Journal of Advanced Concrete Technology Vol.

21, 380-391, May 2023 / Copyright © 2023 Japan Concrete Institute 380

Scientific paper

Effect of Bond Property Between CFRP Bar and Concrete on Nonlinear

Behavior of Circular CFRP Reinforced Concrete Beams
Mamy A. Rabotovao1 and Tomohiro Miki2*

Received 23 December 2022, accepted 29 April 2023 doi:10.3151/jact.21.380

Abstract
Carbon fiber reinforced polymer (CFRP) bars present advantages over conventional steel rebars such as non-corrosiveness
and relatively high tensile strength. CFRP bars cannot directly substitute steel rebars as the main reinforcement for
concrete structures due to a lack of understanding of the behavior of CFRP-reinforced concrete (RC) members. This
research investigates the effect of surface characteristics of deformed FRP bars on the bond between FRP bars and
concrete. Moreover, the research aims to understand how the bond property of FRP bars affects the behavior of CFRP RC
beams under shear load. An analysis was conducted for a better understanding of this matter by producing different bond
models with a reproductive pull-out test that takes into consideration the surface characteristics aspect of the FRP bar.
The structural analysis for CFRP RC beams using the proposed bond models was conducted. It was observed that surface
characteristics of CFRP rebars affect the bond-slip curves in a way that the ribbed type presents higher bond strength and
greater shear stiffness modulus than the dented type. In addition, the analysis for CFRP RC beams using a ribbed bond
model tends to generate steeper diagonal cracks leading to the shear failure in the CFRP RC beams.

1. Introduction rebars on a wide scale as a reinforcing material for

concrete structures are: (i) when subjected to tensile force
Steel reinforcing bars are the main material for the FRP exhibits linear elastic behavior up to failure and does
internal reinforcement of concrete structures. However, not show any yielding point. In other words, the FRP bar
in the presence of corrosive environments, steel is a brittle material. (ii) The modulus of elasticity for
reinforcing bars even in concrete may suffer severe some types of FRP is much lower than steel, hence
deterioration due to corrosion limiting the life expectancy deflection and crack widths may control the design of
of these concrete members and leading to costly repairs reinforced concrete structures.
and rehabilitation. Thus, much research has been carried Whether at the ultimate state (capacity) or at the
out on controlling the corrosion of steel rebars or finding serviceability limit state (crack width and deflection) the
alternative materials that provide the desirable behavior of a reinforced concrete member depends on the
characteristics of steel rebars as well as prevent corrosion. force transfer between the reinforcement and the concrete.
The use of non-corrosive materials such as fiber This force transfer depends on the bond between the two
reinforced polymer (FRP) could be one of the solutions materials. Therefore, the bond property between FRP
to that matter. FRP reinforcing bars present several bars and concrete is a critical aspect of the structural
advantages over conventional steel bars, mainly non- performance of the FRP reinforced members.
corrosiveness, high tensile strength, lightweight, fatigue Still in the topic of concrete structures in aggressive
resistance, nonmagnetic electrical insulation, small creep environments, concrete members with a circular cross-
deformation, and specific gravity (Hao et al. 2009). section are often used in bridges or marine structures as
Therefore, FRP bars have been used as reinforcement for pile or fender since it provides equal strength
different concrete structures subjected to an aggressive characteristics in all directions. In the last decade, the
environment such as marine structures and bridges experimental works have mainly focused on FRP
(Mohamed and Benmokrane 2015). reinforced concrete members with a rectangular cross-
Despite the advantages of FRP bars over conventional section. Hence, research related to circular carbon fiber
steel reinforcement, a direct substitution between FRP reinforced polymer (CFRP) reinforced concrete beams or
and steel rebar is not possible due to differences in the columns is a necessity (Mohamed and Benmokrane 2015).
mechanical and physical properties between the two This study aims at a better understanding of the bond
materials. The main problems that prevent the use of FRP between FRP bars and concrete and its effects on the
structural behavior of FRP RC concrete beams.
Furthermore, this study uses two different surface
1
characteristics of FRP bars for sensitive analysis of the
Graduate Student, Department of Civil Engineering, internal strain distribution of concrete. It was found that
Kobe University, Kobe, Japan. sand-coating surface characteristics were the ones
2
Associate Professor, Department of Civil Engineering, leading to the highest load carrying capacity, and
Kobe University, Kobe, Japan. *Corresponding author, analytical specimens sustain larger deflection than
E-mail : mikitomo@port.kobe-u.ac.jp
M. A. Rabotovao and T. Miki / Journal of Advanced Concrete Technology Vol. 21, 380-391, 2023 381

