Structures 65 (2024) 106688

Contents lists available at ScienceDirect

Structures
journal homepage: [Link]/locate/structures

Parametric study on the flexural behavior of steel fiber reinforced concrete

beams utilizing nonlinear finite element analysis
Anas M.H. Fares *, 1, Burcu Burak Bakir 2
Department of Civil Engineering, Middle East Technical University, Ankara, Turkey

A R T I C L E I N F O A B S T R A C T

Keywords: Reinforced concrete beams undergo significant internal forces and deformations under extensive loading and
Flexural Behavior have flexural or shear failure based on the reinforcement detailing and geometric properties. The addition of steel
Nonlinear Finite Element Analysis fibers increases the deformation and load carrying capacities that leads to ductile behavior. The main contri­
Steel Fiber Reinforced Concrete (SFRC) Beam
bution of fibers is the improvement of the composite tensile capacity, which results in improved flexural and
Fiber Volume Fraction
Fiber Aspect Ratio
shear responses. This study numerically investigates the effect of fiber properties on the flexural behavior of steel
fiber reinforced concrete (SFRC) beams for both compression and tension failure. First, three specimens from
prior experimental studies are modeled using the nonlinear finite element program ABAQUS and the analytical
results are verified by comparison with the test results. Contrary to prior nonlinear models, the developed model
accurately predicts the damage pattern, descending portion of the load-displacement relationship, and ultimate
displacement, which lead to a reliable estimation of energy dissipation capacity and ductility. A comprehensive
parametric study is then conducted to examine the effect of tension reinforcement ratio, fiber volume fraction,
and fiber aspect ratio on the flexural behavior of both reinforced concrete and SFRC beams. The influence of
these key parameters is investigated by comparing the peak load, displacement ductility, peak stiffness, service
stiffness, energy dissipation capacity, and damage pattern for both reinforced concrete and SFRC beams. The
analytical results indicate that the addition of steel fibers up to 2.0 % leads to a slight increase in the load
carrying capacity, while the displacement ductility and energy dissipation capacity are significantly improved.
Furthermore, the utilization of 1.0 % steel fibers is observed to be sufficient to modify the failure mode of over
reinforced concrete beams from brittle compression failure to tension failure.

1. Introduction Steel fibers are the most commonly used fibers in structural appli­
cations, especially for improving the load carrying and deformation
Concrete is a brittle material with low tensile strength and strain capacities of beams. Several experimental studies have been conducted
capacity, which exhibits a strain-softening response after matrix on the flexural behavior of steel fiber reinforced concrete (SFRC) beams.
cracking. The mechanical properties of concrete are enhanced by the Mertol et al. [4] studied flexural behavior of normal strength reinforced
addition of fibers and a composite material with improved tensile concrete and SFRC beams with 0.98 % hooked-end steel fibers. Experi­
response is obtained. A strain hardening response with uniform distri­ mental results indicated the improved response of SFRC beams under
bution of micro and macro-cracks is observed if more than the critical bending when compared to reinforced concrete ones in terms of flexural
volume fraction of fibers is added to composite [1–3]. Even for toughness or energy dissipation capacity, load carrying capacity, and
strain-softening fiber reinforced composites, tensile strain capacity is service stiffness. However, post-peak stiffness of SFRC beams is observed
higher than that of plain concrete, which indicates an improvement in to be lower than those of reinforced concrete beams. Meda et al. [5] used
ductility. Furthermore, addition of fibers could alter the failure mode of 0.38 % and 0.76 % hooked-end steel fibers and observed that the failure
structural members from a sudden and brittle failure mode to a ductile mode was switched from crushing of concrete to yielding of longitudinal
one. steel bars with increasing fiber volume fraction. Moreover, displacement

* Correspondence to: Middle East Technical University, Department of Civil Engineering, 06800 Ankara, Turkey.
E-mail addresses: [Link]@[Link] (A.M.H. Fares), bburcu@[Link] (B. Burak Bakir).
1
Ph.D. Candidate, Department of Civil Engineering, Middle East Technical University, Ankara, Turkey.
2
Assoc. Prof. Dr., Department of Civil Engineering, Middle East Technical University, Ankara, Turkey.

[Link]
Received 5 February 2024; Received in revised form 17 May 2024; Accepted 31 May 2024
Available online 4 June 2024
2352-0124/© 2024 Institution of Structural Engineers. Published by Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and
similar technologies.
A.M.H. Fares and B. Burak Bakir Structures 65 (2024) 106688

Fig. 1. Specimen layout (All dimensions in mm).

Table 1
Reinforcement detailing of SFRC specimens.
Beam ID Bottom reinforcement Top reinforcement Transverse Reinforcement

SFRC 0.81 2Ø14 2Ø10 Ø8/100 mm

SFRC 2.02 3Ø18 2Ø10 Ø8/100 mm
SFRC 2.13 4Ø16 2Ø10 Ø8/100 mm

Table 2
Mechanical properties of SFRC specimens.
Beam ID Compressive strength (MPa) Modulus of elasticity (MPa) Poisson’s ratio

SFRC 0.81 29.6 25570.8 0.2

SFRC 2.02 25.0 23500.0 0.2
SFRC 2.13 31.8 26503.9 0.2

Table 3
Fiber properties.
Fiber length Equivalent diameter Modulus of elasticity Tensile strength
(mm) (mm) (MPa) (MPa)

30 0.55 210000 1345

Table 4
Damage plasticity model parameters.
Parameter Value

ψ 31◦
e 0.1
fb0 /fc0 1.16
k 0.667
μ 0.0001

Fig. 2. Tensile stress-strain model for SFRC.

