2010 Structures Congress 2010 ASCE

Use of Steel Fiber Reinforced Concrete for Enhanced

Performance of Deep Beams with Large Openings
D. R. Sahoo1 and S.-H. Chao2
1

Postdoctoral Fellow, Department of Civil Engineering, University of Texas at

Arlington, TX 76019; PH (817) 272-9163; email: [email protected]
2
Assistant Professor, Department of Civil Engineering, University of Texas at
Arlington, TX 76019; PH (817) 272-2550; FAX (817) 272-2630; email:
[email protected]

ABSTRACT
Reinforced concrete deep beams are used as primary load distribution elements
in various civil engineering structures. Large openings often interrupt the load
transfer by concrete struts in these beams and cause a sharp decrease in strength and
serviceability. Although the strength evaluation and reinforcement details around the
openings are essential considerations, the ACI Building Code does not provide
explicit guidance for designing these elements with openings. Strut-and-tie models
are commonly used for strength evaluation and design of deep beams with openings.
However, reinforcement detailing based on these models can be very complex and the
failure of deep beams may be due to localized damages that could not be predicted by
the strut-and-tie models. In this study, an experimental investigation was conducted
on two concrete deep beam specimens with large single opening, namely, reinforced
concrete (RC) and steel fiber reinforced concrete (SFRC), to evaluate their
performance under monotonically increased load. The reinforcement detailing in the
SFRC specimen was considerably reduced since the steel reinforcement bars were
only used for the tensile longitudinal reinforcements and the boundary elements. Both
test specimens had significantly higher strength than the designed load computed
based on one assumed strut-and-tie model.
INTRODUCTION
Deep beam can be defined as a beam in which either clear span is equal to or
less than four times the overall member depth or concentrated loads are within a
distance equal to or less than two times the depth from the face of support (ACI
Committee 318, 2008). Very often, openings are located through these deep beams to
provide access for conduits or mechanical chases. These openings disrupt the stress
flow from the loading zones to the supports. Code-specified empirical formulas used
to design these members do not explicitly address the design of D-regions with

1981

2010 Structures Congress 2010 ASCE

openings. Instead, various strut-and-tie models have been used for designing these
discontinuous regions.
Past studies (e.g. Maxwell and Breen, 2000; Chen et al., 2002; Kuchma and
Park, 2007; Tan and Zhang, 2007; Brea and Morrison, 2007) showed that the strutand-tie models provide consistent and conservative results in terms of ultimate
strength of deep beams with openings. Further, large differences can exist between
the calculated forces from strut-and-tie model and actual instrumented experimental
specimens. A poorly selected and detailed strut-and-tie model can lead to severe
cracking and damages under service loads (Kuchma et al., 2008). From the
construction point of view, the primary difficulty associated in the design based on
strut-and-tie models is the problem of anchorage and congestion due to large amount
of reinforcement bars in the member.
Steel Fiber Reinforced Concrete (SFRC) has gained increased popularity in
construction industries in the recent years. Reinforcing plain concrete with steel fibers
has been used to reduce conventional steel reinforcement in structural members such
as slabs (ACI Committee 544, 1996). SFRC members can exhibit enhanced shear
strength, more ductile behavior and reduced crack widths (Dupont et al., 2003). In
addition, SFRC offers a multidirectional reinforcement, simple detailing without
congestion, and higher post-cracking residual stress and ductility. Past studies
(Narayanan and Darwish, 1986; Mansur and Ong, 1991) have shown that including
discrete fibers in concrete enhance the strength and the deformation capacities of deep
beams in addition to better cracking control.
This paper presents the behavior of a RC deep beam with single opening under
the action of monotonically increased loading. The deep beam was designed using a
strut-and-tie model and the observed ultimate strength and failure modes were
compared with those predicted by the model. Further, a geometrically similar SFRC
deep beam was tested under the same loading conditions to compare its behavior, in
terms of ultimate strength and failure modes, with that of the RC deep beam. SFRC
deep beam consisted of conventional longitudinal steel bars as flexural tensile
reinforcement only at the bottom face. The effectiveness of reinforcement detailing at
the critical locations and the importance of steel fibers in concrete are recognized in
this study to enhance the performance of concrete deep beams with openings.
TEST SPECIMENS
The specimens tested in this experimental study had the same but 1/4 scale
geometry as the analytical model originally considered by Schlaich et al. (1987).
Earlier, the specimen with the same geometry and dimensions were also tested by
other researchers (Brea and Morrison, 2007). The overall dimension of the
specimens was 74 in. (1875 mm) long, 46 in. (1170 mm) deep, and 4.5 in. (112 mm)
thick. The specimens had a 15 in. (380 mm) square opening near the left-bottom
corner of the beams as shown in Figure 1(a). The position and size of the openings in
the specimens were selected to interfere with the direct load paths that could

