Journal of Building Pathology and Rehabilitation (2022) 7:22

https://doi.org/10.1007/s41024-021-00155-w

RESEARCH ARTICLE

Numerical nonlinear modeling and simulations of high strength

reinforced concrete beams using ANSYS
Pandimani1 · Markandeya Raju Ponnada2 · Yesuratnam Geddada1

Received: 24 November 2021 / Revised: 6 December 2021 / Accepted: 7 December 2021

© The Author(s), under exclusive licence to Springer Nature Switzerland AG 2021

Abstract
This study presents a comprehensive nonlinear material modeling and simulation approach of high-strength concrete (HSC)
beams using the versatile finite element (FE) analysis tool ANSYS. Three reinforced concrete (RC) beams of 102 MPa
strength, comprising three various percentages of tension steel reinforcement are numerically modeled and validated against
the experimentally tested beams available in the literature. The interface bond-slip mechanism between the concrete and
steel reinforcements along with the tension stiffening effects between the cracks is deliberately considered in the developed
FE model. The mesh-sensitivity study is implemented to determine the ideal element density which influences the nonlinear
solution. The load–deflection plots predicted from FEA are corroborated with the experimental curves and the FE models
can able to predict closer response with the experimental curves throughout the loading history with sufficient accuracy.
Subsequently, theoretical analysis based on ACI 318-19 code is executed to verify the cracking, yielding, and ultimate
moments of experimental and numerical models. The ultimate loads and deflections predicted through the FE model show a
maximum discrepancy of − 11.15% and − 3.06%, respectively as compared to the experimental results. The yield and ultimate
loads achieved based on the ACI code exhibit a closer prediction with the experimental data as compared to the numerical
analysis. The graphical contour diagrams of the FE models provide complete and worthy information throughout the beam
length. The developed FE model is a valid and reliable tool to analyze the flexural nonlinear response of high-strength RC
beams and can be further utilized to explore various parametric studies.

Keywords Nonlinear material modeling · Finite element analysis · Theoretical analysis · Stress-contour diagrams · Mesh-
sensitivity analysis · Flexural behavior of HSC beams

1 Introduction defined in the literature. Hence to fill all these gaps, this
study is done to demonstrate the complete nonlinear mod-
The nonlinear behavior of 102 MPa HSC beams using eling approach and the interpretations of the results of HSC
FE analysis software, clear illustrations of the simulation beams using the ANSYS tool and the theoretical compari-
approach, and the output result interpretations are not prop- sons with the predicted FE analysis results. The mesh sensi-
erly illustrated in the past literature. The nonlinear modeling tivity analysis is implemented to ascertain the optimum ele-
approach for various constituent materials of HSC beams ment density which reflects on the ultimate flexural behavior
using different constitutive models and the evolution of of RC beams and the same element density is adopted to
concrete cracking patterns and their causes are not clearly study the symmetry conditions of RC beams which reflect
on the computational effort and time. The achieved optimum
density is later adopted throughout the nonlinear analysis
* Pandimani of all the HSC reinforced beams considered in this study.
Pandimani918@gmail.com
The moment comparisons at critical points achieved through
1
Department of Civil Engineering, Jawaharlal Nehru ACI code provision and the numerical models are verified
Technological University, East Godavari, Kakinada, against the experimental results existing in the literature.
Andhra Pradesh 533003, India The load–deflection plots of the experimental and FE models
2
Department of Civil Engineering, Maharaj Vijayaram exhibit closer agreements with reasonable accuracy.
Gajapathi Raj College of Engineering, Vizianagaram 535005,
Andhra Pradesh, India

13
Vol.:(0123456789)
 22 Page 2 of 23 Journal of Building Pathology and Rehabilitation (2022) 7:22

From the literature study, the following researchers have that the load–deflection curves predicted by the FE models
used the ANSYS numerical software for validation purposes showed reasonable agreement with the experimental data
and conducted various parametric studies. Naser et al. pre- available in the literature. Hawileh et al. created FE mod-
sented a review paper on the FE modeling strategies of RC els to analyze the shear strength of carbon FRP reinforced
beam strengthed with fiber-reinforced polymer (FRP) sheets concrete beams under unsymmetrical loading [12]. They
using ANSYS and ABAQUS software and demonstrated concluded that the numerical outcomes are in good agree-
the techniques related to the complete nonlinear simulation ment with the tested results available in the literature and
strategies [1]. Al-Rousan et al. executed nonlinear FEA of suggested that the FE models are reliable for this purpose.
full-scaled FRP reinforced concrete deck slabs and explored Mustafa et al. performed a numerical simulation on the
parametric study with their validated models and concluded load–deflection response of hybrid steel and FRP reinforce-
that the carbon and glass FRP bars had superior perfor- ment bars [13]. They concluded that the contribution of steel
mance compared to the convention steel bars [2]. Nwankwo bars as top and bottom reinforcement has enhanced the flex-
et al. executed an experimental investigation on RC beams ural ductility and ultimate bearing capacity of RC beams and
strengthened with natural FRP laminate and validated the the FE results show good accuracy with the tested beams.
experimental data with the FE models for ultimate load, Osman et al. examined the shear behavior of RC beams with
crack patterns, and load–deflection plots [3]. Ramesh et al. web openings through experimental testing and FE models.
examined the bending performance of RC beams contain- They identified that the results of numerical models showed
ing hybrid fibers and strengthen with glass FRP laminates reasonable accuracy with the tested beams [14]. Vasudevan
through experimental and numerical simulations [4]. They et al. presented a numerical study on RC beams strengthened
suggested that the ultimate loads predicted by the FE models with external steel bars using batch files in ANSYS and the
showed a discrepancy ranging from 10 to 15% compared load-bearing capacities are validated with the experimental
to the experimental data. Hawileh et al. performed FE data of their previous study [15]. They used COMBIN39
simulation on side bonded FRP laminate RC beams using (spring elements) to model the concrete-steel interface bond.
ANSYS software and by considering the validated models They concluded that the flexural response prediction by the
they extended their study on various factors that affect the numerical models compared well with the experimental data
flexural behavior of strengthened RC beams [5]. Kadhim at all loading points.
et al. executed the FE simulation of basalt FRP strengthened
RC beams and extended their study for different corrosion
rates, basalt FRP wrapping types, and several layers [6]. 1.1 Significance of this study
They conclude that FE models can reasonably predict the
load–deflection response and crack patterns. Ahmed per- This paper demonstrates the numerical nonlinear simula-
formed a numerical simulation on the impact of web opening tions of 102 MPa high strength RC beams under a four-
at the shear-span zone on the shear behavior of RC beams point bending configuration using the robust FE Analysis
using ANSYS FE models [7]. He concluded that the FE tool ANSYS parametric design language (APDL) [16].
models can be used to predict the shear capacity and failure The developed FE model is incorporated with a nonlinear
modes exactly. Choobbor et al. did an experimental inves- concrete stress–strain curve which is a single expression to
tigation on carbon and basalt hybrid FRP strengthened RC define the ascending and descending branch of the curve
beams and the experimental results are validated through FE and is applicable up to 125 MPa strength of concrete [32,
models [8]. They addressed that hybrid FRP strengthening 38, 39]. The developed FE models can precisely capture
techniques are superior compared to the carbon FRP lami- the steel yielding points and can record the crack patterns,
nates alone and stated that the FE models showed a result- stresses, and strains in concrete and steel elements at every
ing discrepancy of less than 12 percent. El-Sayed explored load step. The cracking patterns and the stress contour
through experimental and FE simulation the bending behav- diagram at critical points (cracking, yielding, and ulti-
ior of RC beams incorporated with recycled lath waste as mate loads) predicted through FE analysis provide worthy
steel fiber [9]. He concluded that both studies give simi- knowledge to completely understand the nonlinear flexural
lar responses. Kadhim et al. performed nonlinear numeri- behavior of RC beams in contrast to the expensive experi-
cal simulations on T-beams strengthened with near-surface mental study. The beam symmetry study is also done to
mounted (NSM) carbon FRP bars using ANSYS FE models suggest the importance of geometry and loading symmetry
and did an extended parametric study [10]. They concluded conditions to reduce the computational time. The moment
that the numerical model well predicted the load–deflection of inertias for cracked and uncracked sections calculated
relationship and failure modes. Arafa et al. did FE analysis based on the transformed area method is used to asses the
of RC beams strengthened with various wrapping schemes cracking, service, and yield deflections at mid-span.
of carbon FRP fabric at shear span region [11]. They stated