ductility, flexural toughness, and bearing capacity were significantly Experimental studies were also conducted on the flexural response of
influenced by the fiber content. Furthermore, service limit state high strength SFRC beams. The effect of concrete compressive strength
response of beams was enhanced by adding fibers due to increased and tension reinforcement ratio on the flexural behavior of both rein­
cracking stiffness. Vandewalle [6] investigated cracking of reinforced forced concrete and SFRC beams were investigated by Ashour et al. [8].
concrete and SFRC beams containing both hooked-end steel and Steel fiber volume fractions were 0.0 %, 0.5 %, and 1.0 % and concrete
glued-in bundles of steel fibers with volume fractions of 0.38 % and 0.56 compressive strengths were 49, 79, and 102 MPa in this test series.
%. Experimental results indicated that the use of fibers resulted in Addition of fibers led to delayed crack propagation and increased
reduced crack spacing and crack width. Campione and Letizia Man­ cracking, yield, and ultimate moment capacities independent of the
giavillano [7] conducted a series of reinforced concrete and SFRC beam tension reinforcement ratio. Flexural rigidity also increased with
tests under both monotonic and cyclic loading. The use of 1.0 % increasing fiber volume fraction and compressive strength. Chunxiang
hooked-end steel fibers was observed to improve beam flexural and and Patnaikuni [9] studied the flexural behavior of high strength SFRC
shear responses. Moreover, bearing capacity and ductility increased in beams composed of 1.0 % enlarged-end mild carbon steel fibers with
SFRC beams with a reduced spalling off of cover concrete under cyclic rectangular cross-sections. Addition of fibers increased the flexural ri­
loading. gidity before yielding, displacement at 80 % of the ultimate load in the

2
A.M.H. Fares and B. Burak Bakir Structures 65 (2024) 106688

descending branch of the load-displacement relationship, and

displacement ductility. Dancygier and Savir [10] investigated the flex­
ural behavior of reinforced concrete and high strength SFRC beams
composed of 0.75 % hooked-end steel fibers with minimum reinforce­
ment. It was observed that the displacement ductility of SFRC beams was
lower than that of reinforced concrete beams, which indicated the need
for a higher minimum tension reinforcement ratio definition for the
SFRC beams when compared to reinforced concrete ones.
The effect of fiber aspect ratio, defined as the ratio of fiber length to
diameter, on the mechanical properties of SFRC has also been studied by
several researchers. Sivanantham et al. [11] tested steel fiber reinforced
composite beams with two different fiber aspect ratios, 64 and 73, and
volume fractions ranging from 0.2 % to 1.0 %. Experimental results
indicated that increase in the aspect ratio had an insignificant effect on
compressive strength, while tensile strength increased due to the
Fig. 3. Stress-strain relationship of steel reinforcement. bridging effect of fibers. Fanella and Naaman [12] concluded that
increasing fiber aspect ratio led to an increase in the slope of the
descending branch in compressive stress-strain relationship. Yoo et al.
[13] reported that medium and long straight fibers provided improved
flexural behavior in terms of deformation and load carrying capacities
and toughness compared to short straight fibers.
Prior experimental studies conducted on a limited number of speci­
mens focused on the effect of fiber volume fraction or beam tension
reinforcement ratio with a constant fiber aspect ratio [4–6]. However,
fiber aspect ratio plays a major role in beam flexural behavior in addi­
tion to varying volume fractions. Therefore, it is vital to investigate the
effect of fiber aspect ratio with different volume fractions on the flexural
behavior of SFRC beams.
For investigating the comparative effect of several parameters, nu­
merical studies are beneficial in terms of the required time to complete
the task and cost efficiency. Structural response can be simulated
Fig. 4. Mesh configuration and boundary conditions. accurately by nonlinear analysis and several parameters can be exam­
ined simultaneously, which may be difficult during experimental studies

Fig. 5. Comparison of experimental and analytical results.

3
A.M.H. Fares and B. Burak Bakir Structures 65 (2024) 106688

Fig. 6. Comparison of damage patterns for SFRC 2.02.

ones. Numbers in specimen designations represent the tension rein­

forcement ratio. Reinforcement detailing of these specimens is provided
in Table 1. As can be observed from this table, the only variable is the
amount of tension reinforcement, while compression and transverse
reinforcements are kept constant as 2Ø10 hanger bars and Ø8/100 mm
stirrups used within the beam except at the constant moment region.

2.2. Materials

2.2.1. Steel Fiber Reinforced Concrete (SFRC)

Mechanical properties of SFRC specimens and fiber properties for
7.7 kg/m3 hooked-end steel fibers, corresponding to 0.98 % fiber vol­
ume fraction, utilized in the experimental study conducted by Mertol
et al. [4] are provided in Tables 2 and 3.
Fig. 7. Compression failure of SFRC 2.02. Concrete damage plasticity (CDP) model is employed to reflect the
[14,15]. Therefore, 3D nonlinear finite element models are built in this complexity of SFRC nonlinear behavior. This model detects stiffness
analytical study using the commercial nonlinear finite element analysis degradation of the material by introducing two damage parameters,
software ABAQUS [16] and accuracy of the models was verified by namely dc and dt for compression and tension damage variables [17].
comparing analytical results with experimental ones. Then, a parametric Values of these parameters are between 0 and 1, which corresponds to
study is carried out to investigate the effect of fiber volume fraction, undamaged and fully damaged, respectively. Based on Sidiroff energy
fiber aspect ratio, and tension reinforcement ratio on the flexural equivalent principle, Demin and Fukang [18] proposed Eq. 1 to obtain
behavior of beams constructed with steel fiber-reinforced composites, the material damage parameters for both compression and tension.
the results of which are presented in this article. √̅̅̅̅̅̅̅
σ
d=1− (1)
E0 ε
2. Finite element modeling
where, σ and ε are the stress and the corresponding strain, and E0 is the
2.1. Beam geometry undamaged elastic modulus.
Five parameters are defined for the CDP model: dilation angle (ψ ),
Experiments conducted by Mertol et al. [4] are adopted in this flow potential eccentricity (e), ratio of initial biaxial compressive stress
( )
analytical study. All the subassemblies were SFRC beams that had 180 × to uniaxial one fb0 /fc0 , ratio of the distances between hydrostatic axis
250 mm cross sections and a span length of 3500 mm tested under and both compression and tension meridians in the deviatoric cross-
positive bending, as depicted in Fig. 1. In this analytical study, three section (k), and viscosity parameter (μ). For plain concrete and SFRC
specimens, namely SFRC 0.81, SFRC 2.02, and SFRC 2.13, are modeled materials, Ozkilic et al. [19] recommended using the parameters given
and analytical results are verified by comparison with experimental in Table 4.