experimental ones (Rabotovao and Miki 2022). The bond strength is defined as the maximum bond
Compared to the previous article (Rabotovao and Miki stress observed on the bond-slip curve obtained by the
2022), this study presents a better understanding of the precedent equation. The bond strength between concrete
effect of different bond-slip parameters, such as the shape and FRP bars depends on the embedment length, the
of the bar, the elastic modulus, the diameter, and the local diameter of the bar, the nature of the bar, and its surface
bond property. The present study also focuses on the treatment.
effect of the kinematic shape of FRP rebars as other The following presents the effect of different
parameters were taken as constant depending on the parameters on the bond strength and bond behavior
shape of FRP rebars. between FRP bars and concrete after consideration of
This paper also provides details on the behavior 1002 pull-out tests (Nepomuceno et al. 2021). Unlike
between FRP bars and concrete on a microscopic scale steel reinforcing bars, FRP bars are composed of a
allowing detailed information regarding strain and stress wrapping matrix and fibers leading to a lower modulus
distribution on a small interface element between of elasticity and lower stiffness. Previous research
concrete and FRP bars. Moreover, the first analytical (Aiello et al. 2007; Baena et al. 2009; Nepomuceno et al.
program suggests the concept of a local bond between 2021) has reported that a type of fiber influences the bond
concrete and FRP bars during pull-out test which shows strength between FRP bars and concrete. The collected
strain and stress distribution. Also, this paper presents data has reported the effect of three types of FRP bars,
how the study on a microscopic scale can be applied to a namely BFRP (B: Basalt), CFRP (C: Carbon), and GFRP
real structural component such as circular CFRP beams. (G: Glass).
Furthermore, for the FEM part, this paper investigated The experiment showed that BFRP bars presented a
two different interface elements, namely a surface higher bond strength than CFRP bars and GFRP bars.
interface element and a line interface element. Regarding the modulus of elasticity, CFRP bars presented
a higher modulus of elasticity than those of BFRP and
2. Literature review on the bond property GFRP bars, whereas CFRP bars presented a lower
between FRP and concrete performance than BFRP and GFRP bars since no bar
rupture is observed during pull-out tests (Nepomuceno et
FRP reinforced concrete members rely on the durability al. 2021). This failure mode of CFRP bars in the direct
of their composing materials (concrete and FRP bars) and pull-out test could be advantageous for the FRP RC beam
on the bond property between concrete and FRP bars. since reinforcement would not contribute to additional
Since used for a long period of time in the field of civil cracks in the concrete. The collected data takes into
engineering, the durability of concrete is not questioned consideration 15 different embedment lengths ranging
(fib 2010). On the other hand, FRP bars present from 2ø to 20ø, where ø is the nominal diameter of the
advantageous properties that are considered durable, FRP bars. Among those embedment lengths, the use of
however, the bond between FRP bars and concrete is far 5ø is the most common one, representing up to 48% of
from being understood. specimens. Moreover, this embedment length of 5ø is the
One way to understand the bond between FRP bars and suggested value in international guidelines for bond
concrete is the pull-out test. This setup test consists of a characterization of FRP bars to concrete (ACI 2015;
block of concrete with an embedded FRP bar inside, and ASTM 2020; CNR 2007; CSA 2012).
it allows a better understanding of the short and long- In addition, it was reported that the bond strength
term behavior of the bond between concrete and FRP bars. decreases with the embedment length (Achillides and
The setup consists of pulling the FRP bar from the block Pilakoutas 2004; fib 2007, El Refai et al. 2015) due to a
of concrete while this latter is restrained. Precisely, the decrease in the friction effects along the length. This
bar itself is not bonded to the concrete throughout its all reduction of friction is the consequence of Poisson’s ratio
length but only for a certain length called “embedded effects represented by a reduction in the radial direction
length.” This test setup results in understanding the while the bar elongated along its length (fib 2007).
applied force ( F ) and the relative displacement between The bond stress between concrete and FRP bars can be
the concrete and the slips at the loaded section and at the explained with three different mechanisms: chemical
free end. The applied force leads to the calculation of the adhesion, mechanical interlocking, and frictional bond.
bond stress between concrete and FRP bars with the Chemical adhesion is defined as the resistance that leads
following equation: to the separation of materials of different natures
(concrete, FRP matrix). On the other hand, mechanical
F interlocking is related to the shape and deformation of the
 (1)
 lemb ø FRP bar that generates localized forces when slipping
occurs. And the last mechanism is created by the
where existence of the surface contact of concrete and FRP bars.
 : Bond stress (N/mm2) Hence, the surface treatment plays a significant role in
F : Pull-out tensile force (N) good bond performance. Therefore, Nepomuceno et al.
lemb : Embedded length in concrete (mm) (2021) presented different techniques, including sand
ø : Diameter of the bar (mm)
M. A. Rabotovao and T. Miki / Journal of Advanced Concrete Technology Vol. 21, 380-391, 2023 382