1982

2010 Structures Congress 2010 ASCE

potentially form between the loading point and the supports (Brea and Morrison,
2007). Two types of test specimens (i.e., RC and SFRC) were investigated in the
present study. The RC specimen consisted of reinforcement bars designed and
detailed as per a strut-and-tie model as discussed later, whereas the SFRC specimen
consisted of reinforcement bars for longitudinal tensile reinforcements only at its
bottom. Because of complete absence of complex detailing of reinforcement bars, the
construction of the SFRC specimen was simple and fast.

(a)
(b)
Figure 1. (a) Details of test specimen and location of LVDTs (b) Strut-and-tie
model used in the present study
There is no unique strut-and-tie model to design a particular discontinuous
structure. Strut-and-tie model, originally proposed by Schlaich et al. (1987),
approximately follows the elastic (principal) stress distribution was used in the
present study for the test specimens (Brea and Morrison, 2007). As shown in
Figure 1(b), solid and dashed lines indicate ties and struts in the model, respectively.
The imposed load in the test specimen was transferred directly from the loading point
to the right support through a bottle-shaped strut, but the opening near the lower left
corner impairs the direct load transfer from the load point to the left support. This
model also provided a platform to compare the experimental results with those
available in the literature.
Test specimens were designed for an ultimate load-carrying capacity of 31.3
kips computed as per ACI Committee 318 (2008) procedure for strut-and-tie models.
The nominal compressive strength of concrete was assumed as 5000 psi, whereas the
tensile strength of reinforcement bars was considered as 60000 psi. A strengthreduction factor (s) equal to 0.75 was used for struts and nodes in the strut-and-tie
model. Reinforcing bars in the RC specimen was provided in two layers using No. 3
bars. Clear cover of concrete was provided as 1 in. to the edge of steel reinforcement.
Bars ending at nodes located near the beam edges were hooked at their ends to avoid
pullout. As shown in Figure 2(a), bottom longitudinal reinforcing bars were anchored
at the supports using standard 180-degree hooks. Prior experimental study (Brea and

1983

2010 Structures Congress 2010 ASCE

Morrison, 2007; Flores, 2009) showed that these specimens suffered severe cracking
and crushing of concrete near the supports. To avoid these premature local failures in
the RC specimen, a steel cage formed by four longitudinal reinforcement bas at the
corners and transverse stirrups at a spacing of 4 in. was used as boundary elements on
either side of the RC specimen. Further, no wire meshes were used in the RC
specimen.
Figure 2(b) shows the reinforcement detailing of the SFRC specimens. As stated
earlier, two No. 3 steel reinforcing bars were used only for bottom longitudinal tensile
reinforcement. Similar to the RC specimen, steel cages were used as boundary
elements near the left and right supports of the specimen. Further, two additional No.
6 bars were placed diagonally (normal to the line connecting the loading point to the
corner of opening) in two layers just above the opening so as to restrain the
propagation of cracks emanated from the corner of the opening.

(a)
(b)
Figure 2. Reinforcement detailing in test specimens (a) RC (b) SFRC
A concrete mixture with a nominal 28-day compressive strength equal to 5000
psi was used in all specimens. No chemical admixtures or super-plasticizers were
added in the concrete mix. The mix design was carried out for the target compressive
strength of concrete so that the optimum quantity and similar proportions of materials
could be used in both RC and SFRC specimens. The design mix proportion (by
weight) used for both specimens was 1.0 (cement): 0.5 (Fly ash): 1.7 (fine aggregate):
1.0 (coarse aggregate). Type-1 portland cement and Class-C fly ash were used in the
mix. The maximum size of coarse aggregate used in the concrete mix was limited to
0.5 in. A constant water-to-cementitious materials ratio of 0.4 was used for both
specimens. SFRC specimen consisted of end-hooked steel fibers (diameter = 0.03 in.;
length = 2.4 in.; aspect ratio = 80; tensile strength = 150 ksi) of volume equal to 1.5%
of the total volume of the specimen. The total weight of steel fibers was 89 pounds as
compared to the cement weight of 360 pounds used in the SFRC specimen.