4
A.M.H. Fares and B. Burak Bakir Structures 65 (2024) 106688

Fig. 8. Load-displacement response.

Table 5
Results of the Parametric Study.
Vf P max K SK EDC
Specimen lf /df δmax (mm) δy (mm) δu (mm) μ εt (10-3) FM
(%) (kN) (kN/mm) (kN/mm) (kJ)

S1 0.0 - 81.12 23.97 - 33.80 - 3.38 3.73 2.01 1.41 C

S2 0.5 87.95 25.89 24.47 35.37 1.45 3.40 3.46 2.26 2.27 C
S3 1.0 102.53 32.39 33.00 50.48 1.53 3.17 3.32 3.96 2.65 T
55
S4 1.5 107.39 32.07 27.93 77.15 2.76 3.35 3.34 6.74 3.84 T
S5 2.02 2.0 109.01 31.65 26.84 128.97 4.80 3.44 3.46 11.87 5.16 T
S6 0.5 95.84 30.91 35.57 43.71 1.23 3.10 3.45 3.16 2.60 C
S7 1.0 104.20 34.54 32.25 80.70 2.50 3.02 3.17 6.87 3.14 T
80
S8 1.5 109.58 31.37 26.27 111.75 4.25 3.49 3.54 10.31 5.35 T
S9 2.0 111.66 30.62 25.39 131.53 5.18 3.65 3.68 12.37 5.44 T
S10 0.0 - 55.54 26.75 40.18 52.30 1.30 2.08 2.27 2.31 2.63 T
S11 0.5 56.92 25.89 20.33 49.70 2.44 2.20 2.16 2.25 6.52 T
S12 1.0 58.41 25.35 18.77 84.61 4.51 2.27 2.20 4.11 6.95 T
55
S13 1.5 59.54 25.02 19.33 117.67 6.09 2.38 2.31 6.15 6.33 T
S14 0.90 2.0 60.88 24.38 19.02 156.84 8.25 2.50 2.42 8.57 6.65 T
S15 0.5 57.34 24.10 19.85 62.43 3.14 2.38 2.17 2.96 6.26 T
S16 1.0 59.50 25.22 19.23 109.90 5.71 2.36 2.30 5.67 6.70 T
80
S17 1.5 61.58 23.79 18.82 133.31 7.08 2.59 2.46 7.31 6.25 T
S18 2.0 63.80 24.05 18.62 168.01 9.02 2.65 2.67 9.59 6.32 T

In ABAQUS, SFRC can be modeled as a homogeneous material or modeled as a homogeneous material in this study considering the in­
with discrete and short elements representing fibers distributed fluence of fiber properties and volume fraction on the compressive and
randomly in the matrix. Dispersion and orientation of fibers vary and tensile behavior.
significantly affect the load carrying capacity. Therefore, SFRC is There are numerous constitutive models developed for both

5
A.M.H. Fares and B. Burak Bakir Structures 65 (2024) 106688

Fig. 9. Volume fraction vs. peak load.

Fig. 10. Volume fraction vs. displacement ductility.

compressive and tensile behavior of SFRC. Lok and Pei [20], and Lok and ( )
Xiao [21] concluded that fibers have a negligible effect on composite β εc

compressive strength and recommended using the compressive fc εof

= ( )β (2)
stress-strain relationship of plain concrete for SFRC with a maximum fʹcf
β− 1+ εc
compressive strain of 0.0038. The accuracy of commonly utilized SFRC
εof

compression models developed by Soroushian and Lee [22], Ezeldin and

Balaguru [23], Nataraja et al. [24], and Barros and Figueiras [25] were fcfʹ (MPa) = fcʹ + 2.1604 (RI)W (3)
experimentally investigated by Bencardino et al. [26]. The comparison
indicated that each model agreed well with the experimental data from εof = εco + 6 × 10− 4 (RI)W (4)
which the model was derived, however, accuracy was reduced with
0.7406
additional test data. Abusafaqa [27] stated that Nataraja et al. [24] β = 0.5811 + 1.93 (RI)W − (5)
model has a higher accuracy when compared to other SFRC models.
In the preliminary analytical study, all compressive models lf
(RI)W = Wf , Wf ≅ 3.25 Vf (6)
mentioned previously are utilized and Nataraja et al. [24] model is df
observed to successfully predict the load-displacement relationships
provided in the experimental study. Therefore, Nataraja et al. [24] where, fʹcf is the compressive strength of SFRC, εof is the strain at peak
model, the equations of which are provided through Eqs. 2–6, is adopted compressive stress, fc and εc are the compressive stress and corre­
in the analyses to define the response of SFRC under uniaxial sponding strain values at a point on the curve, fʹc and εco are the peak
compression. compressive stress and corresponding strain values for plain concrete, β
is a material parameter that depends on the shape of the curve, (RI)W is

6
A.M.H. Fares and B. Burak Bakir Structures 65 (2024) 106688

Fig. 11. Volume fraction vs. peak stiffness.

Fig. 12. Volume fraction vs. service stiffness.

fiber reinforcing index by weight, lf , df , and Vf are fiber length, diam­ √̅̅̅̅̅
eter, and volume fraction, respectively. σ cc (MPa) = 0.556 fʹcf (8)
Uniaxial tensile behavior model for SFRC developed by Lok and Pei ( )
[20], which can be utilized for both strain softening and hardening re­ ε0f
εcc = σcc (9)
sponses in post-cracking region, is adopted in this analytical study. Since fʹcf
all the selected specimens have volume fractions less than the critical
value [1–3], only the strain softening response is utilized as presented in 1
( )
lf
Fig. 2 and described through Eqs. 7–11. It is worth mentioning that Lok σ pc = τ d Vf (10)
2 df
and Xiao [21] model is also employed during the preliminary analysis
stage. The difference between the load-displacement relationships ob­ ( )( )
lf τd
tained from each model is insignificant. However, the errors between εpc = (11)
df Ef
the finite element and test results of the SFRC 2.02 specimen for the peak
load and corresponding beam deflection are computed as 5.6 %, and where, σcc is matrix cracking tensile strength, which is assumed to be
√̅̅̅̅̅̅
7.9 % for Lok and Pei [20] model compared to 5.0 %, and 11.6 % for Lok
0.556 fʹcf based on ACI 544.4R-18 [28] recommendations, Ef is steel
and Xiao [21] model.
fiber modulus of elasticity, τd is interfacial bond strength between fiber
σ cc
σ = σ cc − (ε − εcc )2 for ε ≤ εcc (7) and matrix, and ε0f is the compressive strain corresponding to peak
εcc 2
compressive stress, fʹ , of SFRC.
cf

7
A.M.H. Fares and B. Burak Bakir Structures 65 (2024) 106688

Fig. 13. Volume fraction vs. energy dissipation capacity.