coating, surface deformations (helical wrapping, namely R10-In-12-G-P5 and R12-Rb-12-G-P6 (Solyom
indentation), or a combination of both techniques and Balázs 2020). The height and width of the rib of the
recently, and reported the influence of different surface ribbed bar were 0.74 mm and 3.54 mm, respectively, and
treatments of FRP bars. The FRP bars were subdivided will be referred to as a large rib or a large indentation in
into four groups, namely: sand-coating with a deformed the following analysis. The depth and width of the
surface (SC+DS), deformed surface (DS), sand coating indentation of the dented bar were 0.46 mm and 1.84 mm,
(SC), and smooth surface (SM). These different surface respectively, and will be referred to as a small rib or a
treatments were considered with an embedment length of small indentation.
5ø. From the experiments, it was found that FRP bars DIANA ver. 10.5 was used to model the specimens for
with surface treatment present a higher bond strength the pull-out tests. Regarding the meshing of the embedded
than smooth bars, thus a better bond performance. part of the CFRP bar, the ribbed part and dented part were
modeled with a fine mesh size of 0.5 mm. The other
3. Analytical programs surface of the bar presents a mesh size of 1.5 mm. Such
discretization was considered to represent the interlock
3.1 Analysis for pull-out tests of FRP bar mechanics between ribs or indentation and concrete cover.
embedded in concrete The different angles of indentation or rib are considered in
The analytical model of the pull-out test is targeted for the parametric study. The angle corresponds to a spacing
experimental results performed by Solyom and Balázs of indentation or rib, rs as shown in Fig. 1.
(2020). The specimens were designed by taking into For the tensile behavior of concrete, a bilinear tensile
consideration ACI, CSA, and RILEM recommendations. model characterized by a plateau after the tensile strength
Each specimen consists of a cubic mold of 150 mm with is reached, was considered with a tensile strength of 2.5
an FRP bar embedded in the center. The embedment N/mm2. For the compressive behavior, a bilinear model
length was taken as 5ø. This analytical program takes into with a plateau was considered as well with a compressive
consideration the shapes of deformed surfaces. FRP bars strength of concrete as 36 N/mm2 corresponding to an
with ribs and indentation are shown in Fig. 1. elastic modulus of 30000 N/mm2. CFRP bar was modeled
As shown in Fig. 1, the face of the specimen is fixed with an elastic model. The only parameter taken into
along the direction of the loading (X-axis). The bottom consideration was the elastic modulus of CFRP rebars
edge of the specimen is fixed along the Y- and Z-axes to 141000 N/mm2.
prevent any displacement or rotation. In addition, two The interface between concrete and FRP bars was
types of bars are used in the analysis; a bar with rib and a modeled as a 3D surface element. The shear stiffness
bar with indentation referred to as a ribbed bar and a modulus and the normal stiffness modulus of the
dented bar, respectively. The dimensions of rib and interface are calculated by the following equations,
indentation are taken from experimental specimens, which are based on the parametric analysis.

Fig. 1 Representative model for pull-out test.

M. A. Rabotovao and T. Miki / Journal of Advanced Concrete Technology Vol. 21, 380-391, 2023 383

Fig. 2 One interface element of the surface of the FRP bar.

directed in the plane of the surface interface and the z-axis

K nn is perpendicular to that surface. However, displacements
K tt  (2)
1000 and shear tractions are calculated in the global system
XYZ. The local bond was introduced as a bilinear curve
K nn  E /  t×N  (3) in this study.
As shown in Fig. 3, two different local bond diagrams
where were considered here; the first one, “local bond 15” is
K tt : shear stiffness modulus (N/mm3) fitted to the averaged bond stress-slip relationship of the
K nn : normal stiffness modulus (N/mm3) experiment by Solyom and Balázs (2020), and the second
E : elastic modulus of concrete (N/mm2) one, “local bond 45” was selected to observe the effect of
t : thickness of the interface element an increase in local bond on the overall bond-slip curves.
N : number of interface elements The exaggerated value of a local bond of 45 N/mm2 is
For simplicity in the calculation of those stiffnesses considered a trial since in experimental studies, the bond
mentioned above, the thickness of the interface element is defined as a mean value of traction force other than the
is assumed to be 1 mm. surface in contact between FRP bars and concrete. This
The interface between concrete and FRP bars consists implies that the shear traction on every part of the
of a surface interface composed of interface elements, as interface between the concrete and FRP bar is not
shown in Fig. 2. The nonlinearity of the interface uniform or equal to that mean value.
elements is controlled by a diagram that introduces the
concept of a local bond. The local bond is associated with 3.2 Analysis for CFRP RC beams with a circular
each interface element of the surface interface. A cross-section
maximum local shear stress is defined for which the This study aims at understanding the behavior of circular
connection breaks. Once this local shear stress in the cross-sectional CFRP RC beams under shear. As an
element interface has reached the limit in the respective analytical design, specimens consist of circular beams
direction, the stiffness in all directions in the element with a diameter of 500 mm and a total length of 3000 mm
interface will reduce. Figure 2 shows the relationship (Mohamed et al. 2017). Specimens were reinforced with
between the global coordinate system XYZ and the local sand coated CFRP longitudinal rebars (16 mm dia.) and
coordinate system xyz of the surface interface, where, for CFRP spiral (12 mm dia.) with different spacings of 100
each node of an interface element, the x- and y-axes are mm, 150 mm, and 200 mm.
Regarding the boundary conditions, each specimen is
simply supported other a length span of 2400 mm. The
load is applied as displacement control in the Z direction.
The left support is fixed in all directions; however, the
right support is fixed in Y- and Z-axes.
DIANA ver. 10.5 was used to model those specimens.
In the analysis, several bond models of CFRP rebars are
associated with the longitudinal reinforcing bars. The
concrete was modeled with solid elements with a mesh
size of 50 mm. The total strain crack model is considered.
The behavior of concrete in compression was modeled
using a parabolic model (Feenstra 1993). The
compressive fracture energy of concrete GC is one of the
main parameters and was determined using Eq. (4)
(Yamaya et al. 1999). The compressive strain of concrete
Fig. 3 Local bond parameter on each interface element.
M. A. Rabotovao and T. Miki / Journal of Advanced Concrete Technology Vol. 21, 380-391, 2023 384