1984

2010 Structures Congress 2010 ASCE

1985

Standard tests (ASTM C31 and C39; ACI 318 Sect. 5.6.2.4) were carried out to
evaluate the compressive strength of concrete (six 4 in. by 8 in. cylinders) and the
tensile strength of steel reinforcement bars. The nominal and actual properties of
concrete and steel reinforcement bars are compared in Table 1. The average values of
compressive strengths of concrete at the day of testing of the RC and the SFRC
specimens were 6.7 ksi and 6.3 ksi, respectively. The actual yield strength of No. 3
steel reinforcement bars was 81.2 ksi against their nominal value of 60 ksi. Moreover,
the ultimate tensile strength of steel reinforcement bars was found to be 126.7 ksi.
Tensile testing of No. 6 bars were not carried out because they were designed to be
remain elastic. However the state of strain in No. 6 bars was monitored by strain
gauges during the tests.
Table 1: Nominal and measured material properties
Specimen Concrete compressive strength (ksi)
Tensile strength of No. 3
bars (ksi)
28-days nominal Measured at the day Measured
Measured
strength
of testing
yield stress ultimate stress
RC
6.7
5.0
81.2
126.7
SFRC
6.3
TEST SET-UP AND INSTRUMENTATION
Both RC and SFRC specimens were subjected to gradually increased monotonic
loading by using a universal testing machine of 400 kips capacity. The loading
magnitude was increased at an interval of 5 kips until the load-carrying capacity of
specimens reduced significantly. As shown in Figures 1(a) and 3(a), a 2-in. diameter
steel roller was placed between two 1-in. thick steel plates at the supports to avoid
local crushing of the concrete due to bearing. Horizontal restraints were applied to the
roller at one support to simulate the hinge support condition, whereas the roller at
the other support was free to move in both ways to prevent the possible confinement
pressure to the supports.

Load (kips)

150

RC

100
SFRC

Design strength

50

0
0.0

0.2

0.4
0.6
Displacement (in)

0.8

(a)
(b)
Figure 3. (a) Test set-up used in the present study (b) Comparison of loaddisplacement response of test specimens

1.0

2010 Structures Congress 2010 ASCE

Extensive instruments were used to measure the response of test specimens at

different load levels. A 600-kip capacity load cell was used at the load point to
monitor the magnitude of load applied to the specimens. Uniaxial strain-gauges
(resistance = 120 ohms; gauge length = 0.2 in.) were bonded to the surface of steel
reinforcements at specified locations to measure the tie forces at various load levels.
As shown in Figure 1(a), four linear variable differential transformers (LVDTs) were
mounted on the surface of test specimens to measure the deformation of concrete
struts during the testing. Also, a linear potentiometer was used exactly below the load
point to measure the deflection of test specimens and two additional linear
potentiometers were also used to monitor the movement of both supports.
TEST RESULTS
The behavior of test specimens under the action of monotonic loading was
monitored at various stages of loading. The primary parameters investigated in the
present study are load versus deflection response, mode of failure and ultimate
strengths of specimens as discussed in the following sections.
Load-Deflection Response
The deflection of the specimens just below the loading point was monitored at
the different stages of loading. The load-deflection behavior of the RC specimen is
shown in Figure 3(b). The RC specimen showed almost linear response up to 95 kips
beyond which yield plateau and strain-hardening was observed up to peak load of
132.1 kips in the load-deflection response. The design load-carrying capacity of the
specimen was 31.3 kips, which was estimated by using the strut-and-tie model. Thus,
the RC specimen was over-designed by a factor of 4.2. The excellent post-yield
behavior of the RC specimen suggested that the bottom longitudinal reinforcement
bars were yielded in tension. This was also confirmed from the state of strains in
these steel reinforcement measured using uniaxial strain gauges. The delaying of
premature local failures of the specimen near the supports due to steel cages at the
boundaries helped the RC specimen to exhibit excellent displacement ductility of
nearly 4.0. As expected, the descending branch of the load-deflection response of the
RC specimen exhibited by the sudden drop from its peak strength.
Figure 3(b) also shows the load-deflection behavior of the SFRC specimen. The
SFRC specimen showed nearly a linear response up to a peak load of 96.8 kips,
which was much higher than the design load of 31.3 kips that used for the RC
specimen. It should be noted that even though there were absolutely no steel
reinforcement bars used as struts and ties as per strut-and-tie models, the SFRC
specimen reached 3.0 times the design load. Further, the SFRC specimen showed
more gradual post-peak descending branch in the load-displacement response in
contrast to a sudden drop in case of the RC specimen indicating significant
contribution of steel fibers to the residual strength of the specimen. The boundary
elements in addition to diagonal steel reinforcement bars helped the specimen to