Fig. 14. Compression damage for over reinforced beams with short fibers.

For all the examined models developed by Lim et al. [29], Lok and the interfacial bond strength such as fiber aspect ratio and volume
Pei [20], and Lok and Xiao [21], bond strength was assumed to be fraction [32]. Therefore, the interfacial bond strength between fiber and
constant depending only on the shape and type of fibers. There are only a matrix is defined using Yagmur and Burak Bakir equations [32] pro­
limited number of equations that estimate interfacial bond strength for vided as Eqs. 12–13.
different fiber properties. Equations proposed by Voo and Foster and ( )V f
Voo et al. [30,31] consider only the composite compressive strength for lf
τ (MPa) = σ mu (12)
UHPC, which is a major simplification because other factors also affect df

8
A.M.H. Fares and B. Burak Bakir Structures 65 (2024) 106688

Fig. 15. Compression damage for over reinforced beams with long fibers.

supports from the experimental study by Mertol et al. [4]. Translations

are fully restrained along X, Y, and Z axes at the pin side, while restraints
⎧ √̅̅̅̅ are defined in only two directions (X,Y) to represent the roller support.
⎪ 6.7 fʹc (13)
⎨ (psi) Loading is applied through the static analysis method by utilizing a
σmu = √̅̅̅̅ monotonic displacement protocol. To simulate experimental loading
⎪
⎩ 0.556 fʹ (MPa)
c conditions and avoid any localized failure, rigid square steel plates of
50 mm are attached to the beam. The interaction between attached steel
where, σ mu and fʹc are the matrix cracking and compressive strengths, plates and the beam is applied through a tie constraint. Furthermore, the
respectively. interaction between steel cage and the beam is applied using embedded
region for simplicity, which refers to a perfect bond between steel bars
2.2.2. Reinforcement and SFRC with no slip and loss of anchorage. In the experimental study,
Experimentally obtained yield and tensile strengths of both longi­ it was also observed that addition of steel fibers improved the anchorage
tudinal and transverse reinforcement are given as 525 and 640 MPa, between longitudinal steel bars and SFRC.
modulus of elasticity and Poisson’s ratio are also provided as
200,000 MPa and 0.3, respectively [4]. Fig. 3 demonstrates the utilized
2.4. Element type and mesh
engineering stress-strain relationship. However, ABAQUS requires the
true stress-strain relationship, in which the reduction in the cross
Two types of solid elements are adopted in modeling and meshing.
sectional area of reinforcement due to plastic deformations at each time
Concrete and steel plates are modeled by utilizing eight-node linear
step is considered, to define plasticity and accurately obtain the
brick element (C3D8R), whereas 2-node linear 3-D truss element (T3D2)
descending portion of the load-displacement curve. Therefore, engi­
is used for reinforcement. In this study, a mesh size of 20 mm is utilized
neering stress-strain curve is converted to the true one using Eqs. 14–15
since it gives reasonable accuracy in obtaining the load-displacement
[15].
relationship compared to larger mesh sizes utilized in the preliminary
σ T = σ(1 + ε) (14) analysis stage. Fig. 4 illustrates mesh configurations and boundary
conditions of the SFRC beam.
εT = ln(1 + ε) (15)
3. Verification of results
2.3. Boundary conditions, loading, and interactions
Load-displacement relationships obtained from the validation of
Boundary conditions applied in the finite element models are finite element results in comparison with experimental ones are pro­
compatible with the test setup of simply supported beams with pin-roller vided in Fig. 5. Reasonable agreement between numerical results and

9
A.M.H. Fares and B. Burak Bakir Structures 65 (2024) 106688

Fig. 16. Compression damage for under reinforced beams with short fibers.