was 36 N/mm2. The reduction of the compressive restrained in all directions, although the right one is
strength of concrete due to lateral cracking is also taken restrained in the direction of Y- and X-axes. The diameter
into consideration (Vecchio and Collins 1993). After of the longitudinal CFRP rebar (number of longitudinal
cracking, an exponential tension softening model is rebars: 16) was 16 mm with an elastic modulus of 141000
considered (Hordijk 1991). The tensile fracture energy N/mm2 and a tensile strength of 1659 N/mm2. The spiral
GF is calculated with Eq. (5) and was obtained consisted of CFRP bar with a diameter of 12 mm and an
according to the JSCE standard specification for concrete elastic modulus of 124000 N/mm2, and a tensile strength
structures (JSCE 2002). The tensile strength of concrete of 1562 N/mm2.
was set to 2.5 N/mm2.
4. Results of the representative model for
GC  8.77 f c' (4) pull-out test

GF  10  d max  . f c'1/3 with d max  20 mm To understand the bond behavior of those new materials,
1/3
(5)
different parameters are considered: an angle of
where indentation, an angle of rib, a size of indentation and rib,
GC : compressive fracture energy of concrete (N/mm) a diameter of rebar, and the modulus of elasticity of FRP
GF : tensile fracture energy of concrete (N/mm) rebars which represented for GFRP and CFRP rebars. In
f c' : compressive strength of concrete (N/mm2) the parametric analysis, the local bond model with 15
d max : maximum size of aggregate (mm) N/mm2 was used in sections from 4.1 to 4.4.
Figure 4 presents the details of the analytical
specimens, including the boundary conditions, the mesh 4.1 Effect of indentation of FRP bar on bond-
discretization, and the disposition of reinforcements. slip curves
Regarding the boundary conditions, each specimen is This section shows the effect of an indentation of FRP bar
supported by two supports, where the left support is on the bond-slip curves, especially an effect of the angle
and the size of indentation, as shown in Fig. 5. Analytical

Fig. 4 Modeling for CFRP RC beams with different spiral spacings.

Fig. 5 Effect of indentation on the bond-slip curves between FRP (12 mm bar) and concrete.
M. A. Rabotovao and T. Miki / Journal of Advanced Concrete Technology Vol. 21, 380-391, 2023 385

Fig. 6 Effect of ribs on the bond-slip curves between FRP (12 mm bar) and concrete.

results show that dented bar with angles of indentation of modulus. In the model, it is assumed in the analysis that
15° and 30° leads to a large slip at the maximum stress of the angle of the rib or indentation was increased with
1.88 mm and 2.27 mm, respectively, as presented in Fig. decreasing the spacing. For a dented bar, it is also
5(a). This is because of the almost uniform and smooth assumed that the smaller width and depth of the
surface all along the embedded part. On the other hand, indentation, the larger the round surface except for the
larger angles of indentation, such as 45° and 60°, indentation of the bar which is in contact with the
generate smaller slips at the maximum stress of 1.33 mm concrete in the analytical model. The slip at the concrete
and 1.06 mm, respectively. The different angles of interface may occur when large shear traction stresses
indentation in the later cases do not affect the bond lead to large strain reaching the zone of indentation.
strength since no actual difference was observed. Since smaller indentation provides larger round surface
However, it significantly affects the stiffness modulus, contact between concrete and bars, the bond stress-slip
which indicates a slope of the bond-slip curve of those relationship of the dented bar with a large indentation and
specimens. Therefore, in the case of the dented specimen, 60° angle indicated similar to that of the dented bar with
the angle of 60° is considered the suitable angle since it a small indentation and an angle of 15 to 30 degrees.
is the one that leads to the highest stiffness modulus.
To observe the effect of the size of indentation, the 4.2 Effect of ribs on the bond-slip curves between
following two different types of dented bars are FRP and concrete
considered: Among the four different angles of ribs, the angle of
i) Bar with small indentation: a depth of 0.46 mm and a 45° provides the fittest result in the case of ribbed
width of 1.84 mm and, specimens as shown in Fig. 6(a). Figure 6(b) shows
ii) Bar with larger indentation: a depth of 0.74 mm and a the bond between FRP bars and concrete tends to
width of 3.54 mm. increase with the enlargement of the size of the rib. In
Regarding the size of the indentation, a larger and a addition, regarding the size of ribs, the following
smaller indentation were both associated with an angle of dimensions were considered:
indentation of 60° since it presents the highest stiffness i) Bar with small rib: 0.46 mm rib height and 1.84 mm
modulus. The diameter of both bars was set 12 mm in this rib width and,
analysis. ii) Bar with larger rib: 0.74 mm rib height and 3.54 mm
The bond-slip curves at the free end of these bars are rib width.
shown in Fig. 5(b). It is obvious that smaller indentation
presents the lower slip and higher bond strength, 4.3 Effect of diameters on the bond-slip curves
corresponding, of course, to a larger shear stiffness between FRP and concrete
Figure 7 shows that the bond between FRP bars and
concrete decreases with the increase in diameter. The
diameter affects the bond strength of each pull-out test.
Based on Fig. 7, it is analytically found that the bond
between FRP bars and concrete decreases with the
increase in diameter. There is no significant difference
regarding the slip at the maximum bond stress and the
diameter affects mainly the bond strength of each pull-
out test.

