3 1.6 1.0 1.4 1.2
B1N 608 503 300 40.5 1.08 1.70 37.9 637 1.10 0.87 0.80 8.9 8.8 7.7 5.5 4.8
B2N 606 501 300 39.9 1.48 1.71 37.9 400 1.11 0.87 0.68 5.6 5.6 4.9 3.5 3.2
B3N 607 502 300 41.2 2.07 1.71 37.9 216 1.04 0.82 0.52 3.0 3.0 2.6 2.0 1.8
B4N 606 496 300 40.7 1.48 2.13 41.1 415 1.01 0.80 0.69 5.1 5.5 5.0 3.2 3.0
B5H 607 497 300 66.4 1.48 2.12 41.1 531 0.94 0.73 0.54 5.7 5.5 5.0 3.9 3.6
B6H 610 505 300 68.5 2.06 1.70 37.9 188 0.65 0.50 0.27 2.3 2.0 1.7 1.7 1.5
C1N 1003 889 301 51.6 1.10 1.58 42.3 1135 1.11 0.87 0.68 8.2 9.3 7.4 5.8 5.3
C2N 1005 891 304 50.7 1.49 1.56 42.3 662 1.04 0.82 0.53 4.8 5.4 4.3 3.4 3.2
Mean—all specimens 1.03 0.81 0.60 5.53 5.25 4.07 3.47 3.12
Standard deviation 0.20 0.16 0.20 2.75 2.68 2.19 1.63 1.41
Coefficient of variation (COV) 0.20 0.20 0.34 0.50 0.51 0.54 0.47 0.45
Mean—specimens with fc′≈ 40 MPa 1.12 0.88 0.68 6.23 5.98 4.56 3.86 3.46
Standard deviation 0.11 0.09 0.15 2.68 2.51 2.08 1.58 1.35
COV 0.10 0.10 0.22 0.43 0.42 0.46 0.41 0.39
*
Glass fiber-reinforced polymer (GFRP) reinforcement.
Notes: Lb = 100 mm for A specimens, 200 mm for B specimens, and 330 mm for C specimens; 1 mm = 0.0394 in.; 1 MPa = 145 psi.

lish the member capacity due to the linear elastic response the large-scale and moderate-scale specimens, respectively.
of the FRP. At each iteration, the tie force Ft must be The critical section used in the analysis was typically taken
compared to the limiting capacity. However, unlike the CSA at a distance of dv (d for the CAN/CSA S806-0211 model)
A23.3-045 models, the tie strain does not directly influ- from the point load toward the support if a/d > 2.0. For spec-
ence the capacity of the adjacent diagonal strut. The cross- imens with a/d ≤ 2.0, the critical section was assumed to be
sectional dimension of the strut is influenced by the diagonal at the midpoint of the shear span. The relationships between
strut angle q, which varies at each iteration. For all specimens
the experimental-to-predicted shear capacities and a/d are
examined in this study, the maximum predicted capacity Vaci
shown in Fig. 4 for the five sectional models.
was limited by the diagonal strut capacity at the upper end
of the diagonal strut and not by tie failure or nodal failure. In general, all of the sectional shear models were inca-
pable of properly modeling the shear capacity for the
SHEAR CAPACITY PREDICTIONS members with a/d < 2.4. The models all became increasingly
A database of 36 test results was assembled for concrete conservative as the a/d decreased due to their neglect of arch
beams with a/d < 2.4 and containing internal longitudinal FRP action. All models were able to account for decreases in the
reinforcement but no transverse reinforcement. This database shear stress at failure as the member depth increased, but
included the 12 large-scale tests reported in the companion the CAN/CSA S806-0211 model—with different equations
paper14 (Table 1) and 24 moderate-scale tests reported by depending on the magnitude of d—gave higher Vexp/Vc ratios
El-Sayed16 and Nehdi et al.17 (Table 2). Specimens contained as d increased (Fig. 4(c)). Note that Fig. 4(c) only shows
either glass FRP14,16,17 or carbon FRP.16,17 Concrete strengths the specimens from the companion paper14 for ease of size
ranged between 34 and 69 MPa (4.9 and 10 ksi). As-built comparison. The ACI 440.1R-0613 shear model had the
dimensions and measured material properties were used in worst overall predictions with Vexp/Vc ratios of up to 12.8 and
all calculations. The shear capacities of each specimen were
an average value of 5.69 for the 36 members considered. The
determined using the sectional and STM models.
Hoult et al.21 model gave the best overall predictions with
CAPACITY PREDICTIONS USING SECTIONAL a mean Vexp/Vc of 2.97 for the 36 data points. Nevertheless,
SHEAR MODELS this model still demonstrates that substantial improvement in
The experimental-to-predicted shear capacity ratios for the model accuracy is required to properly consider the behavior
five sectional shear models are provided in Tables 1 and 2 for of members with a small a/d.