experimental outcomes can be observed from this figure, which supports reinforcement, which is distinctly captured by analytical damage pat­
the capability of constructed finite element models in predicting the terns. Principal strain contours can also be utilized to monitor the failure
overall response for selected specimens. It is worth mentioning that mode as presented in Fig. 7. The contour provided in this figure is
finite element results yield slightly higher load and stiffness values than illustrated at the time step where the peak load is achieved. Black areas
the test results since perfect bond without any slip is considered between indicate regions in which the principal strain exceeds maximum allowed
reinforcement and SFRC. Interface elements can be utilized to model the compressive strain, which is considered to be 0.0038 based on Lok and
slip of reinforcing bars for improved accuracy, however this will in­ Xiao [21] model.
crease the computational time significantly, which is not suitable for
practical applications. 4. Parametric study
Based on the reinforcement ratios, these beams are classified as over
or under reinforced for which compression and tension failure are After verification of the finite element models, a comprehensive
observed, respectively. ACI 318–19 [33] classifies the sections as parametric study is conducted to examine the effect of tension rein­
compression or tension controlled based on the tension reinforcement forcement ratio, fiber volume fraction, and fiber aspect ratio on beam
strain. If tensile strain is less than the yield strain, which is 0.00263 for behavior. For the tension reinforcement ratio, two different values are
the selected specimens, the section is classified as compression considered, one corresponding to an over reinforced beam with
controlled. If the tension reinforcement strain is more than or equal to ρ= 2.02 % (3Ø18 mm), while ρ= 0.90 % (3Ø12 mm) represents an
the sum of yield strain and ultimate concrete compressive strain, which under reinforced beam for which tension failure is expected. Since the
adds up to 0.00563 for this study, the section is tension controlled. A fiber volume fraction significantly affects the overall response, five
transition zone is considered between these two cases. From finite different values are selected as 0.0 %, 0.5 %, 1.0 %, 1.5 %, and 2.0 %,
element analyses, the tensile strains are obtained as 0.00749, 0.00261, which lead to strain softening response. The 0.0 % volume fraction
and 0.00259 for SFRC 0.81, SFRC 2.02, and SFRC 2.13, respectively. represents reinforced concrete beams with no fibers, which are also
Therefore, SFRC 0.81 has a tension controlled section, while both SFRC investigated to clearly indicate the effect of steel fibers on the behavior.
2.02 and SFRC 2.13 have compression controlled sections. In these models, compressive stress-strain relationship for plain concrete
The comparison of the crack patterns for SFRC 2.02, which demon­ is defined utilizing the modified Popovics model [34], which was veri­
strates the highest correlation with experimental results, is presented in fied by Dere [35], as presented in Eqs. 16–18. The tensile stress-strain
Fig. 6. The damage pattern can be monitored by different methods such relationship for plain concrete is adopted from Dere and Koroglu [36]
as DAMAGET for tension damage, DAMAGEC for compression damage, as described in Eqs. 19–20. Lastly, the fiber aspect ratio is set to 55 for
and PE for plastic strain components. It should be noted that this spec­ short fibers with lf = 30 mm and df = 0.55 mm and 80 for long fibers
imen had a compression failure due to high reinforcement ratio. with lf = 60 mm and df = 0.75 mm to represent existing hooked-end
Therefore, crushing of SFRC occurs before yielding of longitudinal steel fibers commonly used in the construction industry. The cross

10
A.M.H. Fares and B. Burak Bakir Structures 65 (2024) 106688

Fig. 17. Compression damage for under reinforced beams with long fibers.

comparison of these parameters led to construction of 18 finite element 5. Discussion of results

models in which the compressive strength is selected to be 25 MPa, same
as SFRC 2.02. The load-displacement responses obtained from nonlinear finite
( ) element analyses of specimens selected for the parametric study with
n εεcoc different reinforcement ratios, fiber volume fractions, and fiber aspect
fc
= ( )n (16) ratios are illustrated in Fig. 8. From these graphs, key parameters
f ʹc affecting the overall response, such as the peak load, displacement
n − 1 + εεoc
ductility (μ), total energy dissipation capacity (EDC), peak stiffness (K),
√̅̅̅̅ and service stiffness (SK), are extracted and examined. The displacement
εco = 9.37 × 10− 4
fʹc (17)
4
ductility (μ) is computed as the ratio of ultimate deflection to yield
deflection. The ultimate deflection is defined as the deflection corre­
n = 0.058 fʹc + 1.0 (18) sponding to a 15 % reduction in the peak load in the descending portion,
while yield deflection is defined as the deflection corresponding to
where, fʹc (MPa) and εco are the peak compressive stress and corre­ yielding of tension reinforcement [14,15,37]. Total energy dissipation
sponding strain; fc (MPa) and εc are the stress and corresponding strain capacity (EDC) is defined as the area under the load-displacement
at a point on the curve; n is a material parameter that depends on the relationship up to ultimate displacement. The peak stiffness is
shape of the curve. computed as the peak load divided by corresponding displacement. The
service stiffness is considered to be the stiffness of the ascending portion
σ (εt )0.7+1000ε
= (19) between 50 % and 80 % of the peak load [38]. The tension reinforce­
ft ε ment strain (εt ) at the peak load is obtained to validate the failure mode.
Finally, the failure mode (FM) is provided as (T) representing tension
ft
εt = (20) failure, which implies yielding of steel bars before crushing of concrete
Ec
under compression, and (C) representing compression failure for which
where, ft stands for the tensile capacity of plain concrete, which is concrete crushes under compression before yielding of steel bars.
( )2/3 Table 5 summarizes the finite element results of the parametric study.
considered to be ft (MPa) = 0.3 fcʹ based on the study by Ozkilic
et al. [15], while εt is corresponding strain, σ and ε are the stress and
corresponding strain at a point on the curve; and Ec is the modulus of 5.1. Peak load
elasticity of plain concrete.
Fig. 9 demonstrates normalized peak load vs. fiber volume fraction
for over and under reinforced beams with different fiber aspect ratios.

11
A.M.H. Fares and B. Burak Bakir Structures 65 (2024) 106688

Fig. 18. Tension damage for over reinforced beams with short fibers.

