Structures 26 (2020) 785–800

Contents lists available at ScienceDirect

Structures
journal homepage: www.elsevier.com/locate/structures

Behavior of reinforced concrete deep beam with web openings strengthened T

with (CFRP) sheet
Waleed A. Jasima, Yazan B. Abu Tahnatb, , Abdulsamee M. Halahlac
⁎

a
Ministry of Construction and Housing and Public Municipalities, Projects Dept., 10071 Baghdad, Iraq
b
Department of Civil Engineering, Middle East Technical University, Ankara, Turkey
c
Department of Civil Engineering, Fahad Bin Sultan University, Tabuk, Saudi Arabia

ARTICLE INFO ABSTRACT

Keywords: Deep beams are used in various applications in reinforced concrete (R.C.) structures. There have been continuous
Carbon fiber-reinforced polymer (CFRP) efforts to enhance and improve the performance of these crucial elements in (R.C.) structures by using several
Deep beam strengthening techniques such as using the carbon fiber-reinforced polymer (CFRP). However, by exploring the
Finite element literature, none of the previously conducted experimental tests have studied the propagation of cracks beneath
Shear span ratio
the (CFRP) sheets. In this research, the propagation of the first diagonal crack, which takes place beneath the
Concrete damage plasticity (CDP)
(CFRP), is investigated by modeling sixteen (R.C.) deep beams with different opening sizes. Two shear span/
depth ratios (a/h) are studied numerically using the finite element analysis tool (ABAQUS). All models are
validated using the concrete damage plasticity (CDP) model, and their results are found similar to the experi-
mental results obtained by other authors. Results show that the Finite Element models catch the real behavior of
the (R.C.) deep beams. In addition, the (CFRP) sheets are found to enhance the failure load capacity as well as
the flexural crack remarkably. Moreover, the (CFRP) increases the load required to cause the first diagonal crack
for models with (a/h) equals 0.9, while there is a slight change in this load for models with (a/h) equals 1.1.

1. Introduction amount and type of the web reinforcement, width of the support zone,
anchorage of the main reinforcement, concrete strength, inclusion of
Reinforced Concrete (R.C.) deep beams are indispensably used in other materials such as fibers…etc. [6]. ACI-ASCE Committee 426 [7]
many types of structures. Reinforced concrete (R.C.) deep beams have classifies the failure of (R.C.) deep beams without web openings into
been used in high-rise buildings, offshore structures, transfer girders, three types which are, flexural failure, flexural-shear failure, and di-
some walls, and pile caps [1,2]. A beam is considered deep when it has agonal splitting failure. Moreover, there are other types of failure of
a clear span length not greater than four times the beam depth. Adding (R.C.) deep beams without openings such as diagonal compression
to that, in case of an applied concentrated load, the application position failure, bearing failure and anchorage failure. On the other hand,
is at a distance not to exceed double the beam depth measured from the Mansur et al. [8] identified three categories of failure modes – mode A,
support face [3]. Moreover, Hassani et al. [4] define deep beams as the B and C. Mode A is a typical shear failure that occurs as a result of
beams where transference of the load takes place significantly through diagonal cracks that extend through a direct path from the loading
a compression thrust connecting the load with the reaction. Generally, point to the supporting point. Mode B is a collapse caused by cracks that
the term deep beams refers to the beams with high depth-to-span ratios, propagate diagonally along the two critical paths running from the
to the degree that keeps the elastic flexural stress distribution linear loading points as well as the reaction points to the furthest corners of
over the beam depth, and causes non-parabolic shear stress distribu- the opening. Mode C occurs gradually when deformation takes place
tion. Having these types of stresses combined results in inclined cracks mainly in the shear span, and at the same time three different portions
within the shear span. When this occurs, the beam is transformed into a of the beam experience relative rotations.
tied-arch [5]. Several researchers have studied the behavior of (R.C.) deep beams.
The behavior of (R.C.) deep beams is influenced by many factors Demir et al. [9] performed an experimental test to investigate the effect
such as the clear span/depth ratio (Ln/h), shear span/depth ratio (a/h), of several parameters on the behavior of (R.C.) deep beam. The results
type of loading, position of the load, percentage of the tensile steel, showed that by decreasing span to depth ratio (a/d), the shear strength

⁎
Corresponding author.
E-mail address: abutahnat.yazan@metu.edu.tr (Y.B.A. Tahnat).

https://doi.org/10.1016/j.istruc.2020.05.003
Received 16 February 2020; Received in revised form 20 April 2020; Accepted 4 May 2020
2352-0124/ © 2020 Institution of Structural Engineers. Published by Elsevier Ltd. All rights reserved.
W.A. Jasim, et al. Structures 26 (2020) 785–800

Fig. 1. Arrangement of steel reinforcement for the experimental specimens.