13
Journal of Building Pathology and Rehabilitation (2022) 7:22 Page 3 of 23 22

2 Nonlinear FE modeling 2.2 Element types employed

This section demonstrates the summary of the experimental Four distinct elements such as SOLID65, LINK180,
RC beams, the distinct element type used for various compo- COMBIN39, and SOLID185 are employed to construct
nents, a comprehensive material constitutive modeling, and and simulate the three-dimensional (3D) HSC reinforced
the nonlinear simulation techniques of developed numerical beams through a graphical user interface (GUI) in APDL
models for high strength RC beams. [3, 17, 18]. The element types, geometry configuration, and
node locations of these elements are displayed in Fig. 2.
The eight-node 3D brick element SOLID65 is employed
2.1 Description of experimental RC beam to develop the concrete model having degrees of freedom
(DOF) at translational x, y, and z-directions for each node
To build and validate the numerical models, three simply [15, 18–22]. The main feature of this element is that it can
supported steel-reinforced HSC beams experimentally tested efficiently simulate the nonlinear tension and compression
by Ashour (2000) are adopted in this study [16]. He tested behavior of concrete beams in three mutually perpendicu-
nine RC beams with three different grades of concrete and lar directions [2, 8, 21–24]. It also supports smeared rein-
an area of steel reinforcements. Among those only three forcements in three orthogonal directions along with large
experimental beams are selected for modeling purposes with deformation, cracking, crushing capabilities [11, 25, 26].
102 MPa high strength concrete and contains three different The tension reinforcements are created using 3D two-node
areas of steel reinforcements which is enough to justify the LINK180 spar elements with three translational x, y, and z
numerical validations. The tested beams had width, depth, DOF at each node [2, 18, 21, 22]. LINK180 elements can be
effective length, and overall length of 200 mm, 250 mm, modeled through a discrete, smeared, and embedded rein-
3080 mm, and 3400 mm respectively. Each beam had tensile forcement approach but in this study, all the reinforcements
steel percentage of pt = 1.18%, 1.78%, and 2.37% respec- are modeled using a discrete method where the steel bars are
tively, comprising 18 mm diameter two, three, and four directly modeled by connecting the two adjacent nodes of
number of tensile steel bars with a constant yield stress of the concrete elements that exist along the axis of the rein-
530 MPa. All three beams had two 6 mm diameter hanger forcements [22, 26–28]. LINK180 is capable of simulating
bars to support two-legged 8 mm diameter transverse stir- the bilinear elastic–plastic behavior of reinforcements along
rups with 150 mm center to center spacing except at the with supporting the swelling, creep, and stress stiffening
maximum moment zone as depicted in Fig. 1. The beams properties [18–22]. The interface bond-slip relationship
are subjected to monotonic static loading under a four-point between the tension bars and the adjacent concrete is simu-
bending configuration until failure. The beams are desig- lated using a dimensionless unidirectional two-node COM-
nated as R2, R3, and R4 based on the number of tension BIN39 spring element having three translational DOF (x, y,
steel bars provided. and z-direction) at each node [15, 18, 29, 30]. This element

Fig. 1  Longitudinal and cross-

section views of experimental
HSC reinforced beam [16]

13
 22 Page 4 of 23 Journal of Building Pathology and Rehabilitation (2022) 7:22

Fig. 2  Element types used, their configuration, and nodal locations [5, 11]

is provided with the nonlinear force–displacement relation- only the ascending portion of the curve up to the maxi-
ship as real constants to simulate the slip action of tension mum compressive strength ­(fc′) is adopted since beyond
reinforcement related to concrete along the longitudinal (x) this stage the ANSYS program diverges the solution and
direction, while they and z directions are constrained by gets terminated due to the crushing of concrete elements
employing the node coupling method [12, 27]. The loading in the compression zone [33]. The tension stress-strain
and steel plates are created using the 3D SOLID185 ele- relationship of concrete in ANSYS is defined based on
ments having eight nodes with three DOF at each node [2, the William and Warnke model which is a tri-linear curve
18]. This element is employed to reduce the stress concen- as shown in Fig. 3b [23, 24, 28, 38, 39]. According to
trations directly under the load and at the support regions as this model, the curve behaves linearly elastic up to the
that of the experimental setup [5, 15]. maximum tensile strength of concrete (­ ft) at a strain value
equal to (εt) [10, 18, 23, 38, 39]. The concrete (SOLID65)
2.3 Nonlinear material constitutive models elements will crack upon reaching ft value and the tensile
stress suddenly drops to 0.6ft due to stress relaxation fol-
The modulus of elasticity (­ E c) and Poisson’s ratio are lowed by a descending branch linearly decrease to zero
needed to define the linear elastic behavior of SOLID65 strength (due to stress-softening effect) at a strain value of
concrete elements [9, 14, 22, 29, 31, 38, 39]. The expres- 6εt as depicted in Fig. 3b [8, 10, 18, 20, 22, 28]. In addi-
sion given in Eq. 7 is employed to define the ­E c value tion to this, the ANSYS program requires open and closed
whereas 0.3 value is assigned for Poisson’s ratio [20, 29, (βt and βc) shear retention coefficients at the crack inter-
31, 38, 39]. The nonlinear stress-strain constitutive law of face [18, 23, 28]. The βt and βc values are ranging from
concrete in compression developed by Thorenfeldt et al. is 0 to 1 numerical values where zero represents a complete
implemented as an inelastic multi-linear isotropic (MISO) loss of shear transfer (smooth cracks) and one represents
stress-strain relationship in this study as shown in Fig. 3a no loss of shear transfer (rough crack) [3, 28, 34]. In this
[29, 32, 38, 39]. This mathematical model is applicable study, a value of 0.3 and 0.5 is adopted for βt and βc based
up to 125 MPa compressive strength of concrete [29, 32]. on the literature review [5, 8, 24, 38, 39].
In this study, the loading control method is employed and

13
Journal of Building Pathology and Rehabilitation (2022) 7:22 Page 5 of 23 22

Fig. 3  Material constitutive laws: a Compressive stress–strain model of concrete, b Tensile stress–strain model of concrete, c Stress–strain
model for steel reinforcements, and d Bond force–displacement model for concrete-steel interfaces

( ) √
εc
�
n ε0
Ec = 3320 fc� + 6200 (7)
fc = fc ( )nk (1)
εc
(n − 1) + ε0
where ­fc is the compressive strength of concrete (MPa),
f­ t is the maximum tensile strength of concrete (MPa), ɛc is
√ the strain at ­fc, ­fc' is the maximum compressive stress of
ft = 0.62 fc� (2) cylinder (MPa), ɛ0 is the corresponding strain at f­ c', n is the
curve fitting factor, k is the a factor which controls curve
where profile and ­Ec is the initial tangent modulus (MPa).
�
The linear elastic behavior of all the steel reinforce-
fc n ments including tension bar, hanger bar, and stirrups are
ε0 = (3)
Ec (n − 1) defined by assigning the modulus of elasticity (­ E s) and
Poisson’s ratio (μs) [20, 22, 28, 38, 39]. A value of ­Es = 200
fc
� GPa and μs = 0.3 is assigned to LINK180 elements [20,
n = 0.8 + (4) 25]. The nonlinear behavior of steel reinforcements is
17
modeled as bilinear elastic perfectly plastic behavior
( ) which is identical in compression and tension with assign-
ε ing yield stress of reinforcements (­ fy = 530 MPa, in this
K = 1, when c < 1or (5)
ε0 study) as suggested by the past researchers in the liter-
ature [19, 22, 38, 39]. The stress–strain relationship of
εc
) �
fc
( )
ε
( steel reinforcements in compression and tension is shown
K = 0.67 + when c > 1 (6) in Fig. 3c. The stress values between the yield point and
62 ε0 ε0
beyond the yielding of steel reinforcements are calculated
using Eqs. 8 and 9 respectively.