600 ACI Structural Journal/July-August 2013


Table 2—Capacity predictions of specimens from previous research programs using STMs and
sectional models
STMs Sectional models

CAN/
CSA A23.3- ACI 318- ACI 440.1R ISIS CSA CSA Hoult et
045 STM 086 STM -0613 Canada12 S806-0211 S6-064,18 al.21
Refer- d, fc′, FRP Efrp, Vexp, Vexp/ Vexp/
ence Specimen mm MPa a/d r, % type* GPa kN Vcsa1 Vcsa2 Vexp/Vaci Vexp/Vc Vexp/Vc Vexp/Vc Vexp/Vc Vexp/Vc
II-C-0.7/1.6 326 40 1.69 0.78 C 134 179.5 0.98 0.79 0.82 3.76 2.17 3.55 2.14 2.05
II-C-1.2/1.3 326 40 1.3 1.24 C 134 372 1.16 0.95 1.30 6.38 4.50 7.37 3.75 3.61
II-C-1.2/1.6 326 40 1.69 1.24 C 134 195 0.92 0.76 0.89 3.35 2.36 3.86 1.97 1.95
II-C-1.7/1.6 326 40 1.69 1.71 C 134 233.5 1.00 0.84 1.07 3.52 2.82 4.62 2.12 2.14

16 II-G-0.8/1.6 326 40 1.69 0.78 G 42 164.5 1.37 1.05 0.75 5.81 3.55 3.26 3.13 2.72
II-G-1.2/0.9 326 40 0.92 1.24 G 42 450.5 1.33 1.04 1.22 12.83 9.73 8.92 7.05 5.90
II-G-1.2/1.3 326 40 1.3 1.24 G 42 269 1.25 0.97 0.94 7.66 5.81 5.33 4.21 3.67
II-G-1.2/1.6 326 40 1.69 1.24 G 42 175 1.22 0.95 0.80 4.99 3.78 3.47 2.74 2.48
II-G-1.7/1.6 326 40 1.69 1.71 G 42 196 1.21 0.96 0.89 4.84 4.23 3.88 2.70 2.51
CF-B-1 150 34.7 1.55 1.13 C 134 92.6 1.31 1.08 1.49 6.26 4.27 3.49 3.39 3.12
CF-B-1.5 150 38.9 1.83 1.13 C 134 77.5 1.32 1.08 1.34 5.06 3.37 2.76 2.74 2.56
CF-B-2 150 40.6 2.33 1.13 C 134 45.8 1.16 0.96 1.04 2.95 1.95 1.62 1.71 1.59
CF-pl-1.5 150 37.4 1.83 1.69 C 134 81.2 1.26 1.06 1.46 4.53 3.61 2.95 2.53 2.44
CF-pl-2.0 150 39.6 1.83 2.26 C 134 92.8 1.28 1.10 1.57 4.55 4.00 3.28 2.59 2.53
CF-d-250 250 41.7 1.41 1.35 C 134 149.1 1.21 1.01 1.75 5.30 3.76 3.08 3.04 2.95
CF-d-350 350 37.6 1.36 1.21 C 134 234.1 1.77 1.49 2.92 6.34 4.61 7.55 3.84 3.72