Table 1
Mechanical properties of steel reinforcement.
Nominal diameter [mm] Measured Diameter [mm] Area [mm2] Yield stress f y [MPa] Ultimate stress fu [MPa] Elongation [%]

6 5.5 23. 75 623.96 685.5 8.5

16 16 200.96 569.67 668.79 12.5

Table 2
Details of the eighteen tested reinforced concrete deep beams.
Beams No. Specimen Designation Opening location Shear Span/Depth (a/d) Opening Size [mm] Implementation External (CFRP)

B1 DP-S1 – 1.1 – –
B2 DP-S2 – 0.9 – –
B3 DP-S1-C-O1-WS Center 1.1 200 × 200 NO
B4 DP-S1-C-O1-S Center 1.1 200 × 200 YES
B5 DP-S1-C-O2-WS Center 1.1 230 × 230 NO
B6 DP-S1-C-O2-S Center 1.1 230 × 230 YES
B7 DP-S1-E-O1-WS Edge 1.1 200 × 200 NO
B8 DP-S1-E-O1-S Edge 1.1 200 × 200 YES
B9 DP-S1-E-O2-WS Edge 1.1 230 × 230 NO
B10 DP-S1-E-O2-S Edge 1.1 230 × 230 YES
B11 DP-S2-C-O1-WS Center 0.9 200 × 200 NO
B12 DP-S2-C-O1-S Center 0.9 200 × 200 YES
B13 DP-S2-C-O2-WS Center 0.9 230 × 230 NO
B14 DP-S2-C-O2-S Center 0.9 230 × 230 YES
B15 DP-S2-E-O1-WS Edge 0.9 200 × 200 NO
B16 DP-S2-E-O1-S Edge 0.9 200 × 200 YES
B17 DP-S2-E-O2-WS Edge 0.9 230 × 230 NO
B18 DP-S2-E-O2-S Edge 0.9 230 × 230 YES

Fig. 2. Schematic of (CFRP) strengthening.

Table 3 failure may occur as a result of diagonal cracking [10,11].

Results of Hardened Concrete Specimens. In addition to the above, recent studies have focused on strength-
Item ft (MPa) fr (MPa) Ec (MPa)
ening the openings of deep beams by using (CFRP) and found that
f c' (MPa) υ
(CFRP) enhances the behavior of deep beams remarkably [12–17].
Value 27 3.1 3.5 24,667 0.2 Farghaly and Benmokrane [18] investigated the shear behavior of four
large-scale specimens of deep beams reinforced with carbon and glass
fiber reinforced polymer (FRP) bars with different longitudinal re-
of the tested members increases. Moreover, the width of the diagonal inforcement ratios. They noticed that the reinforcement ratio and
crack increases when (a/d) ratio increases. Other researchers showed concrete compressive strength significantly affect the ultimate capacity
that deep beams can retain considerable strength after the development and deflection. Some researchers focused on the effect of the geometry
of diagonal cracks. Once this strength is exceeded, sudden and brittle of concrete deep beams. A study made by Rezaei et al. [19] to

786
W.A. Jasim, et al. Structures 26 (2020) 785–800

Fig. 3. Tension model of concrete.

Table 4
Parameters of damage-plasticity model.
Parameter name Value

Dilation angle (ψ) 36°

Eccentricity (e) 0.1
fb0 / fc 0 (ratio of initial equibiaxial compressive yield stress to initial 1.16
uniaxial compressive yield stress
K (the ratio of the second stress invariant on the tensile meridian) 0.667

investigate the effect of reinforcement development length and beam

geometry on the behavior of discontinuity regions in concrete deep
beams experimentally and numerically. It was noticed that the devel-
opment length influences the generation of tensile stresses near sup-
ports. Moreover, beams of rectangular cross sections were found sub-
jected to diagonal tension stresses developed in the struts, but they were Fig. 5. Schematic of unidirectional (FRP).

Fig. 4. Definition of concrete damage-plasticity model parameters in ABAQUS used for parametric study.

787
W.A. Jasim, et al. Structures 26 (2020) 785–800

Table 5 CDP model in ABAQUS to investigate the effect of reinforcement dis-

Properties of (CFRP) sheets and epoxy. tribution and the number of point loads on the (R.C.) deep beams with
Item Ef [MPa] Vf Ea [MPa] νf νa and without web openings and validated the results using experimental
tests. Other parameters such as the size and geometry of openings and
Value 231,000 0.75 3800 0.2 0.38 the shear span ratio were numerically investigated by Senthil et al. [27]
using ABAQUS. The results showed that the models are capable of
capturing the entire response reasonably.
Table 6 In this research, the numerical analyses are carried out by im-
Properties for combined (CFRP) sheets with epoxy. plementing the (CDP) model in ABAQUS 6.14.1 [28] and verified by the
Item E1 [MPa] E2 [MPa] v23 G12 [MPa] G13 [MPa] G23 [MPa] experimental works which were conducted by Jasim et al. [29]. The
propagation of the first diagonal crack beneath the (CFRP) sheets,
Value 174,200 14485.15 0.245 5280.6 5280.6 5817.3
which is difficult to be inspected experimentally, has been investigated
based on numerical studies. Moreover, the (F.E.) method is used to
study the role of different parameters in changing the ultimate capacity,
cracking load and mid-span deflection response of the tested (R.C.)
deep beams. The studied parameters include the shear span/depth ratio
(a/h), the locations and sizes of the web openings and the strengthening
of the openings by (CFRP).