13
 22 Page 6 of 23 Journal of Building Pathology and Rehabilitation (2022) 7:22

fs = 𝜀s Es , when 𝜀s < 𝜀y (8) SOLID185 brick element is modeled and fixed over the
concrete (SOLID65) element using the volume sweep
command [3, 18, 28]. This element is provided with only
fs = fy = 𝜀y Es , when 𝜀s ≥ 𝜀y (9) linear material properties similar to that of LINK180 ele-
where ­fs is the stress in steel reinforcements (MPa), ­fy is the ments [34, 38, 39]. The loading is applied through the
yield stress (MPa), εs is the strain corresponding to stress centerline nodes of this element with equally distributed
­fs, εy is the yield strain and E ­ s is the modulus of elasticity loads to each node as shown in Figs. 8, 9 and 10.
of steel (MPa).
The interface bond-slip mechanism between the tensile 2.4 Failure criteria and nonlinear solutions
steel bar and the surrounding concrete is simulated using
the spring element COMBIN39 assigned with the nonlinear ANSYS uses the Newton Raphson iteration equilibrium
force–displacement relationship as shown in Fig. 3d [17, method as default to simulate the nonlinear solution of the
18, 29, 30, 38]. This dimensionless spring element is intro- developed FE models [18, 22, 23, 28]. To avoid the diver-
duced between each coincident node of concrete (SOLID65) gence of a solution, the total applied load is divided into
and steel (LINK180) reinforcements along the longitudinal several load steps and each load step is further subdivided
direction to enable the slipping of the tensile steel bar related into several sub-steps defined by the user [18, 22, 26, 28].
to the concrete volume [12, 15, 29]. The other coincident The stiffness matrix of the developed FE model gets auto-
nodes of reinforcements (hanger bars and stirrups) and con- matically updated at every load increment by incorporating
crete elements are constrained in all directions and converted the changes that occur in the structural stiffness before pro-
into a single entity using the node merging command. The ceeding to the next load steps [18, 22, 28]. Small load step
nonlinear bond stress-slip curve based on the CEB-FIP increments with more numbers of sub-steps are needed at
2010 code is adopted in this study [12, 29, 30, 35, 38, 39]. the first cracking and after the steel yielding point to obtain
The bond force is obtained by multiplying the circumfer- the solution convergence. ANSYS uses two default toler-
ential area (πdL) of the tension steel reinforcement with ance limits for force and displacement criteria to check the
the bond stress (τb) as given in Eq. 14 [29]. The expression convergence of the solution [18, 28, 33]. In this study, the
through Eqs. 10–14 is used to completely define the bond- displacement convergence is kept as default but the force
slip behavior at the interfaces of HSC beams. The value tolerance is altered to 0.1 throughout the analysis because
of τmax = 2.5√fc' (MPa), α = 0.4, ­s1 = 1 mm, ­s2 = 2 mm are the default force tolerance limit creates a convergence issue
taken from CEB-FIP code 2010 in this study [35], whereas after the first cracking and yielding loads due to the high
the value of s­ 3 and s­ 4 is taken equal to 4 mm and 6 mm from nonlinear behavior of concrete elements [23, 33]. In this
the literature [12, 27, 35, 38]. study, the failure of each FE model is identified when the
10 N load step increment does not converge and the asso-
ciated mid-span deflection exceeds the ANSYS program
)𝛼
(10)
(
𝜏b = 𝜏max s∕s1 for 0 ≤ s ≤ s1
deflection limits.
𝜏b = 𝜏max for s1 < s ≤ s2 (11)
2.5 Configuration of the FE models
( )( )
𝜏b = 𝜏max − 𝜏bmax − 𝜏bf s − s2 ∕s3 − s2 for s2 < s ≤ s3 To develop the FE models, initially, lines are drawn through
(12) key points to model the top and bottom reinforcements as
𝜏b = 𝜏bf for s ≥ s3 (13) well as a single-end stirrup without meshing. By using the
copy command the required number of stirrups and spac-
ing can be created. All the created lines are transformed
F𝜏 = 𝜏b × 𝜋dl (14)
into LINK180 truss elements through the meshing attributes
where τb is the bond stress (MPa), s is the slip of reinforce- command. A solid volume of the required length, depth, and
ments (mm), τmax is the maximum bond stress (MPa), s­ 1 width is incorporated over the already modeled LINK180
and ­s2 are the associated respective slip values at maximum element and the solid volume is then discretized using con-
stress and the starting of horizontal plateau, α is the curve crete mesh attributes command with identical mesh size as
fitting parameter, τbf is the residual bond stress (MPa), ­Fτ that of steel reinforcements in all directions such that all
is the bond force (kN), d is the diameter of steel bar, l is the the nodes of steel reinforcements are coinciding with the
length of single LINK180 element along length direction. newly developed concrete nodes and shares identical coor-
To reduce the stress concentrations exactly under dinates. To employ the bond-slip action, only the coinci-
the loading point and support locations, and eight-node dent nodes of tensile steel bars and the respective concrete
nodes are selected separately and the spring elements (with

13
Journal of Building Pathology and Rehabilitation (2022) 7:22 Page 7 of 23 22

force–displacement relationship) are introduced in lon- 3 Finite element analysis and validations
gitudinal (x) direction between each shared node of con-
crete (SOLID65) and steel (LINK180) elements. The other To ascertain the final geometry size and element discre-
coordinates (y and z) of these nodes are constrained using tization of RC beams, a preliminary study is performed to
the node coupling command such that the reinforcement examine the efficient geometry configuration and the opti-
can relatively slip in the longitudinal direction only. The mum mesh density of the selected full-scaled RC beams.
other identical (shared) nodes of the concrete and steel bars The effect of symmetry condition and the mesh sensitivity
(hanger bar and stirrups) are constrained in all directions analysis which reflects on the computational effort, time, as
using the node merging command. The loading and support well as numerical results are studied in detail and the result
plates are modeled with the same mesh size as that of con- outcomes are examined and compared through load–deflec-
crete volume and fixed to the concrete elements using the tion plots.
volume sweep command. The configurations of developed
concrete volume, discretized steel reinforcements, concrete 3.1 Mesh sensitivity analysis
volume discretization, and the bond-slip model are shown
in Fig. 4 respectively. The symmetric and support bound- One of the main aspects of FEA is the selection of optimum
ary condition along with loading patterns of full, half, and mesh density which can be achieved when an increase in ele-
quarter size numerical models are illustrated through Figs. 8, ment density shows an insignificant effect on the results [2,
9 and 10, respectively. 22, 38, 39]. The accuracy of the nonlinear solution depends
on the mesh density and convergence criteria [22, 28]. The
convergence criteria for force and displacement were con-
sidered as 0.1 and 0.05, respectively based on the literature
study [26, 33, 38, 39]. A higher value of 0.1 is used for force
criteria to converge the nonlinear solution [33, 38, 39]. A

Fig. 4  Development of quarter beam FE models: a concrete volume, b reinforcemnts detail, c element discretization, d bond-slip model

13
 22 Page 8 of 23 Journal of Building Pathology and Rehabilitation (2022) 7:22

quarter-size beam model with two symmetric planes is mod-

eled to examine the mesh sensitivity on the load-deflection
relationship and crack patterns of the RC beam up to the
failure stage. The mesh size in height and width direction is
kept constant as 31.25 mm and 25 mm, respectively whereas
three distinct element sizes such as 75 mm, 50 mm, and 25
mm are provided along the length direction. Consequently,
three different mesh densities with 934, 1351, and 2581
elements are developed as shown in Fig. 5. In the mesh-
sensitivity study, the experimental RC beam had 48 MPa
compressive strength and consists of two tension reinforce-
ments of 18 mm diameter along with two 6 mm hanger bars
and all the steel reinforcements had a yield strength of 530
MPa. Table 1 shows the predicted load and deflection values
at critical points for the three developed FE models which
reveal that the FE model with 2581 mesh density can able
to predict the ultimate load and deflection values closer to
the experimental one. The predicted load-deflection plots are
drawn and compared with the experimental curve as shown
in Fig. 6b, which evidence that the FE model with 1351 and
2581 elements are capture a closer response to the experi-
mental one but the model with 2581 predicts the ultimate
deflection of 51.23 which is much closer to the experimental
deflection of 61.10 mm as shown in Fig. 6a. Also, the ulti-
mate crack patterns of the FE models are depicted in Fig. 7,
which suggest that the model with 2581 element displays
well-developed crack patterns compared to the others. Hence
the FE model with 2581 elements is finalized as the required
optimum mesh density throughout the analysis.

3.2 Geometry and loading symmetry study

To minimize the computational duration and effort and to

make use of the advantage of geometry and loading sym-
metry conditions of RC beam a quarter and half symmet-
ric as well full-scale FE models are developed as shown in
Figs. 8, 9 and 10, respectively. Most of the researchers in the
past have been developed only a quarter-size model for the
validation purpose which gives similar results as that of the
full-scale beam [1, 5, 8, 12, 17, 23, 25, 26, 38]. An attempt is
made to completely analyze these models up to failure load
and validate the load–deflection plots with the experimental
curve. The boundary conditions at the symmetry plane and
supports of these FE models are depicted in Figs. 8b, 9b
and 10b respectively. Table 2 presents the predicted FEA
load and deflection values at critical loading points along
Fig. 5  FE models with varying mesh density: a 934 elements, b 1351
with the disk memory consumed for each model. Figure 11a elements, c 2581 elements
depicts the comparison of predicted load–deflection plots
with the experimental one. The predicted cracking, yield and
ultimate moments are compared with the experimental data, load and deflection values at critical loading stages and the
which reveals all the models have achieved similar results load–deflection plots are identical and match well with the
as shown in Fig. 11b. From Table 2 and Fig. 11 it is evident experimental plot. Hence the quarter symmetry FE model
and concluded that all the three models have shown similar which captured a similar flexural response as that of the full

13
Journal of Building Pathology and Rehabilitation (2022) 7:22 Page 9 of 23 22

Table 1  Predicted FEA results from mesh sensitivity analysis

FE analysis results EXP Total number of elements
934 1351 2581

Crack loads, ­Pcr (kN) 12.42 18.07 16.63 16.62

Crack deflections, Δcr (mm) 1.21 1.39 1.28 1.24
­ y (kN)
Yield loads, P 77.89 81.98 81.37 80.08
Yield deflections, Δy (mm) 17.97 15.29 17.89 17.78
Ultimate loads, ­Pu (kN) 90.19 90.83 88.85 88.73
Ultimate deflections, Δu (mm) 61.10 54.57 57.49 61.23

Fig. 7  Ultimate crack patterns of FE models with varying element

densities

are validated through comparison with the experimental

data available in the literature. All the reinforcements in the
beams had a yield stress of 530 MPa. The optimum mesh
density obtained through sensitivity study is adopted for all
the FE models. The boundary condition of the developed
quarter-type FE model is shown in Fig. 8b. The geom-
etry configuration of the developed quarter-size FE model
adopted to simulate the RC beams is depicted in Fig. 12.