17 CF-fc′-60 150 63.1 1.83 1.13 C 134 113.5 1.40 1.12 1.21 6.45 3.87 3.18 3.50 3.22
F-B-1 150 35.5 1.55 1.29 G 40.8 67.8 1.36 1.06 1.06 7.21 5.60 3.24 3.70 3.12
F-B-1.5 150 48 1.83 1.29 G 40.8 64.3 1.39 1.08 0.90 6.30 4.56 3.78 3.22 2.75
F-B-2 150 48 2.33 1.29 G 40.8 28.9 0.96 0.74 0.55 2.83 2.05 1.38 1.57 1.33
F-pl-1.5 150 48 1.83 1.72 G 40.8 73.6 1.43 1.12 1.03 6.34 5.22 2.90 3.27 2.86
F-pl-2.0 150 35.2 1.83 2.31 G 40.8 61.0 1.30 1.04 1.17 5.04 5.04 2.41 2.65 2.40
F-d-250 250 42 1.41 1.39 G 40.8 121.6 1.46 1.14 1.42 7.19 5.53 3.22 3.83 3.38
F-d-350 350 48 1.36 1.25 G 40.8 211.3 1.99 1.55 2.06 9.03 6.69 6.04 4.99 4.34
Mean 1.29 1.04 1.24 5.77 4.29 3.92 3.18 2.89
Standard deviation 0.24 0.19 0.50 2.16 1.69 1.86 1.16 0.95
Coefficient of variation (COV) 0.18 0.18 0.40 0.37 0.39 0.47 0.37 0.33
*
C is carbon; G is glass.

Specimens from El-Sayed16 had bw = 250 mm and Lb = 100 mm.

Specimens from Nehdi et al.17 had bw = 150 mm and Lb = 50 mm.
Notes: 1 mm = 0.0394 in.; 1 MPa = 145 psi.

CAPACITY PREDICTIONS USING STRUT-AND- enced by the concrete strength. For the subset of specimens
TIE MODELS with normal-strength concrete (fc′ ≈ 40 MPa [5800 psi]),
Capacity predictions relative to the test results for the mean Vexp/Vcsa1 was 1.12 with a 10% COV and all speci-
the 12 large-scale specimens reported in the companion mens had Vexp/Vcsa1 > 1.0. The specimens with high-strength
paper14 are provided in Fig. 5. Results for the full data set concrete (fc′ ≈ 66 MPa [9570 psi]) had Vexp/Vcsa1 ratios of less
of 36 members are provided in Fig. 6. than 1.0, especially Specimens A4H and B6H, which had
a/d = 2.1. These two specimens exhibited crack patterns that
CSA Approach 1 prevented full direct compression struts from forming.14,15
The mean test-to-predicted capacity ratio Vexp/Vcsa1 of the Specimen B5H with a/d = 1.48 had Vexp/Vcsa1 = 0.94, but
12 specimens reported in Table 1 using CSA Approach 1 (that unlike Specimens A4H and B6H, this beam developed a
is, es = efrp) was 1.03 with a coefficient of variation (COV) complete arching mechanism.14,15
of 20%. It can be observed in Fig. 5(a) that the prediction Predictions for all 36 specimens using this model are
quality was similar for the range of a/d values studied and shown in Fig. 6(a). It can be observed that the prediction
for the member depths, but the average statistics were influ- quality is similar regardless of FRP type (glass or carbon)

ACI Structural Journal/July-August 2013 601


Fig. 4—Shear capacity predictions using: (a) ACI 440.1R-0613; (b) ISIS Design Manual 312; (c) CAN/CSA S806-0211; (d) CSA S6-064,18;
and (e) Hoult et al.21 sectional shear models. Figure 4(c) only shows specimens in current study. (Note: 1 mm = 0.0394 in.; 1 MPa
= 145 psi.)