Peak load is normalized by dividing the peak load of SFRC specimens compression failure to a ductile tension failure for both long and short
(Pmax) by the peak load of reinforced concrete (Pmax0) control specimen fibers. As expected, the ductility values of under reinforced beams are
to investigate the variation in the peak load, which represents load higher than that of over reinforced ones. Moreover, the displacement
carrying capacity, with increasing fiber volume fraction. It can be ductility increases with increasing volume fraction and aspect ratio of
observed from this figure that using steel fibers significantly increases fibers for both over and under reinforced beams, especially for fiber
the peak load depending on fiber aspect ratio and tension reinforcement volume fractions more than 1.0 %. For long fibers with volume fractions
ratio, for over reinforced beams, whereas slightly enhances the load ranging from 0.5 % to 2.0 %, the ductility values are obtained as 0.85,
carrying capacity for under reinforced beams. For over reinforced 1.64, 1.54, and 1.08 higher than those of short fibers for over reinforced
beams, normalized peak load is between 1.08 and 1.34 for short fibers, beams. For under reinforced beams, the ductility values for long fibers
while this ratio ranges from 1.18 to 1.38 for long fibers. For under are computed as 1.29, 1.27, 1.16, and 1.09 higher than those of short
reinforced beams, normalized peak load is between 1.02 and 1.10 for fibers. The effect of fiber aspect ratio is imperative particularly for over
short and 1.03 and 1.15 for long fibers. As expected, the effect of adding reinforced beams and long fibers should be utilized to achieve a higher
steel fibers to the matrix is more pronounced for over reinforced beams, ductility.
especially for fiber volume fractions higher than 0.5 %.
5.3. Peak stiffness
5.2. Displacement ductility
Peak stiffness is computed by dividing the peak load to corre­
Displacement ductility vs. the fiber volume fraction for over and sponding displacement and the rigidity of beams is investigated based
under reinforced beams, with different fiber aspect ratios are illustrated on this parameter. Fig. 11 presents normalized stiffness vs. fiber volume
in Fig. 10. In this figure, ductility values cannot be normalized, because fraction for different fiber aspect ratios. For over reinforced beams,
for the over reinforced control specimen, which is a reinforced concrete variation in normalized stiffness is negligible for different volume frac­
beam with no fibers, concrete crushes under compression before tions of short fibers. However, for high fiber aspect ratios, normalized
yielding of steel bars, which leads to compression failure. Therefore, the stiffness of the specimens having less than 1.0 % volume fraction is less
behavior is brittle and ductility cannot be obtained. With the addition of than 1.0, which implies that the reinforced concrete control specimen
0.5 % steel fibers, the longitudinal bars yield under tension after has a higher rigidity than SFRC beams. When the volume fraction is
crushing of concrete and before reaching ultimate displacement, which more than 1.0 %, SFRC specimens have a rigid response compared to the
is still a brittle behavior. However, for fiber volume fractions of 1.0 % control specimen. For under reinforced sections, SFRC beams are more
and more, steel bars yield before crushing of concrete, therefore, the rigid than the control specimen for both short and long fibers with
ductility of the beam is enhanced, even for over reinforced beams. The different volume fractions. The range of increase is 1.06 - 1.20 for short
addition of fibers alters the failure mode from sudden and brittle fibers, and 1.15 - 1.28 for long fibers.

12
A.M.H. Fares and B. Burak Bakir Structures 65 (2024) 106688

Fig. 19. Tension damage for over reinforced beams with long fibers.

5.4. Service Stiffness aspect ratio on energy dissipation capacity is not as pronounced as for
over reinforced beams and variation in capacity ranges from 0.97 to 3.71
Service stiffness is defined as the slope of the ascending portion of the for short fibers and 1.28 to 4.16 for long fibers.
load-displacement relationship between 50 % and 80 % of the peak
load, which represents cracked section behavior [4,38]. Fig. 12 depicts 5.6. Damage pattern
normalized service stiffness vs. fiber volume fraction for different fiber
aspect ratios. In over reinforced specimens, increasing volume fraction DAMAGEC and DAMAGET contours at the ultimate displacement are
up to 1.0 % results in decreasing normalized service stiffness for both utilized to monitor damage pattern in all specimens. Figs. 14 and 15
short and long fibers. After that point, service stiffness starts to increase, illustrate DAMAGEC for compression damage with crushing of SFRC
however, normalized values are less than 1.0 for all volume fractions. under compression for over reinforced SFRC specimens for short and
For under reinforced specimens, normalized service stiffness follows a long fibers, respectively, while Figs. 16 and 17 present the same com­
decreasing trend up to 0.5 %, from which point on values start to in­ parison for under reinforced SFRC beams. In these figures, red areas
crease, which is more pronounced for long fibers. Normalized service depict regions in which the principal strain exceeds 0.0038, which is
stiffness reaches 1.17 for long fibers compared to 1.06 for short fibers considered to be the ultimate compressive strain of SFRC [21]. As can be
corresponding to a volume fraction of 2.0 %. observed from these figures, SFRC crushes in between the applied loads
at the top, as expected for positive bending. When fiber volume fraction
5.5. Energy dissipation capacity or aspect ratio increases, the extent of the red area reduces slightly. It is
worth mentioning that utilization of 1.0 % steel fibers in over reinforced
Energy dissipation capacity, which is the area under the load- specimens is observed to be sufficient for altering the failure mode from
displacement curve up to ultimate displacement, is a significant indi­ a sudden and brittle compression failure to a tension failure, where the
cator of beam response under seismic loading. Fig. 13 demonstrates tension reinforcement yields before crushing of SFRC. Moreover, the
normalized energy dissipation capacity vs. fiber volume fractions for failure mode of over reinforced specimens is not influenced by fiber
over and under reinforced specimens. For all SFRC specimens, aspect ratio as presented in Table 5.
increasing fiber volume fractions and aspect ratios lead to an increase in Figs. 18 and 19 depict DAMAGET for tension damage for over rein­
energy dissipation capacity. Particularly for over reinforced beams, forced SFRC specimens for short and long fibers, respectively, while
there is a remarkable effect of fiber aspect ratio on the energy dissipation Figs. 20 and 21 provide the same comparison for under reinforced SFRC
capacity. Long fibers have significantly higher energy dissipation ca­ beams. In these figures, red color represents steel fiber pull out which
pacities ranging from 1.57 to 6.14 of the control specimen when corresponds to the ultimate tensile strain of 0.02 [20] as illustrated in
compared to that of short fibers for the same volume fraction ranging Fig. 2. The damage is concentrated at the midspan in the maximum
from 1.12 to 5.89. For under reinforced specimens, the influence of fiber moment region, which indicates the bridging of cracks by steel fibers.

13
A.M.H. Fares and B. Burak Bakir Structures 65 (2024) 106688

Fig. 20. Tension damage for under reinforced beams with short fibers.