3.3.1 Load–deflection response verification

Fig. 6  Mesh sensitivity analysis report: a maximum deflections plot,
b load deflection plot comparison Table 3 and Fig. 13 presents the predicted FEA and experi-
mental moments and deflections result for R2, R3, and R4
beams and their comparison. Figure 13a and b reveal that the
and half size FE model could preferably be the better choice predicted moments and deflections through FEA exhibit rel-
with less analysis time and modeling effort. atively a stiffer response than that of experimental behavior.
The predicted yield moments of R2, R3, and R4 FE models
3.3 Validation of nonlinear FE models show a divergence of − 16.98, 0.09, and − 2.17% respec-
tively with experimental beams. The ultimate moments
Three high strength (102 MPa) RC beams experimentally predicted by R2, R3, and R4 FE models display a devia-
tested by Ashour (2000) are taken to develop the FE models tion of − 9.31, − 11.15, and − 7.33% respectively against
[16]. Instead of a full-size beam, only quarter-size models the experimental beams. The yield and ultimate deflections
are developed having two symmetry planes by considering of R2, R3, and R4 FE modes show a difference of 3.16 and
the advantages of loading and geometry symmetry as shown − 1.16%, 20.02 and − 1.73%, and 14.80 and − 6.43% respec-
in Fig. 12. Three quarter-size models namely R2, R3, and tively with the experimental values. The exact identification
R4 reinforced with 2 or 3 or 4 bars of 18 mm diameter steel of yield points through experimental study is highly difficult
bar respectively are developed, simulated and the results and most often it can be obtained from the load–deflection

13
 22 Page 10 of 23 Journal of Building Pathology and Rehabilitation (2022) 7:22

Fig. 8  Shows developed quarter-symmetry FE models: a element configuration, b boundary condition

Fig. 9  Shows developed half-symmetry FE models: a element configuration, b boundary condition

Fig. 10  Shows developed full-scale FE models: a element configuration, b boundary condition

13
Journal of Building Pathology and Rehabilitation (2022) 7:22 Page 11 of 23 22

Table 2  verification of loads S. no. Critical loading points Experimental ANSYS numerical FE models
and deflections of FE models
with symmetry effect Quarter-beam Half-beam Full-beam

1 Cracking load, P­ cr (kN) 12.42 16.62 17.24 17.18

2 Cracking deflection, Δcr (mm) 1.21 1.24 1.26 1.31
3 Steel yield load, ­Py (kN) 77.89 80.09 82.5 81.98
4 Yield deflection, Δy (mm) 17.97 17.78 16.75 16.72
5 Maximum load, P ­ u (kN) 90.19 88.75 91.17 91.05
6 Maximum deflection, Δu (mm) 61.10 61.23 61.61 61.78
7 Size of FE models (GB) NA 4.44 15.42 31.16

Fig. 12  Geometry configuration of quarter size numerical model for

validation purpose

percentage difference comparisons with the experimental

data. Table 4 also reveals that the ultimate load-bearing
capacity increases with increasing the area of tensile steel
from 509 to 1018 m ­ m2 and as a consequence, the ultimate
deflections are reduced drastically. Overall, the FEA pre-
dicted ultimate loads and deflections value show a maximum
Fig. 11  Comparisons of FE models with symmetry effect: a load–
deflection plot, b moments comparision deviation of − 11.15 and − 3.07% respectively with that of
experimental results which conclude that the simulated FE
model is a reliable tool to perform the numerical simulation
curve. In contrast, the FE models can effectively and pre- of HSC reinforced beams. The predicted load–deflection
cisely capture the load and deflections at the yielding point. plots of R2, R3, and R4 numerical models are compared
Hence large differences in yield deflection were observed with the experimental curve as shown in Fig. 14a–c, which
for R2 and R3 FE models as compared to the experimen- further indicates that the developed FE models captured the
tal values. It is observed that the predicted first cracking flexural response of high strength RC beams closer to the
moments reveal a large discrepancy with the experimental experimental one with reasonable agreement. To demon-
results for all the beams. The predicted cracking moment strate the critical loading points (cracking, yielding, and
of FE models can be obtained based on the assumed val- ultimate) of the load–deflection curves predicted by the FE
ues of ACI code expression and the concrete stress–strain models, the responses captured by the numerical model for
model in tension, but in the experimental beam, there is R2, R3, and R4 is illustrated in Fig. 15, which indicates that
no such assumption. Hence, this could be the reason for the FE can able to predict the crack, yielding and ultimate
large differences in the cracking moments. Table 4 repre- load–deflection response similar to that of the experimental
sents the validations of ultimate load-bearing capacity and plots. Figure 15 also reveals that the percentage of tension
respective deflections predicted through FE analysis and the steel on cracking load has an insignificant effect, yield loads

13
 22 Page 12 of 23 Journal of Building Pathology and Rehabilitation (2022) 7:22

Table 3  Comparisons of moments and deflection at critical points

Beam type Moments (kN-m) Deflections (mm) Ductility
Mcr,EXP Mcr,FEA My,EXP My,FEA Mu,EXP Mu,FEA Δy,EXP Δy,FEA Δu,EXP Δu,FEA μ,EXP μ,FEA

R2 9.18 15.31 48.56 56.81 56.8 62.09 16.77 16.24 62.29 63.01 3.71 3.87
R3 10.35 16.16 78.48 78.55 82.76 91.98 20.52 16.41 49.82 50.68 2.43 3.08
R4 11.82 16.99 100.91 103.1 108.1 116.03 20.94 17.84 37.94 40.38 1.81 2.26

Fig. 13  Validations of FEA results with the experimental: a Critical moments comparison, b yield and ultimate deflections comparisons

Table 4  validations of Beams Ultimate load, P

­ u (kN) % difference Ultimate deflection, Δu % difference
experimental and FEA ultimate (mm)
loads and deflections
EXP FEA EXP FEA

R2 88.06 96.27 − 9.32 62.29 63.02 − 1.16

R3 128.31 142.61 − 11.15 49.82 50.67 − 1.73
R4 167.60 179.89 − 7.33 37.94 39.11 − 3.06

and deflections are increased with increasing steel content, the SOLID65 element exceed the compressive or tensile
whereas the ultimate load has considerably increased but the strength of concrete material, crushing or cracking of con-
ultimate deflections reduced drastically with the increasing crete is developed [18, 26]. Figure 16a, c, e and b, d, f, and
steel percentages. Fig. 17a–c represents the first cracking, cracking at yield
point, and ultimate crack patterns of R2, R3, and R4 FE
3.3.2 Discussion of FEA post‑processing results models respectively. A quarter beam model with a side face
is used for the demonstration of cracking patterns. Flexural
3.3.2.1 Crack patterns The SOLID65 element assumes cracks (first tensile crack) are developed within the maxi-
smeared (distributed) type cracks to model the cracking mum moment zone at the bottom fiber when the principal
in the tension region along with a plasticity algorithm to tensile stresses in x-direction exceed the maximum tensile
simulate the crushing of concrete in the compression region strength of concrete material [18]. Consequently, the crack-
[25, 26, 28]. The ANSYS program can record the crack- ing sign with a circle outline appears perpendicular to the
ing patterns at each load increment. In ANSYS FE mod- principal stresses in the longitudinal (x-axis) direction [28,
els, the cracking and crushing in concrete elements are 36]. Therefore, the crack sign appears as vertical (straight)
represented by a circle outline at each integration point lines formed at the integrated points of concrete (SOLID65)
along with three perpendicular directions. The red, green, elements as shown in Fig. 17a, c, and e. At the shear span
and blue color circle outline indicates the first, second, and location, between the loading and support, both normal
third cracks respectively [17, 18, 28]. When the principal stresses (σ) and shear stresses (τ) develop in the concrete
stresses (tensile or compressive) at the integrated points of elements. Normally the ‘τ’act in the xz plane and ‘σ’ act