2. Summary of the previous experimental tests

Eighteen experimental tests were conducted by Jasim et al. [29] to

investigate the effect of openings on the shear behavior of deep beams
and the effect of opening location whether it is at mid-span or at the
edge of the shear zone. They also investigated using carbon fiber re-
inforced polymer sheets to strengthen the reinforced concrete at the
openings against shear failure. The variables investigated by Jasim
et al. [29] are the size of the openings and their locations, shear span/
depth ratio, and the effect of using (CFRP) sheets at the locations of
high stresses developed due to openings. All beams had the same di-
Fig. 6. Modeling of simply supported boundary condition. mensions. They had an overall length of (1500 mm), a width of
(150 mm) and a height of (500 mm) and they were designed to guar-
antee failure in shear. The deep beams were tested under two-point
symmetric loading with an overall clear span of (1200 mm). This re-
sulted in a ratio of clear span to an overall depth equals to (2.4) which is
less than (4.0). The average cylinder compressive strength of concrete
at 28 days age was 27 MPa. All deep beams had the same interior steel
reinforcement as shown in Fig. 1. The mechanical properties of steel
reinforcement are summarized in Table 1.
Moreover, all details related to the sizes and location of openings,
shear span/depth ratio and the implementation of (CFRP) can be seen
in Table 2. The schematic of (CFRP) strengthening is shown in Fig. 2.

3. Numerical model

In the present study, the nonlinear finite element (F.E.) method

utilizing ABAQUS software environment is used in order to simulate the
simply supported (R.C.) deep beams with and without large openings.
Fig. 7. Meshing of deep beam without openings.
3.1. Materials properties
absent in truss-like specimens. Both stresses that are influenced by the
reinforcement development length and the geometry of the beam were Materials properties used in the model, such as properties of con-
found non additive. crete, support plates, steel reinforcement, and (CFRP) fabric are defined
For the purpose of modeling the concrete material, several con- in ABAQUS. Two types of material properties are defined, linear and
stitutive models were proposed such as Drucker-Prager [20], Ottosen nonlinear. For the linear analysis, it is necessary to define at least two
[21] and Willam-Warnke [22]. However, the concrete damage plasti- linear material properties: Young’s modulus E and Poisson’s ratio υ,
city model (CDP) is among the most preferable because it can reflect the whereas for the nonlinear analysis, the uniaxial behavior of the steel
degradation in strength and stiffness by means of tension and com- and concrete should be determined beyond the elastic range to simulate
pression damages parameters. Many researchers used the (CDP) model their performance at high strains.
in simulating different structural elements and they found that this
model can capture the real behavior of the tested reinforced concrete 3.1.1. Concrete modelling
elements [23–25]. The CDP model is available in ABAQUS software, Mainly, there are two material modeling approaches for concrete in
and many researchers have used this capability in ABAQUS to predict ABAQUS, which are concrete smeared cracking and concrete damaged
the real behavior of (R.C.) deep beams. Mohamed et al. [26] used the plasticity model (CDP). Both models can be used to model plain and

788
W.A. Jasim, et al. Structures 26 (2020) 785–800

Fig. 8. Meshing of deep beam with openings and (CFRP) sheets.

Table 7
Comparison of the deflection for all specimens.
Beam No. Specimen Designation Pu (exp) [KN] Pu (ABAQUS) [kN] Pu (exp)/Pu (ABAQUS) Δu (exp) [mm] Δu (ABAQUS) [mm] Δu (exp)/Δu (ABAQUS)

B1 DP-S1 500 570 0.88 9.00 7.90 1.14

B2 DP-S2 560 577 0.97 8.20 8.48 0.97
B3 DP-S1-C-O1-WS 190 220 0.86 4.00 4.30 0.93
B4 DP-S1-C-O1-S 280 285 0.98 5.30 5.81 0.91
B5 DP-S1-C-O2-WS 175 190 0.92 4.30 4.29 1.00
B6 DP-S1-C-O2-S 210 240 0.88 6.20 6.10 1.02
B7 DP-S1-E-O1-WS 210 220 0.95 4.80 4.98 0.96
B8 DP-S1-E-O1-S 300 275 1.09 6.20 5.70 1.09
B9 DP-S1-E-O2-WS 170 205 0.83 5.10 5.32 0.96
B10 DP-S1-E-O2-S 250 250 1.00 6.80 6.40 1.06
B11 DP-S2-C-O1-WS 240 235 1.02 4.00 4.43 0.90
B12 DP-S2-C-O1-S 290 285 1.02 5.10 5.68 0.90
B13 DP-S2-C-O2-WS 200 200 1.00 4.00 4.27 0.94
B14 DP-S2-C-O2-S 250 245 1.02 5.50 6.12 0.90
B15 DP-S2-E-O1-WS 250 235 1.06 4.00 4.40 0.91
B16 DP-S2-E-O1-S 330 295 1.12 7.00 7.56 0.93
B17 DP-S2-E-O2-WS 210 205 1.02 4.60 5.16 0.89
B18 DP-S2-E-O2-S 300 305 0.98 6.10 5.79 1.05