4.4 Effect of the modulus of elasticity of FRP

bars
Fig. 7 Effect of different diameters on the bond between Figure 8 shows how different moduli of elasticity of FRP
dented GFRP bar and concrete.
M. A. Rabotovao and T. Miki / Journal of Advanced Concrete Technology Vol. 21, 380-391, 2023 386

Table 1 Comparison between experimental and analytical results.

Bond strength Slip at max. stress Stiffness
Designation
(N/mm2) (mm) (N/mm3)
Small indentation – Local bond 15 7.29 1.88 3.88
Small indentation – Local bond 45 12.14 1.11 10.93
Specimen R10-In-12-G-P5 in the study of Solyom and
15.13 1.56 9.88
Balázs (2020)

bars affect the bond-slip curve between FRP bars and shows that the results of bond stress of the dented bar
concrete. The module of elasticity of the GFRP bar and model with a diameter 12 mm changed from 7.3 N/mm2
CFRP bar used in this section are 45000 N/mm2 and up to 12.1 N/mm2 in which the local bond stress has
141000 N/mm2, respectively, and for both bars, a Poisson ranged from 15 N/mm2 and even up to 45 N/mm2. Here,
ratio of 0.2 was set. Those bars are dented bars with the the stress traction distribution of local shear stress shows
same angle of indentation of 60°. Figure 8 shows that that it is not uniformly distributed over the interface
among the different parameters that can affect the bond- element as shown in Fig. 10. This non-uniform
slip curves between concrete and FRP bars, the elastic distribution of the local bond stress is the reason why an
modulus is not a predominant one since almost similar increase in the local bond stress from 15 N/mm2 up to 45
bond-slip curves are observed, though an increase in the N/mm2 results only in the increase of 166% in the bond
elastic modulus was clearly taken into consideration. In strength of the pull-out specimen in the analysis. The
this study, an embedment length of 5ø (ø: nominal analytical results and the results in the case of the
diameter) was considered, which may lead to those experimental specimen R10-In-12-G-P5 in the study of
results. Solyom and Balázs (2020) are summarized in Table 1.
The bond-slip curves of different elastic modulus These results indicate that the stiffness of the analytical
values are identical since the failure of the specimen is a results was similar to that of the experimental one.
concrete failure but not the bar failure. It just shows that Therefore, in the following analysis, this local bond of 45
the elastic moduli of those bars are not a predominant N/mm2 is considered for the bond model for CFRP bars
parameter regarding their bond-slip curves. with a diameter of 16 mm.

4.5 Effect of different local bond curves on the 4.6 Bond models of CFRP bar with a diameter of
overall bond-slip curves 16 mm and different surface characteristics
In the parametric analysis shown previously, the local This section presents the bond models obtained from this
bond was considered in the analysis, where the bond representative model for the pull-out test, which will be
strength ranges from around 7 to 8 N/mm2. Figure 9 used as parameters in new sets of analyses. Based on the

Fig. 8 Effect of elastic modulus on the bond between Fig. 9 Effect of local bond on the bond slip curve of dented
dented FRP bar and concrete (angle of indentation: 60°). bar (diameter 12 mm).