13
Journal of Building Pathology and Rehabilitation (2022) 7:22 Page 13 of 23 22

Fig. 15  Demonstration of critical loading points captured by the FE

models

to the tensile principal strains in the y-direction. These are

said to be compressive cracks as shown in Fig. 17 [28, 36].
From the numerical analysis, the first tensile flexural crack
is formed at 23.73kN, 25.05 kN, and 26.34 kN for R2, R3,
and R4 FE models respectively and are concentered at the
bottom fiber within the maximum moment zone as shown
in Fig. 16a, c and e. With a further increment of the applied
load, numerous flexural and diagonal tensile cracks are
developed and propagated towards the loading points. With
higher loads, the yielding of R2, R3, and R4 models occurs
at 88.08 kN, 121.78 kN, and 159.84 kN respectively, and
the respective crack patterns are depicted in Fig. 16b, d, and
f. when the beam reaches its ultimate stage, large diagonal
cracks are appeared along with several compressive cracks
that are formed at the top surface near the loading zone due
to Poisson’s effect, and ultimately the beam is said to fail as
shown in Fig. 17a–c for the respective beams. The ultimate
failure of R2, R3, and R4 FE models occurs at 96.26 kN,
142.60 kN, and 179.89 kN respectively and the respective
cracking pattern is illustrated in Fig. 17.
Fig. 14  Validation of experimental and FEA load–deflection curves
of HSC reinforced beams: a ­Ast = 509 ­mm2, b ­Ast = 763 ­mm2, and c
3.3.2.2 Stresses analysis in concrete and steel reinforce‑
­Ast = 1018 ­mm2 ments The predicted FEA stresses in concrete and steel
reinforcements are illustrated in Table 5. At yielding
and ultimate loads the longitudinal (axial) compressive
in the x-direction. Consequently, the principal tensile stress stresses (at top fiber) in concrete at the mid-span of the
(σ1) direction becomes inclined from horizontal. At the beam are increased from 56.87 to 82.22 MPa and 102.8
integrated point of the concrete elements, once the tensile to 111.07 MPa, respectively with increasing the p­ t content
principal stresses surpass the maximum tensile strength of from 1.18 to 2.37% as shown in Table 5. In contrast, the
concrete ­(ft) inclined circles appear as straight lines perpen- tensile stresses (bottom fiber) in concrete at mid-span at
dicular to the principal stresses direction. These are referred yield and ultimate load reduces from 7.19 to 4.01 MPa and
to as diagonal tensile cracks as shown in Figs. 16 and 17 6.39 to 4.05 MPa respectively with increasing pt content.
[28, 36]. As the loading of the FE model is applied in the Also, Table 5 reveals that the axial stresses in steel rein-
width (z-axis) direction which develops tensile strain in the forcements are more or less similar at cracking load but at
y-direction due to Poisson’s effect. Hence, at the integrated yielding of tension reinforcement the stresses in compres-
points of concrete elements near the loading location (top sion steel (hanger bar) exhibit an increment from 100.61
compression zone) the circle outline appears perpendicular to 178.23 MPa as the pt content increases. Figures 18a,

13
 22 Page 14 of 23 Journal of Building Pathology and Rehabilitation (2022) 7:22

Fig. 16  Concrete crack patterns at cracking and yield load: a, b for R2 FE model, c, d for R3 FE model, and e, f for R4 FE model

19a, and 20a represent the axial stress contour plot at first 4 Theoretical analysis
cracking (modulus of rupture) of concrete, which shows
that tensile stress softening occurs at the formation of According to the ACI:318–19 code the RC members are
a flexural crack at the tension fiber within in maximum classified as tensioned controlled, transition and compres-
moment zone. Figures 18b, 19b, and 20b show the princi- sion controlled sections based on the strain limit in the
pal compressive stress (S3) diagrams at yield load which extreme tension reinforcements (εt) at the nominal strength
reveal that the compressive stresses are maximum near [37]. When the strain in extreme tensile reinforcement is
the loading location. It is also evident from Table 5 and greater than or equal to 0.005 (εt ≥ 0.003 + εy), the sec-
Figs. 18b, 19b, and 20b that the compressive stresses (Sx) tion is said to be tension controlled. Where, εy = yield
in concrete at yielding will be equal to the third principle strain which is taken as 0.002 as per ACI 318–19 code.
(S3) stress. The yield stress contour plots for steel rein- The section is said to be compression controlled when
forcements are depicted through Figs. 18c, 19c, and 20c the strain in tensile reinforcement is less than or equal to
for R2, R3, and R4 beams which show that the yield load 0.002 (εt ≤ εy). If the strain limit lies between the tension
increases with increasing steel content, and the FE mod- and compression controlled section (0.002 ≤ εt ≤ 0.005),
els accurately capture the yielding (fy = 530 MPa) of steel then it is called transition controlled. The limiting strain
reinforcements for all the cases. Figures 18d, 19d, and 20d distribution diagram for the tension and compression con-
show the ultimate deflection contour plots which reveal trolled section is shown in Fig. SM7. The beams (R2, R3,
that increasing the tensile steel content drastically reduces and R4) considered in this study are tension-controlled in
the ultimate deflection. The vertical mid-span deflec- which ductile type failure prevails where the yielding of
tions at first cracking and yield load for all the three types tensile steel reinforcement occurs first before the crushing
of high strength RC beams are depicted in Fig. SM1 to of concrete. Figure 21 illustrates the stress–strain distri-
SM3 which is evidence that the deflection at cracking and bution profile and the internal forces acting over the RC
yielding increases with increasing tension steel area. The beam for tension controlled section as per ACI 318–19
axial (Sx) and von-mises stress (SEQV) contour diagrams code provision.
at the ultimate failure of R2, R3, and R4 beams are shown The following are the expressions used to analyze the
through Fig. SM4 to SM6, which reveals that the com- flexural behavior of the tensioned controlled high-strength
pressive stresses at failure is maximum and concentrated RC beam section based on the ACI 318–19 code provision.
near the applied loading locations.

13
Journal of Building Pathology and Rehabilitation (2022) 7:22 Page 15 of 23 22

Fig. 17  Shows the ultimate

cracking patterns of concrete: a
for the R2 FE model, b for the
R3 FE model, and c for the R4
FE model

Table 5  Numerical stress FEA model Axial stresses in concrete at mid-span (MPa) Reinforcement axial stresses
analysis of concrete and steel (MPa)
reinforcements at critical
loadings Cracking Yielding Ultimate Cracking Yielding
fc,top fc,bottom fc,top fc,bottom fc,top fc,bottom σst σsc σst σsc

R2 9.28 6.27 57.87 7.19 102.8 6.39 23.96 24.02 530 100.61
R3 10.89 6.65 68.16 4.58 110.42 5.61 23.92 25.03 530 138.37
R4 10.40 5.66 82.22 4.01 111.07 4.05 23.97 26.01 530 178.23

Depth of rectangular stress block (a) = fs × As ∕ 0.85 × fc� × b The minimum area of tension reinforcement ­(As,min)
(15) �√ � � � �
As,min = 0.25 × fc ∕fy × b × d or 1.4∕fy
Neutral axis depth (C) = a∕𝛽1 (16) (19)
× b × d (whichever is higher)
�
Stress intensity factor 𝛽1 = 0.65 for fc ≥ 56 MPa (17) �
(20)
√
Modulus of rupture stress ft = 0.62 × fc (MPa)
( )
Strain limit in tension reinforcement 𝜀t = 0.003 × (d − C)∕C
(21)
( )
Cracking moment Mcr = ft × It ∕yt
(18)

13
 22 Page 16 of 23 Journal of Building Pathology and Rehabilitation (2022) 7:22

Fig. 18  FEA post-processing results of R2 beam model: a Axial stresses (­ Sx) in concrete at first crack, b principal stresses in concrete at yielding
(S3), c Axial stresses (σx) in reinforcements at yielding, and d ultimate deflection contour plot

Fig. 19  FEA post-processing results of R3 beam model: a Axial stresses (­ Sx) in concrete at first crack, b principal stresses in concrete at yielding
(S3), c Axial stresses (σx) in reinforcements at yielding, and d ultimate deflection contour plot

4.1 Theoretical deflection calculations

( )
Nominal moment of resistance Mn = fs × As × (d − 0.5 × a)
(22)
( ) To determine the mid-span deflections of RC beams sub-
Design moment Md = 𝜙 × Mn , jected to four-point bending configuration, the traditional
(23)
(𝜙 = 0.9 for Tension controlled section) elastic bending theory is utilized to obtain the theoretical
deflection values at first cracking, serviceability, and steel

13
Journal of Building Pathology and Rehabilitation (2022) 7:22 Page 17 of 23 22

Fig. 20  FEA post-processing results of R4 beam model: a Axial stresses ­(Sx) in concrete at first crack, b principal stresses (S3) at yielding in
concrete, c Axial stresses (σx) in reinforcements at yielding, and d ultimate deflection contour plot

Fig. 21  Stress–strain distribution diagram for tensioned-controlled RC beam section as per ACI:318–19 code provision

Table 6  Theoretical calculations of the moment of inertia (I) and neutral axis depth (C)
Beam type Moment of inertia (I) Neutral axis depth Lever-arm (mm) Tension rebar strain Internal forces kN
× ­106 ­mm4 (C) mm
Ig Itr Icr Ieff Ctr Ccr Cu Z = d-0.5a εt Fi

R2 260.42 273.35 67.11 71.17 129.44 67.65 23.84 207.25 0.02405 269.77
R3 260.42 279.58 92.39 94.56 131.51 79.47 35.74 203.38 0.01504 404.39
R4 260.42 285.69 114.77 116.11 133.47 88.61 47.68 199.50 0.01052 539.54

Where, ­Ig = ­Itr = ­Icr = ­Ieff = gross, transformed, cracked and effective moment of inertia, ­Ctr = ­Ccr = ­Cu = transformed, cracked and ultimate depth of
neutral axis, εt = strain at failure in extreme tensile steel bar, F ­ i = internal forces in concrete or tension steel at ultimate failure.