Fig. 9. Comparison of the ultimate load for all specimens.

reinforced concrete. In the present study, concrete damage plasticity compression stress. Some of these examples, such as those introduced
model is used to model the concrete. Modulus of elasticity Ec , concrete by Mander et al. [30] and Yong et al. [31], have nevertheless failed to
tensile strength ft , rupture stress fr , compressive strength f 'c and provide the full stress-strain curve of concrete. To overcome this
Poisson’s ratio υ are taken as presented in Table 3 for all models. shortcoming, the stress-strain equation suggested by Saenz [32] and
A number of equations have been presented by several researchers verified by Asran et al. [33], Tahnat, et al. [34] and Halahla et al. [35]
to reflect concrete behavior when it is subjected to uniaxial are adopted to give the entire behavior of concrete when subjected to

789
W.A. Jasim, et al. Structures 26 (2020) 785–800

Fig. 10. Comparison of the Deflection for all specimens.

Fig. 11. Comparison of load-deflection relation for all specimen.

uniaxial compressive stress as given by Eqs. (1)–(7). Ec = 4700 f c' , for normal strengh concrete [3] (2)
Ec c
c = RE (R 1) 1
() ()
2 3
1 + (R + RE 2) c
(2R 1) c
+R c R=
0 0 0 (1) (R 1)2 R (3)

790
W.A. Jasim, et al. Structures 26 (2020) 785–800

Fig. 11. (continued)

RE =
Ec R : Strain ratio, which is equal 4 as reported by Hu and Schnobrich
E0 (4) [36].
For the stress-strain curve of concrete under tension used in the
f
R = c model, it is the one proposed by Nayal and Rasheed [37] and modified
f (5) by Wahalathantri et al. [38] as shown in Fig. 3. (a) and (b) respectively.
This model has been used by many researchers and proved its effec-
f
R = tiveness [35] and [39].
(6)
0
Modeling of concrete requires considering four parameters ac-
f c
cording to the (CDP) model in order to capture the behavior of concrete
E0 = accurately. These parameters are summarized in Table 4. In addition,
0 (7)
the compression damage parameter dc represents the decay in the
Where: elastic stiffness due to compressing the concrete, while the tension
c :Concrete compressive stress (MPa) damage parameter dt represents the decay in the elastic stiffness due to
Ec :Modulus of elasticity of concrete (MPa) tensioning the concrete. Both parameters are calculated following the
E0 : Secant modulus of concrete (MPa) method proposed by Lima et al. [40].
fc' : Maximum compressive strength of concrete (MPa) This model uses concrete with compressive strength of 27 MPa and
c : Compression strain elasticity modulus of 24,422 MPa. However, Fig. 4 shows the required
0 : Strain corresponding to f c which approximately equals 0.0025 as
'
data that must be used in defining concrete material in ABAQUS. Fig. 4
reported by Hu and Schnobrich [36]. (a) shows the uniaxial compression stress versus the inelastic strain of
f : Maximum strain. concrete, while Fig. 4 (b) shows the tension stress versus the cracking
f : Stress at maximum strain (MPa). strain of concrete. The curve in Fig. 4 (c) displays the relation between
R : Ratio relation the compression damage parameter and the inelastic strains, while
RE : Modular ratio. Fig. 4 (d) shows the change in the tension damage parameter at various
R : Stress ratio, which is equal 4 as reported by Hu and Schnobrich cracking strains.
[36].

791
W.A. Jasim, et al. Structures 26 (2020) 785–800

Fig. 11. (continued)

3.1.2. Steel modelling modeled as a lamina linear elastic element as displayed in Fig. 5.
The steel reinforcement and stirrups are modeled analytically as Moreover, Table 5 shows the properties of the (CFRP) sheets and the
bilinear isotropic materials. For the linear isotropic part, it is defined by epoxy material used in this study.
the modulus of elasticity of the reinforcement and the Poisson’s ratio, Eqs. (8)–(13), suggested by Mallick [41], are used to evaluate the
which are taken as 200 × 103 MPa and 0.3, respectively. For the bi- mechanical properties of the combined (CFRP) sheet and adhesion. A
linear isotropic part, it is defined by the yield stress fy and the ultimate summary of the properties of the combined (FRP) sheets and the ad-
stress fu as presented in Table 1. hesive material is given in Table 6.
The steel bearing plates and support plates are modeled as linear
isotropic materials. For the linear isotropic part, it is defined by the
E1 = Ef Vf + Ea (1 Vf ) (8)
modulus of elasticity of steel plates multiplied by 10. It therefore equals
200 × 103 MPa. Multiplication by 10 implies that plates are considered Ef Ea
E2 =
rigid, and this prevents deflections accordingly. Ea Vf + Ef (1 Vf ) (9)