Fig. 5—Comparison of experimental shear capacity to predicted capacity from CSA A23.3-045 STM using (a) Approach 1 and
(b) Approach 2; and (c) ACI 318-086 STM for specimens reported in Table 1. (Note: 1 MPa = 145 psi.)
or member size. The average Vexp/Vcsa1 ratio was 1.20 with studied. While the average statistics were influenced by the
a 21% COV for the 36-member data set and 1.24 and 18%, concrete strength, all specimens except for Specimen A1N
respectively, for the subset of 32 normal-strength concrete had Vexp/Vcsa2 ≤ 1.0. The subset of specimens with normal-
(fc′ ≈ 40 MPa [5800 psi]) members. strength concrete (fc′ ≈ 40 MPa [5800 psi]) had an average
Vexp/Vcsa2 of 0.88 and a COV of 10%.
CSA Approach 2
The mean test-to-predicted capacity ratio Vexp/Vcsa2 of the Predictions for all 36 specimens using this model are
12 specimens reported in Table 1 using CSA Approach 2 (that shown in Fig. 6(b). It can be observed that the prediction
is, es = 0.5efrp) was 0.81 with a 20% COV. Similar to CSA quality is similar regardless of FRP type (glass or carbon)
Approach 1 and as shown in Fig. 5(b), the prediction quality or member size. The average Vexp/Vcsa2 ratio was 0.96 with
was similar for the range of a/d values and member depths a 22% COV for the 36-member data set and 0.99 and 18%,

602 ACI Structural Journal/July-August 2013


Fig. 6—Comparison of experimental shear capacity to predicted capacity for all specimens using CSA A23.3-045 STM using
(a) Approach 1 and (b) Approach 2; and (c) ACI 318-086 STM. (Note: 1 MPa = 145 psi.)