Fig. 10 Distribution of the shear traction on the surface interface element for two different local bond models at the peak
load.
M. A. Rabotovao and T. Miki / Journal of Advanced Concrete Technology Vol. 21, 380-391, 2023 387

previous experiments, a larger rib associated with an different models. For both surface characteristics, the
angle of 45° is presented as the highest stiffness modulus bond-slip curves are composed of an initial slope
for the ribbed bar. On the other hand, a smaller followed by a plateau. These bond-slip curves were used
indentation with an angle of 60° shows the highest as the linear interface parameters between CFRP bars and
stiffness modulus for the dented bar. Therefore, those two concrete in the structural analysis of CFRP RC beams.
types are considered with the modulus of elasticity of
CFRP rebars and a diameter of 16 mm. 5. Results of the CFRP RC beams
Regarding the initial stiffness of those two models, the
dented bond model presents an initial slip of 0.162 mm at 5.1 Load carrying capacity
the bond stress of 2.13 N/mm2. On the other hand, the Figure 12 shows the load-deflection curves of different
ribbed bond model presents an initial slip of 0.006 mm at BC specimens, including the experimental results
the bond stress of 2.16 N/mm2. Therefore, initial (Mohamed et al. 2017). It is obvious that there is a
stiffnesses used in the analyses for those models were set difference in terms of stiffness between the experimental
as 336 N/mm3 for the ribbed bond model and 13 N/mm3 results and analytical results. However, the elastic
for the dented bond model, respectively. behavior of the specimens was observed in both cases
Figure 11 shows the bond-slip curves of the two (experimental and analytical). The experimental
specimen BC100 reached a load carrying capacity of
1270 kN corresponding to a deflection of 26 mm.
In general, the behavior of CFRP RC beams can be
divided into two stages. Prior to the flexural cracking, all
the beams present the same behavior, corresponding to
the load of 125 kN. This stage represents the behavior of
uncracked beams with the gross moment of inertia of the
concrete cross-section. After cracking, an almost linear
behavior is observed in almost all cases, regardless of the
nature of the surface characteristics of the CFRP rebars.
At this stage, a reduction of stiffness up to failure can be
observed. Assuming that in the post cracking stage, the
beam stiffness mainly depended on the axial stiffness of
the reinforcing bars. The strength of the beam depended
Fig. 11 Bond models used in the analysis for the CFRP on its reinforcements after the cracking.
RC beams.

Fig. 12 Deflection curves of all BC specimens (BC100, BC150 and BC200).

M. A. Rabotovao and T. Miki / Journal of Advanced Concrete Technology Vol. 21, 380-391, 2023 388

Table 2 Tensile strains of longitudinal reinforcements at the midspan (tensile reinforcement).

Experimental results of Mohamed et al.
Specimens Perfect bond Ribbed Dented
(2017)
BC100 0.0090 0.0086 0.0076 0.0089
BC150 0.0071 0.0071 0.0070 0.0085
BC200 0.0063 0.0061 0.0060 0.0069

In the case of BC100, the load carrying capacities in stiffness modulus leads to larger slip between concrete
the analysis were respectively 1155 kN, 1117 kN, and and CFRP longitudinal rebars which reduces the force
970 kN for rebars with the perfect bond, the ribbed rebar, transfer between those materials, leading to a low load
and the dented rebar models. Those load-deflection carrying capacity of BC100 reinforced with dented CFRP
curves of BC 100 and BC specimens show that the nature rebar.
of the bond between CFRP bars and concrete affects The internal strain at the midspan of the longitudinal
relatively the flexural stiffnesses of the specimen and CFRP bar with a perfect bond model is similar to the
therefore the nature of the reinforcement should be taken strain of CFRP bars in the experiment. However, the
into great consideration in the design of FRP RC load-deflection curve of the analytical specimen shows a
structures. low stiffness if compared to the experimental results.
This pattern is also observed in the cases of BC150 and
BC200 since the perfect bond model cases present the 5.3 Cracking patterns of BC specimens (BC100,
highest peak load followed by the ribbed model case, and BC150, and BC200)
the lowest load carrying capacity is observed for the case This section focuses on analyzing the effect of different
of the dented model. surface characteristics of longitudinal CFRP rebar on the
crack propagation of circular RC beams with spiral CFRP.
5.2 Tensile strain of longitudinal CFRP rebars To observe that crack propagation, the cracking of circular
(BC100, BC150, and BC200) beams at two different loading stages is shown in this
This section presents the tensile strain of longitudinal section. Those two different loading stages are namely: the
CFRP bars obtained in the analysis at the peak load in the loading of 400 kN corresponding to the opening of the
midspan of the tensile reinforcement. As shown in Table diagonal crack and the peak load of each specimen.
2, the analytical tensile strains of longitudinal rebars are Figure 13 shows the cracking patterns of BC100 at a
0.0090, 0.0086, and 0.0076 for the perfect bond model, loading of 400 kN for different surface characteristics.
ribbed model, and dented model, respectively. As for the BC100 with the ribbed model showed several flexural
other specimens BC150 and BC200, the difference in the cracks with the opening of diagonal cracks in the shear
tensile strain distribution shows the unequal contribution span of the beam. On the other hand, BC100 with the
of the longitudinal reinforcements after the cracking of dented model showed few large flexural cracks due to the
concrete. Those values suggested that the perfect bond poor bond between CFRP and concrete. Regarding
model presented a large force transfer among the specimen BC100 with the perfect bond model, several
different surface characteristics. On the other hand, flexural cracks are also observed, as well as the opening
dented longitudinal rebars show less contribution to the of diagonal cracks. However, if the results of the ribbed
shear strength of those specimens. model and the perfect-bond model are compared, the
Regarding the different surface characteristics, the former represents more flexural cracks.
dented bond model which presents the lowest shear Figure 14 shows the cracking patterns of specimen