Gf Ga
G12 = G13 =
3.1.3. (CFRP) modelling Ga Vf + Gf (1 Vf ) (10)
Unidirectional (FRP) sheets are used as a strengthening material to
the (R.C.) deep beam model. Existence of fibers increases the stiffness as E2
well as the capacity of the (FRP) composite steel to carry higher loads. G23 =
2(1 + v23 ) (11)
The purpose of using the matrix is to distribute the load over all fibers,
and to provide protection of fibers against the environmental impacts.
v23 = vf Vf + va (1 Vf ) (12)
Fibers possess a linear elastic behavior up to a rupture failure. (CFRP) is

792
W.A. Jasim, et al. Structures 26 (2020) 785–800

(1 Vf ) Ea rupture failure. For the contact between reinforcement and concrete it

co = Vf u + u is considered embedded contact. This type of contact allows friction to
Ef (13)
act between concrete and reinforcement surfaces.
Where:
E1: Elastic modulus in the longitudinal direction 3.3. Load and boundary conditions
E2 : Elastic modulus in the transverse direction
G12, G13 : Plane shear moduli Two simple supports are assigned to the deep beam ends as shown
G23: Normal to plane shear modulus in Fig. 6. One support prevents translations in the X,Y, and Z directions
v : Poisson’s ratio to act as a pin support, while the second support restrains translation in
co : Ultimate tensile strength the X and Y directions only, and therefore it acts as a roller support.
Ef : Elastic modulus of (CFRP)
Vf : Volume fraction of (CFRP), provided by the manufacturer 3.4. Meshing type and sensitivity study
Ea : Elastic modulus of the adhesive material
Gf : Shear modulus of (CFRP) The meshing process is applied on a part-by-part basis to the com-
Ga : Shear modulus of the adhesive material ponents of the deep beam. Neither the global nor the sweep meshing are
used. Eight-noded linear brick elements (C3D8R) are used to model the
3.2. Modelling of interfaces solid elements; concrete and the loading plate. 2-node linear 3-D truss
elements are used to model the main and the transverse reinforcement
The interfacial behavior between two surfaces can be simulated (T3D2), whereas 4-noded shell elements (S4R) are used to model the
using different contact models. The proper model is selected based on (CFRP) as shown in Fig. 7. Fig. 8 also shows the meshes of the model
the actual behavior and the expected accuracy level. Many researchers with (CFRP) used around openings. A sensitivity study is performed to
studied the interaction between concrete and (CFRP) [42–46]. Chen determine the optimum mesh size which has minimum impact on the
et al. [47] investigated the effect of different contact models between accuracy of results. Accordingly, mesh size of 20 mm is adopted in
concrete and (CFRP) on the predicted shear behavior of (R.C.) beams, every model.
shear-strengthened with (FRP). The interfacial element between (FRP)
and concrete is used to catch the proper behavior of interaction be- 4. Verification of models
tween concrete and (CFRP). In this paper the interfacial (CFRP)-con-
crete contact is assumed as a tie contact due to the fact that the failure The numerical analysis results for all specimens are presented and
mode reported in the previous experimental works on (CFRP) was compared with experimental results extracted from the literature. (F.E.)

Fig. 12. Comparison between tension damage from (F.E.) and experimental test for all specimens.

793
W.A. Jasim, et al. Structures 26 (2020) 785–800

Fig. 12. (continued)

analysis is applied using the static analysis. in the F.E analysis results. For example, the connection between con-
In this section, several comparisons are established with the ex- crete and steel bars is assumed fully bonded which reduces the actual
perimental results. These include the first diagonal crack Pd , the flexural specimen stiffness. Therefore, the overall stiffness of the real specimen
crack Pf , and the ultimate load capacity Pu . In addition, the load-de- is found lower than that expected during the (F.E.) analysis [48]. Some
flection response curves are also compared with the experimental re- assumptions in the numerical model are presumed standing behind the
sults. difference between the experimental and the numerical results, in-
Results extracted from the static (F.E.) analysis for the load versus cluding boundary conditions, homogeneity of concrete material, post
deflection relation are presented in Table 7. The ultimate force Pu and failure phenomena (softening behavior of the concrete) and time de-
deflections Δu for all specimens which are obtained from the (F.E.) pendent effects [49].
analysis are compared with those obtained from the experimental tests
as shown in Figs. 9 and 10, respectively. The results indicate a good 5. Numerical results and discussion
agreement between Finite Element and the experimental tests results.
Moreover, the load versus deflection curves obtained from the (F.E.) 5.1. Failure mode and damage evolution
analysis are found close to those obtained from the experimental tests
as shown in Fig. 11., with the numerical models observed having higher The Finite Element model shows that there are two modes of failure.
stiffness. This is due to some influences that may cause higher stiffness The first mode is the diagonal splitting failure which occurs when the