respectively, for the subset of 32 normal-strength concrete as a/d changes, indicating that the ACI 318-086 STM was
(fc′ ≈ 40 MPa [5800 psi]) members. incapable of accurately modeling FRP-reinforced concrete
members for all practical a/d.
ACI 318-086 predictions In addition, CSA Approaches 1 and 2 were able to account
The mean test-to-predicted capacity ratio Vexp/Vaci of for the different FRP types, while the ACI 318-086 STM
the 12 specimens reported in Table 1 using the ACI 318-086 was not able to provide similar prediction quality as the
STM was 0.60 with a 34% COV. All specimens had reinforcement stiffness changed. The ACI 318-086 strut
Vexp/Vaci ≤ 1.0. It is shown in Fig. 5(c) that the prediction capacity equation (Eq. (17)) was empirically derived for
quality decreased as the a/d increased, indicating poor steel-reinforced concrete deep beams and does not directly
modeling correlation with the a/d. The average statistics account for the strain in the tie crossing a strut (that is, the
were also influenced by the concrete strength. For the subset stiffness of the reinforcement). The influence of the reinforce-
of specimens with normal-strength concrete (fc′ ≈ 40 MPa ment axial stiffness EfrpAfrp normalized by the a/d on the
[5800 psi]), the mean Vexp/Vaci was 0.68 with a 22% COV. capacity predictions is shown in Fig. 7 for all specimens.
The three specimens with high-strength concrete (fc′ ≈ CSA Approaches 1 and 2 were able to account for the influ-
66 MPa [9570 psi]) had a mean Vexp/Vaci of 0.36. ence of the reinforcement axial stiffness (Fig. 7(a) and (b)),
Predictions for all 36 specimens are shown in Fig. 6(c). while the ACI 318-086 STM (Fig. 7(c)) showed poor corre-
It can be observed that the prediction quality is poor, with lation between the capacity prediction and axial stiffness of
a large amount of scatter, and is dependent on the FRP type the reinforcement.
(glass or carbon). The average Vexp/Vaci ratio was 1.02 with All models gave poor predictions for the specimens with
a 51% COV for the 36-member data set and 1.08 and 47%, high-strength concrete. As noted previously and in the
respectively, for the subset of 32 normal-strength concrete companion paper,14 some of these specimens had crack
(fc′ ≈ 40 MPa [5800 psi]) members. The average Vexp/Vaci patterns that prevented full direct compression struts from
forming. In addition, research on steel-reinforced concrete
ratio was 1.40 with a 39% COV for the carbon fiber-reinforced
deep beams has established that the ACI 318-086 STM
polymer (CFRP) reinforced members and 0.83 and 46%,
and the AASHTO LRFD7 STM, which is similar to the
respectively, for the subset of glass fiber-reinforced polymer
CSA A23.3-045 STM, give unconservative predictions when
(GFRP) reinforced members.
fc′ exceeds 41 and 48 MPa (6000 and 7000 psi), respectively.26
Comparison of strut-and-tie models RECOMMENDED ANALYSIS APPROACH
Based on Tables 1 and 2 and Fig. 6, the CSA A23.3-045 STM FOR DEEP BEAMS
where the diagonal strut capacity is calculated using the full As discussed previously, the sectional shear models became
midspan tie strain (that is, CSA Approach 1 with es = efrp), increasingly conservative as the a/d decreased and were inca-
gave the best predictions for all specimens analyzed. This pable of modeling the shear capacity of members with a/d <
model had a mean Vexp/Vcsa1 ratio of 1.20 with a COV of 2.4. In general, the CSA STM models were able to effectively
21% for the 36-member data set. Only 17% of the specimens account for the a/d, member height, and reinforcement stiff-
had Vexp/Vcsa1 < 1.0. CSA Approach 2 with es = 0.5efrp had ness, with CSA Approach 1 providing the best prediction
a mean Vexp/Vcsa2 ratio of 0.96 with a COV of 22%. While quality. The ACI 318-086 STM gave poor predictions and was
this gave an average value closer to unity than the CSA incapable of producing consistent capacity predictions as the
Approach 1 results, the majority of the specimens (56%) a/d or reinforcement stiffness changed.
had Vexp/Vcsa2 < 1.0 with 28% having Vexp/Vcsa2 < 0.85. The Because the intent of this study was to provide a method
ACI 318-086 STM had a mean Vexp/Vaci ratio of 1.02 with for predicting the shear capacity of full-scale members, the
a COV of 51%. Using the ACI 318-086 model resulted in experimental data from the large-scale specimens should be
predictions where 53% of the specimens had Vexp/Vaci < given greater consideration. Therefore, this study considered
1.0 and 42% had Vexp/Vaci < 0.85. Further, Fig. 5(c) and members with h > 300 mm (11.8 in.) and up to 1000 mm
6(c) demonstrate the variability in the prediction quality (39.4 in.) to be more representative of the beam sizes used

ACI Structural Journal/July-August 2013 603


Fig. 7—Influence of axial stiffness of reinforcement on experimental shear capacity to predicted capacity for all specimens
using CSA A23.3-045 STM using (a) Approach 1 and (b) Approach 2; and (c) ACI 318-086 STM. (Note: 1 MPa = 145 psi; 1 N
= 0.225 lb.)