Similar to concrete crushing, increasing both fiber volume fraction and volume fractions, Vf, ranging from 0.5 % to 2.0 % and fiber aspect ratios,
aspect ratio reduces the extent of red areas. Cracks initiate at the bottom, lf /df, of 55 and 80 for hooked steel fibers, the following conclusions are
then propagate towards the applied loads. It is observed that increasing drawn:
both fiber volume fraction and fiber length significantly minimize the
propagation of cracks for both over and under reinforced sections with a 1. Constructed nonlinear finite element models predict the load-
reduction in the areas of high tensile strains. displacement relationships with adequate accuracy. Moreover,
Figs. 22 and 23 demonstrate DAMAGEC for compression damage and damage patterns, crack distributions, compression and tension fail­
DAMAGET for tension damage for over and under reinforced concrete ure zones and principal strain contours are successfully generated,
beams with no fibers (Vf = 0.0 %) at the ultimate displacement, which validate the utilization of the proposed procedure in obtaining
respectively. In Fig. 22, the red areas indicate regions in which the the flexural response of reinforced concrete and fiber reinforced
principal strain exceeds a value of 0.003, which is the ultimate concrete beams with strain softening response.
compressive strain of plain concrete based on ACI 318–19 [33] re­ 2. Not only the fiber volume fraction but also the fiber aspect ratio plays
quirements. The regions where the ultimate tensile strain, which is a major role in the flexural behavior of SFRC beams. Due to enhanced
assumed to be 0.001 as recommended by Dere and Koroglu [36], is fiber bridging provided by longer fibers during fiber pull out, the
exceeded are depicted in red in Fig. 23. As can be observed from these utilization of high fiber aspect ratios leads to improved flexural
figures, the damage intensifies when fibers are not utilized. response.
3. The addition of 1.0 % steel fibers is observed to be sufficient for
6. Conclusions altering the failure modes of over reinforced beams with 2.02 %
tension reinforcement ratios from sudden and brittle compression
The flexural behavior of steel fiber reinforced concrete (SFRC) beams failure to tension failure.
is investigated in this analytical study. First, 3D nonlinear finite element 4. For over reinforced beams, addition of steel fibers to the matrix
models are constructed utilizing the commercial software ABAQUS and significantly increases the peak load up to 38 %, whereas for under
the results are verified by comparison with experimental results of three reinforced beams, there is only a slight improvement.
specimens tested under four-point loading. Then, a total of 18 FE models 5. The displacement ductility is improved significantly by steel fibers
are constructed to conduct a parametric study to investigate the effect of for all beams with any volume fraction over 0.5 %, This improve­
tension reinforcement ratio, fiber volume fraction, and fiber aspect ratio ment is more pronounced when fiber volume fraction is increased
on the flexural behavior. The comparison of peak load, displacement from 1.0 % to 1.5 %. Fiber aspect ratio also influences the ductility.
ductility, peak stiffness, service stiffness, energy dissipation capacity, Over reinforced beams with long fibers, for which volume fractions
and damage pattern provide valuable insight for both reinforced con­ are ranging from 0.5 % to 2.0 %, have 0.85, 1.64, 1.54, and 1.08
crete and SFRC beams. Based on the analytical results obtained for fiber

14
A.M.H. Fares and B. Burak Bakir Structures 65 (2024) 106688

Fig. 21. Tension damage for under reinforced beams with long fibers.

Fig. 22. Damage pattern for over reinforced beam.

higher ductility than those with short fibers, while these ratios are 7. In terms of service stiffness, over reinforced specimens are less rigid
computed as 1.29, 1.27, 1.16, and 1.09 for under reinforced beams. than control specimens for all volume fractions and fiber aspect ra­
6. In terms of peak stiffness, for over reinforced specimens with volume tios. Same trend is observed for under reinforced specimens with
fractions less than 1.0 %, SFRC beams are less rigid than reinforced volume fractions less than 1.0 %, however for higher volume frac­
concrete beams. However, over reinforced SFRC beams with higher tions, SFRC beams have higher service stiffness than reinforced
volumetric ratios exhibit a more rigid response than the control concrete beams.
specimen. Under reinforced SFRC beams are more rigid when 8. As fiber volume fraction and aspect ratio increase, energy dissipation
compared to reinforced concrete beams for all fiber lengths regard­ capacity increases, while tensile strains and the extent of concrete
less of the volume fraction. crushing under compression reduce.

15
A.M.H. Fares and B. Burak Bakir Structures 65 (2024) 106688

Fig. 23. Damage pattern for under reinforced beam.