794
W.A. Jasim, et al. Structures 26 (2020) 785–800

Fig. 12. (continued)

diagonal crack starts from the corners of the opening and propagates 5.2. The effect of (CFRP) on the behavior of reinforced concrete deep beam
towards the applied load and support. The second mode is the shear- with opening
compression failure which takes place when the diagonal crack de-
velops in the shear span and results in high strains in the compression (CFRP) sheets increase the capacity of deep beams in general. The
chord of the openings close to the location of loads. Fig. 12 shows how strength of such members are dramatically affected by the size of
the (F.E.) simulation results in failures that are in good agreement with openings. Finite element results lead to the conclusion that as the size of
those observed experimentally. opening is increased from 200 × 200 mm to 230 × 230 mm, the first
However, for specimens covered with (CFRP) sheets, (CFRP) con- diagonal cracking, the first flexural cracking and the failure loads all
tributes to minimize the cracks significantly. The Finite Element model decrease. (CFRP) sheets enhance the failure load capacity and raise it
can show the cracks in concrete under the (CFRP) sheet. It is obvious from (21.3%–48.7%) as well as the flexural crack from almost
that there are very mild diagonal cracks developed over the surface of (7.14%–33.3%). However the Finite Element results show that the first
the (CFRP) sheets at the top and bottom chords of the openings. diagonal crack appears earlier in the specimens with openings of size

795
W.A. Jasim, et al. Structures 26 (2020) 785–800

Fig. 12. (continued)

230 × 230 mm. On the other hand, (CFRP) sheets increase the load 6. Conclusion
required for the first diagonal crack from almost (6%–30%) for all
models with low shear span ratio, while there is a slight change for In this article, a three-dimensional (3-D) non-linear Finite Element
models with high shear span ratio, as shown in Table 8. (F.E.) model of a reinforced concrete (R.C.) deep beam is verified.
Eighteen experimental tests are used to conduct the verification.
5.3. The effect of span ratio on the first diagonal crack Concrete damage plasticity model (CDP) is used to simulate the com-
plex nonlinear behavior of concrete. The final conclusion remarks of
The finite element results lead to the conclusion that as (a/h) ratio this study are listed below:
increases, the load required for the first diagonal crack, flexural cracks
and the failure loads all decrease. Results show that in the models with 1) Results show that the Finite Element models catch the real behavior
a shear span/depth ratio (a/h) equals 0.9, the load that causes the first of (R.C.) deep beam. Moreover, results indicate the capability of the
diagonal crack is more than in the models with shear span/depth ratio F.E. model to predict to a high accuracy the probable cracks that
(a/h) equals 1.1. This is clearly noticed in Table 8. would occur in experiments. In addition, from the comparisons

796
W.A. Jasim, et al. Structures 26 (2020) 785–800

Fig. 12. (continued)

between the experimental and the numerical results, it can be ob- decrease. When (CFRP) sheets are added to the models with (a/
served that the numerical models are stiffer than the experimental h = 1.1), they cause a slight change in the value of the load required
ones at both linear and nonlinear parts of the behavior, but there is a to initiate the first diagonal crack. However, (CFRP) causes this load
good agreement between them, which makes it acceptable to use the to increase from almost (6%–30%) in models with (a/h = 0.9).
Finite Element models to perform the parametric studies related to 4) The strength of (R.C.) deep beam is dramatically affected by the size
this topic. of openings. Based on the Finite Element results it can be concluded
2) Results show that the (CFRP) sheets increase the failure load capa- that, as the size of opening is increased from 200 × 200 mm to
city and the load at flexural crack from about (21.3%–48.7%) and 230 × 230 mm, the first diagonal cracking, the first flexural
(7.14%–33.3%), respectively. cracking and the failure loads all decrease.
3) Results show that as (a/h) ratio increases, the load required to cause
the first diagonal crack, the flexural cracks and the failure loads all

797
W.A. Jasim, et al. Structures 26 (2020) 785–800

Fig. 12. (continued)

798
W.A. Jasim, et al. Structures 26 (2020) 785–800

Table 8
Comparison of the first cracks and the ultimate load capacity (at failure load).
Beam NO. Specimen Designation (CFRP) (a/h) Pd [KN] Improvement Pd % Pf [KN] Improvement Pd % Pu [KN] Improvement Pu %

B1 DP-S1 NO 1.1 100 – 120 – 570 –

B2 DP-S2 NO 0.9 110 – 135 – 577 –
B3 DP-S1-C-O1-WS NO 1.1 65 REF 70 REF 220 REF
B4 DP-S1-C-O1-S YES 1.1 65 0 70 0 285 29.5
B5 DP-S1-C-O2-WS NO 1.1 50 REF 55 REF 190 REF
B6 DP-S1-C-O2-S YES 1.1 50 0 60 9 240 26.3
B7 DP-S1-E-O1-WS NO 1.1 60 REF 70 REF 220 REF
B8 DP-S1-E-O1-S YES 1.1 60 0 75 7.14 275 25
B9 DP-S1-E-O2-WS NO 1.1 50 REF 60 REF 205 REF
B10 DP-S1-E-O2-S YES 1.1 65 30 75 25 250 22
B11 DP-S2-C-O1-WS NO 0.9 70 REF 80 REF 235 REF
B12 DP-S2-C-O1-S YES 0.9 70 0 90 12.5 285 21.3
B13 DP-S2-C-O2-WS NO 0.9 55 REF 65 REF 200 REF
B14 DP-S2-C-O2-S YES 0.9 60 10 75 15.3 245 22.5
B15 DP-S2-E-O1-WS NO 0.9 70 REF 80 REF 235 REF
B16 DP-S2-E-O1-S YES 0.9 75 7.1 85 6.3 310 32
B17 DP-S2-E-O2-WS NO 0.9 55 REF 60 REF 205 REF
B18 DP-S2-E-O2-S YES 0.9 65 18.18 80 33.3 305 48.7

Declaration of Competing Interest 1977;103(4):527–35.