in the industry. The best STM predictions for the large- an alternative method to consider the arch action. The
scale specimens tested in this study were obtained by using ACI 318-086 STM provisions overpredicted the shear
CSA Approach 1 (es = efrp). capacity of the specimens by up to 250% because they do not
account for the influence on capacity related to the reinforce-
CAUTIONS AND LIMITATIONS ment strain magnitudes.
Unlike steel reinforcement, which exhibits a yield plateau The CSA A23.3-045 STM was able to directly account
and substantial elongation prior to rupture, FRP reinforcing for the influences of the reinforcement strain and the a/d
bars exhibit a linear stress-strain response prior to brittle (through the strut angle)—two parameters that have a large
failure. Further, most design codes for the use of internal impact on the shear capacity of deep members. To properly
FRP reinforcement in concrete structures require that apply the CSA A23.3-045 STM, the elastic stress-strain
the stress in the FRP be kept below a relatively low frac- response should be directly computed using an equilibrium-
tion of the guaranteed tensile strength due to concerns based approach.
regarding long-term material performance and/or degrada- Two methods for calculating the CSA A23.3-045 strut
tion. Thus, the FRP will behave elastically, even at failure capacity were discussed in this paper. The strut capacity can
of the member, and the dominant mechanism to achieve a either be calculated using efrp or 0.5efrp. Using efrp in calcu-
ductile structural response must be through the deformation lating the strut capacity gave experimental-to-predicted shear
capacity of the concrete. Because the recommended STM capacity (Vexp/Vcsa1) ratios greater than 1.0 for the majority of
approach for deep concrete members reinforced with FRP the specimens tested. Using 0.5efrp resulted in the majority
evaluates the performance at the ultimate limit state condi- of the specimens having Vexp/Vcsa2 < 1.0.
tion, this approach gives rise to several cautions and limita- The CSA A23.3-045 STM, where the strut capacity is
tions. The linear elastic nature of the FRP must be taken into calculated by assuming that the strain in the FRP passing
account when using the strut-and-tie modeling technique. through the strut is the same as the midspan FRP strain (efrp),
In steel-reinforced concrete members, the steel provides is the recommended model for calculating the shear capacity
the required plastic deformation to allow redistribution of deep beams reinforced with FRP and having no distrib-
of internal stresses into the adopted truss model. In FRP- uted reinforcement. Using sectional models to determine the
reinforced concrete members, the concrete can only provide shear capacity of FRP-reinforced deep beams is a safe alter-
limited ductility, which may not be enough in truss models native but can have adverse effects on the member economy
more complex than those discussed in this study. The simple as the a/d decreases.
truss model used in this study (Fig. 1(b)) was determinant
and closely followed the internal forces, as indicated by ACKNOWLEDGMENTS
the theory of elasticity. Limited ductility was required in Funding for this project was provided by the Natural Sciences and Engi-
this instance, which may not be the case in more complex neering Research Council of Canada (NSERC), Alberta Ingenuity, and the
STMs. Additional research is required to ensure that the University of Alberta.
STM can be adopted for more complex members containing
FRP reinforcement. NOTATION
Acs = cross-sectional area at one end of strut, taken perpendicular to
strut axis
RECOMMENDATIONS AND CONCLUSIONS Afrp = area of FRP reinforcing bars on flexural tension side of member
All of the sectional shear models analyzed in this study At = cross-sectional area of tension tie reinforcement
gave poor predictions of capacity for specimens having an a = dimension of shear span measured between centerlines of loading
a/d of less than 2.5. The sectional shear models increas- and reaction plates
ingly underpredicted the shear capacity as the a/d decreased. ag = nominal maximum size of coarse aggregate
bw = member width
These models are incapable of accounting for arch action. c = distance from extreme compression fiber to neutral axis
As is the case with steel-reinforced members, the d = effective depth measured from extreme compression fiber to
shear capacity of deep members must be computed using centroid of longitudinal tension reinforcement