Fig. 13 Cracking patterns of BC100 at P = 400 kN. Fig. 14 Cracking patterns of BC100 at the peak load.
M. A. Rabotovao and T. Miki / Journal of Advanced Concrete Technology Vol. 21, 380-391, 2023 389

BC100 for experimental and analytical programs. The figures.

cracking patterns of analytical BC100 specimens at the
peak load are similar to the cracking pattern of 5.4 Discussion on the internal strain of concrete
experimental BC100. Compared to the previous stage, This section describes the effect of different surface
there is no appearance of new diagonal cracks for all the characteristics of longitudinal rebars on the internal strain
specimens. Moreover, flexural cracks keep opening for of concrete in the principal direction. To do so, the
all the specimens. Specimens BC 100 with perfect bond internal strain of concrete at a certain loading is shown
and ribbed bond model failed after the appearance of below, in this case, loading of 800 kN. Moreover, the
several diagonal cracks. However, specimen BC100 with internal strain of concrete strain are localized at three
the dented model shows a brittle shear failure or even a different plane sections as shown in Fig. 17:
bending failure since the non-existence of real diagonal i) Plane 1: The middle axial section of concrete
cracks joining the support and the loading point. There is ii) Plane 2: The vertical section located at the distance of
no real difference between the cracking pattern of the D/4, and
perfect bond case and the ribbed model case since the iii)Plane 3: The vertical section located at the distance of
latter provides enough bond and force transfer between D/8, where D is the diameter of the circular beam.
CFRP bars and concrete to allow a similar cracking In terms of the shape of the internal strain, the perfect
pattern. bond model and ribbed bond model led to a similar
Figures 15 and 16 show the cracking patterns of consequence. The internal strain of concrete is mainly
BC150 and BC200 at peak load. The same conclusion concentrated in the shear span of the beam.
observed from Fig. 14 can be drawn again from these Figures 18, 19, and 20 present the maximum principal
strain of concrete for BC100. Regarding the internal
strain of concrete in the web of the CFRP RC beam, it
was observed that specimens strengthened by CFRP
longitudinal rebar with the perfect bond model showed

Fig. 15 Cracking patterns of BC150 at the peak load. Fig. 17 Section planes of internal strain of concrete.

Fig. 18 Internal strain of concrete (BC100 with perfect

Fig. 16 Cracking patterns of BC200 at the peak load. bond).
M. A. Rabotovao and T. Miki / Journal of Advanced Concrete Technology Vol. 21, 380-391, 2023 390

diagonal shear cracks is clearly observed, as shown in Fig. beam contributes more than that at Plane 2 due to the
18. On the other hand, the internal tensile strain of unequal contribution of longitudinal CFRP rebars. In
concrete of the beams with the ribbed model, the diagonal Plane 2, longitudinal reinforcements from the middle
shear cracks are observed as shown in Fig. 19. The axial plane to the extremity of the beam side surface
perfect bond model and the ribbed model present almost contribute to the resistance of that section when the
the same pattern regarding the internal strain which perfect bond and the ribbed bond model. It is more
indicates a diagonal shear strain distribution joining the evident that the internal strain of concrete presents a
support and the loading point as shown in Figs. 18 and uniform distribution once Plane 3 is observed. It is the
19, respectively. However, in the case of dented models, effect of the circular cross-sectional shape of the beam.
the diagonal shear strain presented a steeper angle.
Moreover, those diagonal strains started from the 6. Conclusions
midspan of the beam to the loading point, as shown in
Fig. 20. The main findings of this article can be summarized as
Another aspect of the surface characteristic of the bond follows:
model is mentioned hereafter. In the case of the perfect 1. Regarding the representative model for the pull-out
bond model, its mainly represented by an almost zero- test, it was observed that smaller indentation and large
slip displacement between longitudinal CFRP rebars and rib led to a better bond between concrete and FRP bars.
concrete. The perfect bond model can be associated with Between the two types of nature of FRP
a bond model with an infinite shear stiffness modulus. reinforcements, reinforcements with ribs present a
Therefore, the internal strain of those two surface better bond since it leads to high initial stiffness of 336
characteristics is the consequence of that high shear N/mm3 although a poor initial stiffness of 13 N/mm3
stiffness modulus. for FRP reinforcement with indentation.
Compared to the two other surface characteristics, the 2. The elastic modulus of CFRP reinforcements, on the
effect of the indentation surface characteristic is other hand, does not present a significant effect on the
somehow different. In this case, as shown in Fig. 20, it is bond-slip curves between FRP bars and concrete,
found that only a small part of the shear span contributes since CFRP bars and GFRP bars present similar bond-
to the resistance of the specimen. The poor bond between slip curves for the same diameter.
concrete and CFRP bars leads to a bending failure of the 3. Bond strength which is the maximum strength in the
specimens. bond-slip curve reduces with the increase of the
Moreover, those internal strains of concrete show the diameter of FRP bars. In this study, specimens with a
part of the gross area of concrete that contributes to the diameter of 12 mm presented the higher bond strength,
performance of those specimens. From Figs. 18, 19, and followed by a diameter of 16 mm and 20 mm.
20, the non-uniform internal strain of concrete web is 4. The parametric analysis shows that a poor bond
obvious such as a middle axial at section Plane 2 (D/4). between CFRP bars and concrete leads to bending
The non-uniform pattern of the internal strain of concrete cracks, which was the case with dented CFRP bars.
in the web of the beam is the consequence of the circular However, the ribbed CFRP bars led to a better bond
cross-sectional shape leading to different contributions of between concrete and CFRP, therefore those
CFRP longitudinal reinforcement. Concrete at Plane 1, a specimens presented a diagonal shear failure.
middle axial plane of the cross-section of the circular