[22] William K, Warnke E, Constitutive model for the triaxial behaviour of concrete. In:
Proc. int. ass. for bridge and struct. engng.; 1975, 1–30.
The authors declare that they have no known competing financial [23] Lee SH, Abolmaali A, Shin KJ, Lee HD. ABAQUS modeling for post-tensioned re-
interests or personal relationships that could have appeared to influ- inforced concrete beams. J Build Eng 2020;1(30):101273.
[24] Kenawy M, Kunnath S, Kolwankar S, Kanvinde A. Concrete uniaxial nonlocal da-
ence the work reported in this paper. mage-plasticity model for simulating post-peak response of reinforced concrete
beam-columns under cyclic loading. J Struct Eng 2020;146(5):04020052.
References [25] Feng DC, Wu G, Lu Y. Finite element modelling approach for precast reinforced
concrete beam-to-column connections under cyclic loading. Eng Struct
2018;1(174):49–66.
[1] Zhang Z, Hsu C-TT, Moren J. Shear strengthening of reinforced concrete deep beams [26] Mohamed AR, Shoukry MS, Saeed JM. Prediction of the behavior of reinforced
using carbon fiber reinforced polymer laminates. ASCE 2004;8(5):403–14. concrete deep beams with web openings using the finite element method.
[2] Ashour AF, Alvarez LF, Toropov VV. Empirical modelling of shear strength of RC Alexandria Eng J 2014;53(2):329–39.
deep beams by genetic programming. Comput Struct 2003;81(5):331–8. [27] Senthil K, Gupta A, Singh SP. Computation of stress-deformation of deep beam with
[3] ACI. Building Code Requirements for Structural Concrete. Committee report 318-14, openings using finite element method. Adv Concr Constr 2018;6(3):245.
American Concrete Institute, Detroit, Michigan 2014. [28] ABAQUS version 6.14.1 (2014). [Computer software]. Assault Systems.
[4] Mohammadhassani M, Nezamabadi-Pour H, Suhatril M, Shariati M. Identification of Waltham. MA.
a suitable ANN architecture in predicting strain in tie section of concrete deep [29] Jasim WA, Allawi AA, Oukaili NK. Strength and serviceability of reinforced con-
beams. Struct Eng Mech 2013;46(6):853–68. crete deep beams with large web openings created in shear spans. Civil Eng J
[5] Yagendra P. Serviceability considerations for reinforced concrete deep beams. J 2018;4(11):2560–74.
Struct Eng 1976;4(1):33–8. [30] Mander JB, Priestley MJ, Park R. Theoretical stress-strain model for confined
[6] Subedi NK, Vardy AE, Kubotat N. Reinforced concrete deep beams some test results. concrete. J Struct Eng 1988;114(8):1804–26.
Mag Concr Res 1986;38(137):206–19. [31] Yong YK, Nour MG, Nawy EG. Behavior of laterally confined high-strength concrete
[7] ACI-ASCE Committee 426. The shear strength of reinforced concrete members under axial loads. J Struct Eng 1988;114((2):332–51.
Chapters 1 to 4. Proceedings. J Struct Div ASCE 1973;99:1091–187. [32] Saenz LP. Equation for the stress-strain curve of concrete. ACI J
[8] Mansur MA, Alwis WAM. Reinforced fiber concrete deep beams with web openings. 1964;61(9):1229–35.
Int J Cem Compos Lightweight Concrete 1984;4:263–71. [33] Asran AG, El-Esnawi HH, Fayed SE. Experimental investigation of RC exterior beam
[9] Demir A, Caglar N, Ozturk H. Parameters affecting diagonal cracking behavior of column connection with eccentric beam subjected to reversible quasi static loads.
reinforced concrete deep beams. Eng Struct 2019 Apr;1(184):217–31. Civil Eng J 2016;3(4):221–36.
[10] Wight JK. Reinforced concrete mechanics and design. New Jersey: Pearson [34] Tahnat YBA, Dwaikat MM, Samaaneh MA. Effect of using CFRP wraps on the
Education Inc.; 2016. strength and ductility behaviors of exterior reinforced concrete joint. Compos Struct
[11] Dogangun A. Design and calculation of reinforced concrete structures. Istanbul: 2018;201:721–39.
Birsen Publishing House; 2012. [35] Halahla AM, Tahnat YB, Almasri AH, Voyiadjis GZ. The effect of shape memory
[12] Ibrahim M, Wakjira T, Ebead U. Shear strengthening of reinforced concrete deep alloys on the ductility of exterior reinforced concrete beam-column joints using the
beams using near-surface mounted hybrid carbon/glass fibre reinforced polymer damage plasticity model. Eng Struct 2019;200:109676.
strips. Eng Struct 2020;1(210):110412. [36] Hu HT, Schnobrich WC. Constitutive modeling of concrete by using nonassociated
[13] Albidah A, Abadel A, Abbas H, Almusallam T, Al-Salloum Y. Experimental and plasticity. J Mater Civ Eng 1989;1(4):199–216.
analytical study of strengthening schemes for shear deficient RC deep beams. Constr [37] Nayal R, Rasheed HA. Tension stiffening model for concrete beams reinforced with
Build Mater 2019;20(216):673–86. steel and FRP bars. J Mater Civ Eng 2006;18(6):831–41.
[14] Lee HK, Cheong SH, Ha SK, Lee CG. Behavior and performance of RC T-section deep [38] Wahalathantri BL, Thambiratnam DP, Chan TH, et al. A material model for flexural
beams externally strengthened in shear with CFRP sheets. Compos Struct crack simulation in reinforced concrete elements using ABAQUS. In: First interna-
2011;93(2):911–22. tional conference on engineering, designing and developing the built environment
[15] Zhang Z, Hsu CT, Moren J. Shear strengthening of reinforced concrete deep beams for sustainable wellbeing, Brisbane, Australia 2011, pp. 260-264. Queensland
using carbon fiber reinforced polymer laminates. J Compos Constr University of Technology.
2004;8(5):403–14. [39] Chen A, Yossef M, Hopkins P. A comparative study of different methods to calculate
[16] Abdel-Jaber MS, Walker PR, Hutchinson AR. Shear strengthening of reinforced degrees of composite action for insulated concrete sandwich panels. Eng Struct
concrete beams using different configurations of externally bonded carbon fiber 2020;1(212):110423.
reinforced plates. Mater Struct J 2003;36(5):291–301. [40] Ho NM, Lima MM, Doh JH. Axially loaded three-side restrained reinforced concrete
[17] Nie XF, Zhang Shi Shun, Teng Jin Guang, Chen GM. Experimental study on RC T- walls: a comparative study. In: Mechanics of Structures and Materials XXIV:
section beams with an FRP-strengthened web opening. Compos Struct Proceedings of the 24th Australian Conference on the Mechanics of Structures and
2018;185:273–85. Materials (ACMSM24, Perth, Australia, 6-9 December 2016) 2016 Nov 30 (p. 65).
[18] Farghaly Ahmed Sabry, Benmokrane Brahim. Shear behavior of FRP-reinforced CRC Press.
concrete deep beams without web reinforcement. J Compos Constr 2013:04013015. [41] Mallick PK. Fiber-reinforced composites. New York: Marcel Dekker; 1993.
[19] Rezaei Nazanin, Klein Gary, Garber David B. Effect of development and geometry [42] Godat A, Neale KW, Labossière P. Numerical modeling of FRP shear strengthened
on behavior of concrete deep beams. ACI Struct J 2019;116(3). reinforced concrete beams. J Compos Constr 2007;11(6):640–9.
[20] Drucker DC, Prager W. Soil mechanics and plastic analysis or limit design. Q Appl [43] Smith ST, Gravina RJ. Modeling debonding failure in FRP flexurally strengthened
Math 1952;10(2):157–65. https://doi.org/10.1090/qam/48291. RC members using a local deformation model. J Compos Constr
[21] Ottosen NS. A failure criterion for concrete. Am Soc Civil Eng Eng Mech Div J 2007;11(2):184–91.