604 ACI Structural Journal/July-August 2013


da = top strut height 9. Rogowsky, D. M., and MacGregor, J. G., “Design of Reinforced
dv = effective shear depth Concrete Deep Beams,” Concrete International, V. 8, No. 8, Aug. 1986,
Efrp = modulus of elasticity of FRP reinforcement pp. 49-58.
Es = modulus of elasticity of steel reinforcement 10. Schlaich, J.; Schäfer, K.; and Jennewein, M., “Towards a Consis-
Fs = strut capacity tent Design of Reinforced Concrete Structures,” PCI Journal, V. 32, No. 3,
Ft = tie reinforcement axial capacity 1987, pp. 74-150.
fc = limiting compressive stress in strut 11. CAN/CSA S806-02, “Design and Construction of Building Compo-
fc′ = compressive strength of concrete nents with Fibre-Reinforced Polymers,” Canadian Standards Association,
fcr = cracking strength of concrete Mississauga, ON, Canada, 2002, 177 pp.
ft = stress in reinforcement 12. ISIS Canada Research Network, “Reinforcing Concrete Structures
h = overall member height with Fibre Reinforced Polymers—Design Manual 3, Version 2,” ISIS
ha = height of bottom node Canada, Winnipeg, MB, Canada, 2007, 129 pp.
k = ratio of depth of neutral axis to reinforcement depth 13. ACI Committee 440, “Guide for the Design and Construction of
Lb = bearing plate dimension in direction of span Structural Concrete Reinforced with FRP Bars (ACI 440.1R-06),” Amer-
M = moment ican Concrete Institute, Farmington Hills, MI, 2006, 44 pp.
nf = modular ratio of FRP to concrete 14. Andermatt, M. F., and Lubell, A. S., “Behavior of Concrete Deep
P = total load applied on member Beams Reinforced with Internal Fiber-Reinforced Polymer—Experimental
sze = crack spacing parameter that allows for influence of aggregate size Study,” ACI Structural Journal, V. 110, No. 4, July-Aug. 2013, pp. 585-594.
V = shear force 15. Andermatt, M. F., “Concrete Deep Beams Reinforced with Internal
Vc = shear resistance attributed to concrete FRP,” MSc thesis, University of Alberta, Edmonton, AB, Canada, 2010,
Vexp = experimental shear resistance of reinforced concrete member 266 pp.
bn = factor to account for effect of anchorage of ties on effective 16. El-Sayed, A. K., “Concrete Contribution to the Shear Resistance of
compressive strength of nodal zone FRP-Reinforced Concrete Beams,” PhD dissertation, Université de Sher-
bs = factor to account for effect of cracking and confining reinforce- brooke, Sherbrooke, QC, Canada, 2006, 252 pp.
ment on effective compressive strength of concrete in strut 17. Nehdi, M.; Omeman, Z.; and El-Chabib, H., “Optimal Efficiency
e1 = principal tensile strain in cracked concrete due to loads Factor in Strut-and-Tie Model for FRP-Reinforced Concrete Short Beams
efrp = tensile strain in tie reinforcement at midspan with (1.5 < a/d < 2.5),” Materials and Structures, V. 41, No. 10, 2008,
efrpu = rupture strain of FRP pp. 1713-1727.
es = tensile strain in tie reinforcement at center of strut due to loads 18. CSA S6S1-10, “Supplement #1 to CAN/CSA S6-06 Canadian
ex = longitudinal strain at middepth of member due to loads Highway Bridge Design Code,” Canadian Standards Association, Missis-
q = angle of inclination of diagonal compressive stresses to axis of sauga, ON, Canada, 2010, 278 pp.
longitudinal reinforcement 19. Vecchio, F. J., and Collins, M. P., “The Modified Compression
r = longitudinal reinforcement ratio Afrp/bdw Field Theory for Reinforced Concrete Elements Subjected to Shear,” ACI