1986

2010 Structures Congress 2010 ASCE

achieve the design strength without premature local crushing and cracking of concrete
at the boundaries.
Mode of failure
Figure 4(a) shows the observed crack pattern in the RC specimen. As expected,
the local failure at the supports of the RC specimen was not observed during the
entire loading because of sufficient confinement action proved by the steel cage to the
compressive stresses. The major flexural crack running from the bottom face of the
specimen to the loading point was observed at the failure stage. The specimen
collapsed due to shear failure of concrete in the horizontal segment of the opening
due to lack of shear reinforcement as shown in Figure 4(b). This ultimately led to the
fracture of bottom longitudinal tensile reinforcing bars after reaching their failure
strains. It is interesting to note that the major cracks were developed away from the
opening region indicating that the flow of force was least affected by the opening due
to local strengthening of the boundaries near supports.

(a)
(b)
Figure 4. Mode of failure of the RC specimen (a) Overall state of specimen at the
failure stage (b) Shear failure of concrete in bottom segment of the opening

Plastic Hinges

(a)
(b)
Figure 5. Failure mode of the SFRC specimen (a) Specimen at the failure stage
(b) Failure of horizontal segment of the opening

1987

2010 Structures Congress 2010 ASCE

As shown in Figure 5(a), the failure mode of the SFRC specimen was
completely different from that of the RC specimen. A major crack (compressive strut)
was developed just above the opening of the SFRC specimen. Due to presence of
diagonal steel reinforcement bar, the major crack deviated from the corner of the
opening and formed at the end of this bar. A plastic mechanism was developed after
several plastic hinges formed at the specimens (Figures 5(a) and 5(b)). The failure
was fairly ductile as evidenced by the large deformation of the specimen shown in
Figure 5 (a).
Ultimate strength
As stated earlier, the design strength of test specimens was 31.3 kips. The
specimen was analyzed by a strut-and-tie computer program, CAST (Tjhin and
Kuchma, 2002). This program has a feature that allows the analysis of nodes to
ensure that geometry and stress limits are not exceeded. Using nominal (specified)
values of material strength and assuming a strength reduction factor as 0.75, the
nominal ultimate strength of the RC specimen was estimated to be 41.2 kips. The
expected actual capacity of the specimens was 70.3 kips considering the actual
material properties obtained from coupon testing of concrete and steel bars and the
strength-reduction factor as unity. The ultimate strength of the RC specimen was
observed as 132.1 kips which was 1.9 times the expected strength predicted by the
computer model. The main reason for the large difference between measured and
predicted strengths is the redistribution of stress in the specimen after the yielding of
steel bars acting as tie members. Similarly, the SFRC specimen had ultimate strength
of 96.8 kips which was about 1.4 times the predicted ultimate strength of the RC
frame. It should be noted that the ultimate strength of the SFRC specimen was higher
than the predicted ultimate strength (70.3 kips) of the RC specimen even though no
steel reinforcement bars were used as tie members. Further, the fiber bridging effects
limited the widening of cracks and allowed significant stress redistribution and a
plastic mechanism formation upon failure.
SUMMARY & CONCLUSIONS
This paper investigates the behavior of deep beams with large opening that were
designed using strut-and-tie model. One reinforced concrete (RC) and one steel fiber
reinforced concrete (SFRC) specimens were tested under monotonically-increased
loading. The main objectives of this paper were: (a) to investigate the effect of local
strengthening on load-transferring mechanism and failure modes of test specimens;
(b) to study the behavior of SFRC specimen and compare with the RC specimens
designed using strut-and-tie models; and (c) to identify critical regions that are not
identified by strut-and-tie models and to suggest reinforcing detailing to avoid
localized failures.
The following conclusions were drawn in this present study:

1988

2010 Structures Congress 2010 ASCE

1. Design strut-and-tie models significantly underestimated the strength of

specimens. RC specimen showed the load-carrying capacity greater than four
times of the nominal strength predicted using the design strut-and-tie model.
2. SFRC specimen reached the three times of design strength of RC specimen
even though steel reinforcement bars were not used as tie members except
only bottom longitudinal tensile bars.
3. Local strengthening of test specimens using steel cage at the supports
significantly enhanced load-carrying capacity and changed the mode of failure
to a much ductile manner. As a consequence, significant bending (flexural)
action was noticed in the RC specimen without any premature local failures.
4. Steel fibers in the SFRC specimen restrained the widening of crack growth
and increased the number of cracks. This helped in better stress redistribution
in the specimen even if the steel reinforcement bars were not present. A
ductile plastic mechanism was developed after the formation of plastic hinges.
Further research is needed to investigate the effects of volume fraction of steel
fibers on strength of members designed as per strut-and-tie method.
5. The construction of RC specimen was usually time-consuming and laborintensive due to complicated detailing of reinforcing bars in contrast to the
SFRC specimen. Hence, the complete replacement of conventional reinforcing
bars by deformed steel fibers at a volume of 1.5% can be a feasible alternative
to the current practice.
REFERENCES
ACI Committee 318 (2008), Building Code Requirements for Structural Concrete
(ACI 318-08) and Commentary (ACI 318-08), American Concrete Institute,
Farmington Hills, Mich.
ACI Committee 544 (1996), Report on Fiber Reinforced Concrete (ACI 544.1 R-96),
American Concrete Institute, Farmington Hills, Mich.
Brea, S. F,; and Morrison, M. C. (2007), Factors Affecting Strength of Elements
Designed Using Strut-and-Tie Models, ACI Structural J., American Concrete
Institute, 104(3), 267-277.
Flores, C. A. (2009), Performance of large scale steel fiber reinforced concrete deep
beam with single opening under monotonic loading, M. E. Thesis, Department
of Civil Engineering, University of Texas at Arlington, USA
Chen, B. S., Hagenberger, M. J., and Breen, J. E. (2002), Evaluation of Strut-andTie Modeling Applied to Dapped Beam with Opening, ACI Structural J.,
American Concrete Institute, 99(4), 445-450.
Dupont, D., and Vandewalle, L. (2003), Shear Capacity of Concrete Beams
Containing Longitudinal Reinforcement and Steel Fibers, Ed. Banthia, N.;
Criswell, M.; Tatnall, P.; and Folliard, K., Innovations in Fiber-Reinforced
Concrete for Value, American Concrete Institute, Farmington Hills, Mich.

1989

2010 Structures Congress 2010 ASCE

Kuchma, D., and Park, J. W. (2007), Strut-and-Tie Model Analysis for Strength
Prediction of Deep Beams, ACI Structural J., American Concrete Institute,
104(6), 657-666.
Kuchma, D., Yindeesuk, S., Nagle, T., Hart, J., Lee, H. H. (2008), Experimental
Validation of Strut-and-Tie Method for complex Regions, ACI Structural J.,
American Concrete Institute, 105(5), 578-589.
Mansur, M. A., and Ong, K. C. G. (1991), Behavior of Reinforced Fiber Concrete
Deep Beams in Shear, ACI Structural J., American Concrete Institute, 88(1),
98-105.
Maxwell, B. S., and Breen, J. E. (2000), Experimental Evaluation of Strut-and-Tie
Model Applied to Deep Beam with Opening, ACI Structural J., American
Concrete Institute, 97(1), 142-148.
Narayanan, R., and Darwish, I. Y. S. (1988), Fiber Concrete Deep Beams in Shear,
ACI Structural J., 85(2), 141-149.
Schlaich, J., Schfer K., and Jennewein, M. (1987), Toward a Consistent Design of
Structural Concrete, PCI J., Precast Prestressed Concrete Institute, 32(3), 74150.
Tan, K. H., and Zhang, N. (2007), Size effect in RC deep beams: Experimental
investigation and STM verification, Engineering Structures, 29, 3241- 3254.
Tjhin, T. N., and Kuchma, D. A. (2002), Computer-Based Tools for Design by Strutand-Tie Method: Advances and Challenges, ACI Structural J., American
Concrete Institute, 99(5), 586-594.

1990