799
W.A. Jasim, et al. Structures 26 (2020) 785–800

[44] Obaidat YT, Heyden S, Dahlblom O. The effect of CFRP and CFRP/concrete inter- bond behavior. J Compos Constr 2015;19(1):04014026.
face models when modelling retrofitted RC beams with FEM. Compos Struct [47] Chen GM, Chen JF, Teng JG. On the finite element modelling of RC beams shear-
2010;92(6):1391–8. strengthened with FRP. Constr Build Mater 2012;1(32):13–26.
[45] Tahnat YB, Samaaneh MA, Dwaikat MM, Halahla AM. Simple equations for pre- [48] Kachlakev DI, Miller TH, Potisuk T, Yim SC, Chansawat K. Finite element modeling
dicting the rotational ductility of fiber-reinforced-polymer strengthened reinforced of reinforced concrete structures strengthened with FRP laminates (No. FHWA-OR-
concrete joints. Structures 2020;24:73–86. https://doi.org/10.1016/j.istruc.2020. RD-01-XX). Oregon. Dept. of Transportation. Research Group 2001.
01.010. [49] Kotsovos MD. Finite-element modelling of structural concrete: short-term static and
[46] Tao Y, Chen JF. Concrete damage plasticity model for modeling FRP-to-concrete dynamic loading conditions. CRC Press; 2015.

800