Journal, V. 83, No. 2, Mar.-Apr. 1986, pp. 219-231.
20. Bentz, E. C., and Collins, M. P., “Development of the 2004 Cana-
REFERENCES dian Standards Association (CSA) A23.3 Shear Provisions for Reinforced
1. Joint ACI-ASCE Committee 445, “Recent Approaches to Shear Concrete,” Canadian Journal of Civil Engineering, V. 33, No. 5, 2006,
Design of Structural Concrete (ACI 445R-99) (Reapproved 2009),” Amer- pp. 508-520.
ican Concrete Institute, Farmington Hills, MI, 1999, 55 pp. 21. Hoult, N. A.; Sherwood, E. G.; Bentz, E. C.; and Collins, M. P.,
2. Zsutty, T., “Beam Shear Strength Prediction by Analysis of Existing “Does the Use of FRP Reinforcement Change the One-Way Shear Behavior
Data,” ACI Journal, V. 65, No. 11, Nov. 1968, pp. 943-951. of Reinforced Concrete Slabs?” Journal of Composites for Construction,
3. Kani, M. W.; Huggins, M. W.; and Wittkopp, R. R., Kani on Shear in ASCE, V. 12, No. 2, 2008, pp. 125-133.
Reinforced Concrete, University of Toronto, Toronto, ON, Canada, 1979, 22. CSA A23.3-94, “Design of Concrete Structures,” Canadian Stan-
225 pp. dards Association, Mississauga, ON, Canada, 1994, 199 pp.
4. CAN/CSA S6-06, “Canadian Highway Bridge Design Code,” Cana- 23. Tureyen, A. K., and Frosch, R. J., “Concrete Shear Strength: Another
dian Standards Association, Mississauga, ON, Canada, 2006, 788 pp. Perspective,” ACI Structural Journal, V. 100, No. 5, Sept.-Oct. 2003,
5. CSA A23.3-04, “Design of Concrete Structures,” Canadian Standards pp. 609-615.
Association, Mississauga, ON, Canada, 2004, 232 pp. 24. Wight, J. K., and MacGregor, J. G., Reinforced Concrete: Mechanics
6. ACI Committee 318, “Building Code Requirements for Structural and Design, fifth edition, Pearson Prentice Hall, Upper Saddle River, NJ,
Concrete (ACI 318-08) and Commentary,” American Concrete Institute, 2009, 1126 pp.
Farmington Hills, MI, 2008, 473 pp. 25. Collins, M. P., and Mitchell, D., Prestressed Concrete Structures,
7. AASHTO, “LRFD Bridge Design Specifications, SI Units, 4th Response Publications, Toronto, ON, Canada, 1997, 720 pp.
Edition,” American Association of State Highway and Transportation Offi- 26. Bahen, N., and Sanders, D. H., “Investigation of Strut Strength Using
cials, Washington, DC, 2007, 1518 pp. a Deep-Beam Database,” Thomas T.C. Hsu Symposium: Shear and Torsion
8. Marti, P., “Basic Tools of Reinforced Concrete Beam Design,” ACI in Concrete Structures, SP-265, A. Belarbi, Y. L. Mo, and A. Ayoub, eds.,
Journal, V. 82, No. 1, Jan.-Feb. 1985, pp. 46-56. American Concrete Institute, Farmington Hills, MI, 2009, pp. 385-404.

ACI Structural Journal/July-August 2013 605


One Click... One Entire Journal

Introducing
a new feature on ACI’s website!
VOL. 108, NO. 6 VOL. 108, NO. 6
NOVEMBER-DECEMBER 2011 NOVEMBER-DECEMBER 2011

ACI ACI
MATERIALS STRUCTURAL J O U R N A L
J O U R N A L

A JOURNAL OF THE AMERICAN CONCRETE INSTITUTE A JOURNAL OF THE AMERICAN CONCRETE INSTITUTE

215 Long-Term Behavior of Cracked Steel Fiber-Reinforced 215 Long-Term Behavior of Cracked Steel Fiber-Reinforced
Concrete Beams under Sustained Loading Concrete Beams under Sustained Loading

Now, with one click of the mouse, subscribers of the


ACI Structural and Materials Journals digital editions can
download an Adobe Acrobat PDF of an unabridged
issue at www.concrete.org/pubs/electronicjournals/
esjindexnew.asp and ww.concrete.org/pubs/
electronicjournals/emjindexnew.asp.

You might also like