yielding points. The theoretical moment of inertias and neu- sections using the transformed area method is illustrated in
tral axis depth calculated based on cracked and uncracked Table 6. It is evident from Table 6 that the transformed,

13
 22 Page 18 of 23 Journal of Building Pathology and Rehabilitation (2022) 7:22

cracked, and effective moment of inertias (­ Itr, ­Icr, and I­ eff ( )2

b × C2t ∕2 = b × D − Ct ∕2 + (m − 1) × As × d − Ct
( )
respectively) are increased with an increasing tension-
(25)
steel percentage. Also, the transformed, cracked, and ulti-
mate depth of the neutral axis (­ Ct, ­Ccr, and C ­ u respectively) Moment of inertia of the uncracked section
increases with the increasing tensile steel content. The )3 )2
It = b × C3t ∕3 + b × D − Ct ∕3 + (m − 1) × As × d − Ct
( (
increasing pt content lowers (increase) the ­Cu value which in
(26)
turn declines the ultimate deflection and displacement duc-
tility ultimately as shown in Table 6. Table 6 shows the strain Cracked depth of neutral axis
values in tensile reinforcements of R2, R3, and R4 beams
b × C2cr ∕2 = m × As × d − Ccr (27)
( )
at nominal strength are higher than the strain limit of 0.005,
which reveals that all the beams are tension-controlled sec- Cracked moment of inertia
tions. Table 6 also indicates that the lever arm (Z) and the
internal resisting forces ­(Fi) of the RC beams are increased
)2
Icr = b × C3cr ∕3 + m × As × d − Ccr (28)
(
with the increasing tensile reinforcement percentage. The
theoretical downward deflections at mid-span for cracking, Effective moment of inertia ­(Ieff)
serviceability, and yielding stage are calculated using the ( )2 ( )))
(29)
( (
expression based on the elastic bending equation as given Ieff = Icr ∕ 1 − 0.67Mcr ∕Ma × 1 − Icr ∕Ig
in Eq. 24 [16]. In this equation, the value of the moment of
inertia is taken as I­ t, ­Ieff, and I­ cr for determining the cracking, Gross moment of inertia ­(Ig)
serviceability, and yield load deflections respectively. The
transformed area method is used to calculate the transformed Ig = BD3 ∕12 (30)
and cracked moment of inertia as shown in Fig. SM8 and
given through Eqs. 25–28 [34]. For the smooth transition
4.2 Results comparison between experimental,
from the maximum moment of inertia ­(Ig) to the minimum
numerical, and theoretical analysis
moment of inertia ­(Icr), the ACI code presents an expression
for effective moment of inertia ­(Ieff) for the calculation of
This section deals with the comparisons between the FEA
deflection at serviceability limit as given in Eq. 29, such that
predicted, experimental and theoretical results, and the
­Icr ≤ ­Ieff < ­Ig [16, 32, 37].
salient points are discussed. The predicted numerical and
Deflection, Δ = Ma × 3 × l2 − 4 × a2v ∕ 24 × Ec × I
( ) ( ) theoretical moments at first cracking show large variation
(24) with the experimental findings since it is highly difficult to
exactly denote the first cracking through the experimental
where ­Ma is the applied moment, l is the effective length, a­ v
investigation as shown in Table 7. On the other hand, the
is the shear span, E
­ c is the elstic modulus of concre, and I is
theoretical and numerical cracking moments are well corre-
the respective moment of inertia.
lated with a maximum deviation of 11%, which justifies that
Uncracked neutral axis depth ­(Ct)
the responses captured by the FE models are more relevant

Table 7  Comparisons of Beam types Cracking moments, M

­ cr Yield moments, M
­ y Ultimate moments, M
­ u
experimental, numerical, and (kN-m) (kN-m) (kN-m)
theoretical moments at critical
points EXP FEA ACI EXP FEA ACI EXP FEA ACI

R2 9.18 15.31 14.23 48.56 56.81 51.92 56.8 62.09 55.91

R3 10.35 16.16 14.80 78.48 78.55 76.23 82.76 91.98 82.25
R4 11.82 16.99 15.38 100.91 103.1 100.06 108.1 116.03 107.64

Table 8  The yield and ultimate Beam types Yield load capacity, ­Py (kN) Ultimate load capacity, ­Pu (kN)
loads comparison between
experimental, FEA, and EXP FEA % Diff ACI % Diff EXP FEA % Diff ACI % Diff
theoretical results
R2 75.29 88.06 − 16.97 80.49 − 6.91 88.06 96.26 − 9.31 86.68 1.57
R3 121.67 121.78 − 0.09 118.19 2.86 128.31 142.61 − 11.13 127.52 0.62
R4 156.44 159.83 − 2.17 155.13 0.84 167.6 179.89 − 7.33 166.88 0.43

13
Journal of Building Pathology and Rehabilitation (2022) 7:22 Page 19 of 23 22

than the experimental results at the cracking stage. The yield expressions predict closer response as compared to the FE
loads and moments obtained from FEA show less than − 3% analysis. The developed FE model reasonably predicts the
difference with the experimental result except for the R2 ultimate load-bearing capacity of high-strength RC beams
beam that shows − 16.97%, whereas the ultimate (load and with less than 12% disparity against both experimental
moment) capacity shows less than 12% deviation with the and ACI code provisions. Figure 22a–c depicts the graphi-
experimental values as shown in Tables 7 and 8. The the- cal comparison of moments at critical points between the
oretical yield and ultimate capacity (loads and moments) experimental, FEA, and ACI, which suggests that the FEA
based on the ACI code show a maximum discrepancy of and ACI predicted yield and ultimate moments are in good
− 6.97% and 1.57% respectively with the experimental agreement with the experimental data. Figure 22d demon-
results as shown in Tables 7 and 8, which reveals that the strates that the cracking, yielding, and ultimate loads pre-
theoretical calculations based on ACI code provision give dicted by the developed FE models show a closer agreement
reliable results. with the ACI code provisions.
Table 9 presents the comparison of the results between
the numerical and theoretical load-bearing capacities at the 4.3 Limitation and recommendations
crack, yield, and ultimate stages. The predicted results of
the FE model show a maximum deviation of less than 12 The limitation of the developed FE model is that the post-
percent against the theoretical results, which suggests that peak behavior (descending branch) of the load-deflection
the predicted results are reliable and within reasonable lim- curve can not be assessed with the loading control method.
its. Table 10 illustrates the comparison of theoretical and Non-convergence of the solution occurs at the first crack and
numerical predicted deflections at cracking, service, and beyond the yielding loads due to the high nonlinearity of
yielding stages. The serviceability deflection is obtained by concrete material. The proposed models can be directly used
dividing the ultimate load by a load factor of 1.6 for both to investigate the various intensive parameters that influence
the cases (EXP and FEA) because as per the ACI 318-19 the flexural behavior of HSC beams. This study can be fur-
code the service live load is multiplied by a factor of 1.6 to ther extended to do a cyclic load analysis, thermal analysis,
convert into a limit sate. The theoretical cracking deflection steel reinforcement corrosion study, and dynamic analysis.
shows a large variation with the numerical value because
the FE model is assigned with different constitutive mod-
els (like concrete stress–strain model, bond-slip model), 5 Conclusions
stress relaxation phenomenon, and large load increments
at the cracking stage. Table 10 also reveals that the crack- This study aims to present an exhaustive nonlinear FE
ing, service, and yield deflections of FEA show a maximum analysis of HSC beams using the ANSYS FE analysis
deviation of − 20.31, − 12.27, and − 9.87% respectively with package. A wide range of graphical displays of contour
the theoretical results. Overall the ultimate moment carry- diagrams provides a good understanding of the nonlin-
ing capacity predicted based on FEA and ACI code show a ear behavior of RC beams and provides worthy informa-
maximum discrepancy of − 11.13 and − 1.57% respectively tion about the flexural behavior of beams up to the failure
with the experimental data, which reveals that the ACI code stage. Theoretical calculations using ACI code provision,

Table 9  Results comparison Beam type Cracking loads, P

­ cr (kN) Yielding loads, P
­ y (kN) Ultimate loads, ­Pu (kN)
between FEA and theoretical
loads at critical points ACI FEA % Diff ACI FEA % Diff ACI FEA % Diff

R2 22.06 23.73 − 7.57 80.49 88.07 − 9.42 86.68 96.26 − 11.05

R3 22.95 25.06 − 9.19 118.19 121.78 − 3.03 127.52 128.32 − 0.62
R4 23.84 26.34 − 10.47 155.13 159.84 − 3.03 166.88 167.60 − 0.43

Table 10  Comparisons of FEA Beam type Cracking deflections, Serviceability deflection, Yielding deflection,
and theoretical deflections at the Δcr (mm) Δs (mm) Δy (mm)
crack, service, and yield points
FEA ACI % Diff FEA ACI % Diff FEA ACI % Diff