CRediT authorship contribution statement [13] Yoo DY, Kim S, Park GJ, Park JJ, Kim SW. Effects of fiber shape, aspect ratio, and
volume fraction on flexural behavior of ultra-high-performance fiber-reinforced
cement composites. Compos Struct 2017;174:375–88. [Link]
Anas M. H. Fares: Writing – original draft, Writing – review & compstruct.2017.04.069.
editing, Visualization, Validation, Software, Methodology, Conceptual­ [14] Abu Tahnat YB, Dwaikat MMS, Samaaneh MA. Effect of using CFRP wraps on the
ization, Formal analysis, Investigation, Data curation. Burcu Burak strength and ductility behaviors of exterior reinforced concrete joint. Compos
Struct 2018;201:721–39. [Link]
Bakir: Writing – review & editing, Validation, Supervision, Methodol­ [15] Ozkilic YO, Aksoylu C, Arslan MH. Numerical evaluation of effects of shear span,
ogy, Conceptualization, Formal Analysis, Investigation, Project stirrup spacing and angle of stirrup on reinforced concrete beam behaviour. Struct
administration. Eng Mech 2021;79:309–26. [Link]
[16] ABAQUS. Abaqus/CAE user’s manual. Dassault Syst Simulia Corp 2014:1–1174.
[17] Chi Y, Yu M, Huang L, Xu L. Finite element modeling of steel-polypropylene hybrid
fiber reinforced concrete using modified concrete damaged plasticity. Eng Struct
Declaration of Competing Interest 2017;148:23–35. [Link]
[18] Demin W, Fukang H. Investigation for plastic damage constitutive models of the
The authors declare that they have no known competing financial concrete material. Procedia Eng 2017;210:71–8. [Link]
proeng.2017.11.050.
interests or personal relationships that could have appeared to influence [19] Ozkilic YO, Aksoylu C, Arslan MH. Experimental and numerical investigations of
the work reported in this paper. steel fiber reinforced concrete dapped-end purlins. J Build Eng 2021;36:102119.
[Link]
[20] Lok T-S, Pei J-S. Flexural behavior of steel fiber reinforced concrete. J Mater Civ
References Eng 1998;10:86–97. [Link]
[21] Lok TS, Xiao JR. Flexural strength assessment of steel fiber reinforced concrete.
[1] Naaman AE, Reinhardt HW. Characterization of high performance fiber reinforced J Mater Civ Eng 1999;11:188–96. [Link]
cement composites— HPFRCC. In: Naaman AE, Reinhardt HW, editors. High (1999)11:3(188).
Perform. Fiber Reinf. Cem. Compos., 2. Ann Arbor, USA: CRC Press; 1995. p. 1–24. [22] Soroushian P, Lee C-D. Constitutive modeling of steel fiber reinforced concrete
[Link] under direct tension and compression. In: Swamy RN, Barr B, editors. Fibre Reinf.
[2] Naaman AE, Reinhardt HW. Proposed classification of HPFRC composites based on Cem. Concr. Recent Dev. 1st Editio, Tylor & Francis; 1990. p. 363–77.
their tensile response. Mater Struct 2007;39:547–55. [Link] [23] Ezeldin AS, Balaguru PN. Normal- and high-strength fiber-reinforced concrete
s11527-006-9103-2. under compression. J Mater Civ Eng 1992;4:415–29. [Link]
[3] Naaman AE. Deflection-softening and deflection-hardening FRC composites: (ASCE)0899-1561(1992)4:4(415).
characterization and modeling. Deflection Stiffness Issues FRC Thin Struct. Elem, [24] Nataraja MC, Dhang N, Gupta AP. Stress–strain curves for steel-fiber reinforced
vol. SP-248. American Concrete Institute; 2007. p. 53–66. [Link] concrete under compression. Cem Concr Compos 1999;21:383–90. [Link]
10.14359/19010. org/10.1016/S0958-9465(99)00021-9.
[4] Mertol HC, Baran E, Bello HJ. Flexural behavior of lightly and heavily reinforced [25] Barros JAO, Figueiras JA. Flexural behavior of SFRC: testing and modeling. J Mater
steel fiber concrete beams. Constr Build Mater 2015;98:185–93. [Link] Civ Eng 1999;11:331–9. [Link]
10.1016/[Link].2015.08.032. (331).
[5] Meda A, Minelli F, Plizzari GA. Flexural behaviour of RC beams in fibre reinforced [26] Bencardino F, Rizzuti L, Spadea G, Swamy RN. Stress-strain behavior of steel fiber-
concrete. Compos Part B Eng 2012;43:2930–7. [Link] reinforced concrete in compression. J Mater Civ Eng 2008;20:255–63. [Link]
compositesb.2012.06.003. org/10.1061/(ASCE)0899-1561(2008)20:3(255).
[6] Vandewalle L. Cracking behaviour of concrete beams reinforced with a [27] Abusafaqa FR. Improving ductility behavior of sway-special exterior beam-column
combination of ordinary reinforcement and steel fibers. Mater Struct 2000;33: joint using ultra-high performance fiber-reinforced concrete. An-Najah National
164–70. [Link] University,; 2020.
[7] Campione G, Letizia Mangiavillano M. Fibrous reinforced concrete beams in [28] ACI Committee 544. ACI 544.4R-18: Guide to Design with Fiber-Reinforced
flexure: experimental investigation, analytical modelling and design Concrete. Farmington Hills, Michigan: 2018.
considerations. Eng Struct 2008;30:2970–80. [Link] [29] Lim TY, Paramasivam P, Lee SL. Analytical model for tensile behavior of steel-fiber
engstruct.2008.04.019. concrete. Acids Mater J 1987;84:286–98. [Link]
[8] Ashour SA, Wafa FF, Kamal MI. Effect of the concrete compressive strength and [30] Voo JYL, Foster SJ. Variable engagement model for the design of fibre reinforced
tensile reinforcement ratio on the flexural behavior of fibrous concrete beams. Eng concrete structures. Adv Mater Constr Bridg Build Other Struct III 2003:1–10.
Struct 2000;22:1145–58. [Link] [31] Voo YL, Poon WK, Foster SJ. Shear strength of steel fiber-reinforced ultrahigh-
[9] Chunxiang Q, Patnaikuni I. Properties of high-strength steel fiber-reinforced performance concrete beams without stirrups. J Struct Eng 2010;136:1393–400.
concrete beams in bending. Cem Concr Compos 1999;21:73–81. [Link] [Link]
10.1016/S0958-9465(98)00040-7. [32] Yagmur E. Analytical Modeling of Fiber Reinforced Composite Deep Beams. Middle
[10] Dancygier AN, Savir Z. Flexural behavior of HSFRC with low reinforcement ratios. East Technical University 2018.
Eng Struct 2006;28:1503–12. [Link] [33] ACI Committee 318. 318–19 Building Code Requirements for Structural Concrete
[11] Sivanantham PA, Prabhu GG, Vimal Arokiaraj GG, Sunil K. Effect of fibre aspect- and Commentary. 2019th ed. Farmington Hills, Michigan: American Concrete
ratio on the fresh and strength properties of steel fibre reinforced self-compacting Institute; 2019. [Link]
concrete. Adv Mater Sci Eng 2022;2022:1–8. [Link] [34] Popovics S. A numerical approach to the complete stress-strain curve of concrete.
1207273. Cem Concr Res 1973;3:583–99. [Link]
[12] Fanella DA, Naaman AE. Stress-strain properties of fiber reinforced mortar in
compression. ACI J Proc 1985;82:475–83. [Link]

16
A.M.H. Fares and B. Burak Bakir Structures 65 (2024) 106688

[35] Dere Y. Assessing a retrofitting method for existing RC buildings with low seismic reinforced concrete. Structures 2022;36:979–96. [Link]
capacity in Turkey. J Perform Constr Facil 2017;31. [Link] istruc.2021.12.059.
(asce)cf.1943-5509.0000969. [38] Mertol H. Flexural behavior of reinforced concrete beams with various layers of
[36] Dere Y, Koroglu MA. Nonlinear FE modeling of reinforced concrete. Int J Struct Civ conventional and steel fiber reinforced concrete. Polite Derg 2022;25:273–80.
Eng Res 2017;6:71–4. [Link] [Link]
[37] Abusafaqa FR, Samaaneh MA, Dwaikat MBM. Improving ductility behavior of
sway-special exterior beam-column joint using ultra-high performance fiber-

17