R2 1.19 0.994 − 19.71 10.61 9.68 − 9.60 16.24 14.78 − 9.87

R3 1.21 1.011 − 19.68 11.98 10.67 − 12.27 16.41 15.76 − 4.12
R4 1.29 1.028 − 20.31 12.54 11.37 − 10.29 17.85 16.65 − 7.20

13
 22 Page 20 of 23 Journal of Building Pathology and Rehabilitation (2022) 7:22

Fig. 22  Results comparisons between experimental, FEA, and ACI: a cracking moments, b yield moments, c ultimate moments, and d load-
carrying capacity at critical points (ACI Vs FEA)

estimation of the cracked and uncracked moment of iner- • Overall the ultimate load carrying capacity and the asso-
tias based on transformed area method, and the assess- ciated central-span deflection achieved through FE analy-
ment of mid-span deflections using elastic bending theory sis show a disparity of less than 12 and 4% respectively
provides closer agreement with the numerical and experi- against the experimental findings.
mental results. Through this investigation, the following • The developed numerical models well captured the
conclusions can be produced. evolution of concrete crack patterns, load–deflection
responses, and exact prediction of steel reinforcement
• The developed ANSYS numerical models can able to yielding points which is evidence that the ANSYS FE
predict the complete flexural response of 102 MPa HSC model is a reliable tool to validate and explore the para-
beams with sufficient accuracy. metric studies of high strength RC beams.
• The developed FE models reasonably predicted the load-
bearing capacity and the associated mid-span vertical
deflection at critical loading points and the predicted Appendix
load–deflection plot clearly indicates the three-stage
profile similar to the experimental curve and correlates Theoretical analysis of tension controlled high
well with each other. strength RC beam using American concrete institute
• The developed numerical models efficiently captured the (ACI 318‑19) code provisions
stress distribution in concrete and steel reinforcements
until failure and can able to record and display the crack- Moments and forces calculations of high strength RC (R2)
ing patterns at each load increment, which demonstrate beam
the practicability of ANSY numerical simulations.
• The theoretical mid-span deflections obtained using the At the first tensile crack The modulus of elasticity of con-
elastic bending theory at cracking, service, and yielding crete ­(Ec)
stages are well correlated with the experimental and FE
results.

13
Journal of Building Pathology and Rehabilitation (2022) 7:22 Page 21 of 23 22

(( � �
Ec = 4700 f�c = 4700 × 102.40 = 47.56 × 103 MPa
√ √ ) )
𝛽1 = 0.85 − 0.005 fc − 28 ∕7 for 28 < fc < 56 MPa.

The modulus of rupture ­(ft) ( � )

𝛽1 = 0.65 since fc = 102.40 MPa ≥ 56 MPa
ft = 0.62 f�c = 0.62 × 102.40 = 6.27 MPa
√ √
The depth of stress block (a)
The cracking moment ­(Mcr)
a = fy As ∕0.85f�c b
6 6
Mcr = ft It ∕yt = 6.27 × 273.34 × 10 ∕120.57 = 14.21 × 10 N − mm
a = 530 × 509∕0.85 × 102.40 × 200 = 15.50 mm
The cracking load
The depth of neutral axis (C)
Pcr = 2Mcr ∕av = 2 × 9.79 × 106 ∕1290 = 22.03 × 103 N C = a∕ 𝛽1 = 15.50∕0.65 = 23.85 mm

The tensile strain in reinforcement (εt)

At yielding of tension steel reinforcement The minimum area
of tension reinforcement (­ As,min) 𝜀t = 𝜀cu (d − C)∕ C = 0.003 × (215 − 23.85)∕23.85
�√ � � � � = 0.0240 > 0.005
As,min = 0.25 fc ∕fy bw d or 1.4∕fy bw d (whichever is greater)
Hence the section is tensioned controlled.
The nominal moment at failure ­(Mn)
102.40∕530 × 200 × 215 = 205.25 mm2
�√ �
As,min = 0.25 ×
Mn = fy As (d − 0.5a)
or
= 530 × 509(215 − 0.5 × 15.50) = 55.91 × 106 N − mm
As,min = (1.4∕530) × 200 × 215 = 113.58 mm2
The ultimate load at failure
2
Hence as per ACI 318–19, ­As,min = 205.25 ­mm .
Pu = 2 × 55.91 × 106 ∕1290 = 86.68 × 103 N
Therefore, ­(As)provided = 509 ­mm2 > ­As,min = 205.25 ­mm2.
The yielding moment ­(My).
By assuming the linear stress-distribution at yielding load Design moment at failure Strength reduction factor (ϕ), As
instead of actual parabola profile per ACI 31-19
( )
My = fy Ast d − xcr ∕3 Φ = 0.9, for tension controlled section
(( ) )
= 530 × 509 × (215 − (67.65∕3) = 51.92 × 10 N − mm 6 Φ = 0.65 + 0.25 𝜀t ∕𝜀ty ∕0.003 , for transition controlled section
Φ = 0.65, compression controlled section
The yield load ­(Py)
The design moment ­(Md)
Py = 2My ∕av = 2 × 51.92 × 106 ∕1290 = 80.49 × 103 N
Md = 𝜙Mn = 0.9 × 55.91 × 106 = 50.32 × 106 N − mm

At ultimate failure The stress block intensity factor (β1) The design load at failure ­(Pd)
�
𝛽1 = 0.85 for 18 ≤ fc ≤ 28 MPa, Pd = 2 × 50.32 × 106 ∕1290 = 78.01 × 103 N

Deflection calculations of high strength RC (R2) beam

�
𝛽1 = 0.65 for fc ≥ 56 MPa,

and At first cracking load The cracking deflection (Δcr)

Δcr = Mcr ∕24 Ec It × 3l2 − 4a2

( ) ( )

Δcr = (14.21 × 106 ∕24 × 47.56 × 103 × 273.34 × 106 ) × 3 × 30802 − 4 × 12902
( )

Δcr = 0.994 mm

At service load The service deflection (Δser)

13
 22 Page 22 of 23 Journal of Building Pathology and Rehabilitation (2022) 7:22

Mser = Mn ∕1.6 = 55.91 × 106 ∕1.6 Compos Part B Eng 173:106952. https://​d oi.​o rg/​1 0.​1 016/j.​
compo​sitesb.​2019.​106952
= 34.94 × 106 N − mm 6. Kadhim AMH, Numan HA, Özakça M (2019) Flexural strength-
ening and rehabilitation of reinforced concrete beam using
= Mser ∕24 Ec Ieff × 3l2 − 4a2
( ) ( )
Δser BFRP composites: finite element approach. Adv Civ Eng.
Δser = (34.94 × 106 ∕24 × 47.56 × 103 × 70.36 × 106 ) https://​doi.​org/​10.​1155/​2019/​49817​50
7. Sayed AM (2019) Numerical study using FE simulation on rectan-
× 3 × 30802 − 4 × 12902 gular RC beams with vertical circular web openings in the shear
( )
zones. Eng Struct 198:109471. https://d​ oi.o​ rg/1​ 0.1​ 016/j.e​ ngstr​ uct.​
Δser = 9.49 mm 2019.​109471
8. Choobbor SS, Hawileh RA, Abu-Obeidah A, Abdalla JA (2019)
Performance of hybrid carbon and basalt FRP sheets in strength-
At yielding load The yield deflection (Δy) ening concrete beams in flexure. Compos Struct 227:111337.
https://​doi.​org/​10.​1016/j.​comps​truct.​2019.​111337
Δy = My ∕24 Ec Icr × 3l2 − 4a2 9. El-Sayed TA (2019) Flexural behavior of RC beams containing
( ) ( )
recycled industrial wastes as steel fibers. Constr Build Mater
Δcr = (51.92 × 106 ∕24 × 47.56 × 103 × 67.11 × 106 ) 212:27–38. https://​doi.​org/​10.​1016/j.​conbu​ildmat.​2019.​03.​311
× 3 × 30802 − 4 × 12902
( ) 10. Kadhim MMA, Adheem AH, Jawdhari AR (2019) Nonlinear finite
element modelling and parametric analysis of shear strengthening
Δcr = 14.78 mm RC T-beams with NSM CFRP technique. Int J Civ Eng 17:1295–
1306. https://​doi.​org/​10.​1007/​s40999-​018-​0387-8
11. Arafa M, Alqedra M, Hammad M (2018) Effect of fiber orienta-
tion on shear behavior of RC beams externally strengthened with
Supplementary Information The online version contains supplemen- CFRP. Int J Compos Mater 8:64–71. https://​doi.​org/​10.​5923/j.​
tary material available at https://d​ oi.o​ rg/1​ 0.1​ 007/s​ 41024-0​ 21-0​ 0155-w. cmate​rials.​20180​803.​03
12. Hawileh RA, Abdalla JA, Naser MZ (2019) Modeling the shear
strength of concrete beams reinforced with CFRP bars under
Acknowledgements The authors wish to express their sincere thanks unsymmetrical loading. Mech Adv Mater Struct 26:1290–1297.
to Dr. Maheswaran Rathinasamy, associate professor civil engineer- https://​doi.​org/​10.​1080/​15376​494.​2018.​14328​03
ing department, and the management of MVGR engineering college 13. Mustafa SAA, Hassan HA (2018) Behavior of concrete beams
for facilitating the workstation in the research laboratory and being reinforced with hybrid steel and FRP composites. HBRC J
permitted to conduct the simulation work. 14:300–308. https://​doi.​org/​10.​1016/j.​hbrcj.​2017.​01.​001
14. Osman BH, Wu E, Ji B, Abdulhameed SS (2017) Shear behavior
Funding No external funding was received for this work. of Reinforced Concrete (RC) beams with circular web openings
without additional shear reinforcement. KSCE J Civ Eng 21:296–
Declarations 306. https://​doi.​org/​10.​1007/​s12205-​016-​0387-7
15. Vasudevan G, Kothandaraman S (2014) Finite element analy-
sis of bearing capacity of RC beams retrofitted with external
Conflict of interest No potential conflict of interest was reported by bars. Strength Mater 46:831–842. https:// ​ d oi. ​ o rg/ ​ 1 0. ​ 1 007/​
the authors. s11223-​014-​9617-8
16. Ashour SA (2000) Effect of compressive strength and tensile
reinforcement ratio on flexural behavior of high-strength concrete
beams. Eng Struct 22:413–423. https://​doi.​org/​10.​1016/​S0141-​
0296(98)​00135-7
References 17. Yousaf M, Siddiqi ZA, Sharif MB, Qazi AU (2017) Force- and
displacement-controlled non-linear FE analyses of RC beam with
1. Naser MZ, Hawileh RA, Abdalla J (2021) Modeling strate- partial steel bonded length. Int J Civ Eng 15:499–513. https://​doi.​
gies of finite element simulation of reinforced concrete beams org/​10.​1007/​s40999-​016-​0076-4
strengthened with FRP: a review. J Compos Sci 5:19. https://​ 18. ANSYS (2007) ANSYS release version 12, a finite element com-
doi.​org/​10.​3390/​jcs50​10019 puter software and user manual for nonlinear structural analysis.
2. Al-Rousan RZ, Alhassan M, Al-wadi R (2020) Nonlinear finite ANSYS, Canonsburg
element analysis of full-scale concrete bridge deck slabs rein- 19. Barour S, Zergua A (2020) Numerical analysis of reinforced con-
forced with FRP bars. Structures 27:1820–1831. https://​doi.​org/​ crete beams strengthened in shear using carbon fiber reinforced
10.​1016/j.​istruc.​2020.​08.​024 polymer materials. J Eng Des Technol. https://​doi.​org/​10.​1108/​
3. Nwankwo CO, Ede AN (2020) Flexural strengthening of rein- JEDT-​03-​2020-​0099
forced concrete beam using a natural fibre reinforced polymer 20. Jawdhari A, Harik I (2018) Finite element analysis of RC beams
laminate: an experimental and numerical study. Mater Struct strengthened in flexure with CFRP rod panels. Constr Build Mater
Constr. https://​doi.​org/​10.​1617/​s11527-​020-​01573-x 163:751–766. https://​doi.​org/​10.​1016/j.​conbu​ildmat.​2017.​12.​139
4. Ramesh B, Eswari S, Sundararajan T (2021) Experimental and 21. El-Shaer MAA, Samaan MF (2018) Experimental and numeri-
numerical studies on the flexural behaviour of GFRP laminated cal investigation of flexural behaviour of new DSG reinforced
hybrid-fibre-reinforced concrete (HFRC) beams. Innov Infra- concrete mixes reinforced with GFRP bars. Ain Shams Eng J
struct Solut. https://​doi.​org/​10.​1007/​s41062-​020-​00374-z 9:3437–3449. https://​doi.​org/​10.​1016/j.​asej.​2018.​07.​002
5. Hawileh RA, Musto HA, Abdalla JA, Naser MZ (2019) Finite 22. Banjara NK, Ramanjaneyulu K (2017) Experimental and numeri-
element modeling of reinforced concrete beams externally cal investigations on the performance evaluation of shear deficient
strengthened in flexure with side-bonded FRP laminates. and GFRP strengthened reinforced concrete beams. Constr Build

13
Journal of Building Pathology and Rehabilitation (2022) 7:22 Page 23 of 23 22

Mater 137:520–534. https://​doi.​org/​10.​1016/j.​conbu​ildmat.​2017.​ 32. Wight JK, MacGregor JG (2020) Reinforced concrete mechanics
01.​089 and design, 6th edn. Prentice hall
23. Sakar G, Hawileh RA, Naser MZ, Abdalla JA, Tanarslan M (2014) 33. Hawileh RA, Abdalla JA, Tanarslan MH, Naser MZ (2011) Mod-
Nonlinear behavior of shear deficient RC beams strengthened with eling of nonlinear cyclic response of shear-deficient RC T beams
near surface mounted glass fiber reinforcement under cyclic load- strengthened with side bonded CFRP fabric strips. Comput Concr
ing. Mater Des 61:16–25. https://​doi.​org/​10.​1016/j.​matdes.​2014.​ 8:193–206. https://​doi.​org/​10.​12989/​cac.​2011.8.​2.​193
04.​064 34. Abu-Obeidah A, Hawileh RA, Abdalla JA (2015) Finite element
24. Hawileh RA (2015) Finite element modeling of reinforced con- analysis of strengthened RC beams in shear with aluminum plates.
crete beams with a hybrid combination of steel and aramid 704 Comput Struct 147:36–46. https://​doi.​org/​10.​1016/j.​comps​truc.​
reinforcement. Mater Des 65:831–839. https://​doi.​org/​10.​1016/j.​ 2014.​10.​009
matdes.​2014.​10.​004 35. beton fib-federation internationale du (2013) Model code for con-
25. Pathak P, Zhang YX (2019) Numerical study of structural behav- crete structures 2010. 1st ed. Ernst & Sohn
ior of fiber-reinforced polymer-strengthened reinforced concrete 36. Özcan DM, Bayraktar A, Şahin A, Haktanir T, Türker T (2009)
beams with bond-slip effect under cyclic loading. Struct Concr Experimental and finite element analysis on the steel fiber rein-
20:97–107. https://​doi.​org/​10.​1002/​suco.​20180​0035 forced concrete (SFRC) beams ultimate behavior. Constr Build
26. Dahmani L, Khennane A, Kaci S (2010) Crack identification in Mater 23:1064–1077. https://d​ oi.o​ rg/1​ 0.1​ 016/j.c​ onbui​ ldmat.2​ 008.​
reinforced concrete beams using ANSYS software. Strength Mater 05.​010
42:232–240. https://​doi.​org/​10.​1007/​s11223-​010-​9212-6 37. ACI Committee 318 (2019) Building code requirements available
27. Bitencourt LAG, Manzoli OL, Trindade YT, Rodrigues EA, Dias- for public review (ACI 318-19)
da-Costa D (2018) Modeling reinforced concrete structures using 38. Pandimani, Markandeyaraju P, Yesuratnam G (2021) Nonlinear
coupling finite elements for discrete representation of reinforce- modelling and finite element analysis on the load bearing capac-
ments. Finite Elem Anal Des 149:32–44. https://d​ oi.o​ rg/1​ 0.1​ 016/j.​ ity of RC beams with considering the bond–slip effect. J Eng
finel.​2018.​06.​004 Des Technol. https://d​ oi.o​ rg/1​ 0.1​ 108/J​ EDT-0​ 6-2​ 021-0​ 310-4​ 80-2​ 1
28. Godínez-Domínguez EA, Tena-Colunga A, Juárez-Luna G (ahead of printing)
(2015) Nonlinear finite element modeling of reinforced concrete 39. Pandimani, Markandeyaraju P, Yesuratnam G (2021) A compre-
haunched beams designed to develop a shear failure. Eng Struct hensive nonlinear finite element modelling and parametric analy-
105:99–122. https://​doi.​org/​10.​1016/j.​engst​ruct.​2015.​09.​023 sis of reinforced concrete beams. World J Eng. https://​doi.​org/​10.​
29. Adawi A, Youssef MA, Meshaly ME (2016) Finite element mod- 1108/​WJE-​04-​2021-​483-​0212 (ahead of printing)
eling of the composite action between hollowcore slabs and the
topping concrete. Eng Struct 124:302–315. https://​doi.​org/​10.​ Publisher's Note Springer Nature remains neutral with regard to
1016/j.​engst​ruct.​2016.​06.​016 jurisdictional claims in published maps and institutional affiliations.
30. Sajedi S, Huang Q (2017) Load-deflection behavior prediction of
intact and corroded RC bridge beams with or without lap splices
considering bond stress-slip effect. J Bridg Eng 22:04016102.
https://​doi.​org/​10.​1061/​(asce)​be.​1943-​5592.​00009​81
31. Naser MZ, Hawileh RA (2019) Predicting the response of continu-
ous RC deep beams under varying levels of differential settlement.
Front Struct Civ Eng. https://d​ oi.o​ rg/1​ 0.1​ 007/s​ 11709-0​ 18-0​ 506-2

13