Fig. 19 Internal strain of concrete (BC100 with ribbed Fig. 20 Internal strain of concrete (BC100 with dented
bond model). bond model).
M. A. Rabotovao and T. Miki / Journal of Advanced Concrete Technology Vol. 21, 380-391, 2023 391

5. The internal strain of the circular cross-sectional beam Construction, 19(3), 04014050.
with different bond properties of the longitudinal Feenstra, P. H., (1993). “Computational aspects of
CFRP rebars showed a non-uniformity as it goes from biaxial stress in plain and reinforced concrete.” Thesis
the surface to the middle axial cross-section of the (PhD). Delft University of Technology.
beam. fib, (2007). “FRP reinforcement in RC structures (fib
Bulletin 40).” Lausanne, Switzerland: Fédération
Acknowledgements Internationale du Béton.
This research was partially supported by the Japan Hao, Q., Wang, Y., He, Z. and Ou, J., (2009). “Bond
Society for the Promotion of Science (JSPS), Grant-in- strength of glass fiber reinforced polymer ribbed
Aid for Scientific Research (B), 19H02226, and Research rebars in normal strength concrete.” Construction and
Committee on Technological Development for Building Materials, 23(2), 865-871.
Infrastructure in the Ministry of Land, Infrastructure, Hordijk, D. A., (1991). “Local approach to fatigue of
Transport, and Tourism Kinki Regional Development concrete.” Thesis (PhD). Delft University of
Bureau. Technology.
JSCE, (2002). “Standard specifications for concrete
References structures-2002: Structural performance verification.”
Achillides, Z. and Pilakoutas, K., (2004). “Bond behavior Tokyo: Japan Society of Civil Engineers.
of fiber reinforced polymer bars under direct pullout Mohamed, H. M., Ali, A. H. and Benmokrane, B., (2017).
conditions.” Journal of Composites for Construction, “Behavior of circular concrete members reinforced
8(2), 173-181. with carbon-FRP bars and spirals under shear.”
ACI, (2015). “Guide for the design and construction of Journal of Composites for Construction, 21(2),
structural concrete reinforced with fiber-reinforced 04016090.
polymer (FRP) bars (ACI440.1R-15).” Farmington Mohamed, H. M. and Benmokrane, B., (2015). “Torsion
Hills, Michigan, USA: American Concrete Institute. behavior of concrete beams reinforced with glass fiber-
Aiello, M. A., Leone, M. and Pecce, M., (2007). “Bond reinforced polymer bars and stirrups.” ACI Structural
performances of FRP rebars-reinforced concrete.” Journal, 112(5), 543-552.
Journal of Materials in Civil Engineering, 19(3), 205- Nepomuceno, E., Sena-Cruz, J., Correia, L. and
213. D’Antino, T., (2021). “Review on the bond behavior
ASTM, (2020). “Standard test method for bond strength and durability of FRP bars to concrete.” Construction
of fiber reinforced polymer matrix composite bars to and Building Materials, 287, 123042.
concrete by pullout testing (ASTM D7913).” West Rabotovao, A. M. and Miki, T., (2022). “Influence of
Conshohocken, Pennsylvania, USA: ASTM surface characteristics of CFRP rebars on behavior of
International. CFRP strengthened RC beams with a circular cross
Baena, M., Torres, L., Turon, A. and Barris, C., (2009). section.” Proceedings of the Japan Concrete Institute,
“Experimental study of bond behaviour between 44(2), 115-120.
concrete and FRP bars using a pull-out test.” Solyom, S. and Balázs, G. L., (2020). “Bond of FRP bars
Composites Part B, 40(8), 784-797. with different surface characteristics.” Construction
CNR, (2007). “Guide for the design and construction of and Building Materials, 264, 119839.
concrete structures reinforced with fiber-reinforced Vecchio, F. J. and Collins, M. P., (1993). “Compression
polymer bars (CNR 203-06).” Rome: National response of cracked reinforced concrete.” ASCE
Research Council of Italy (CNR). Journal of Structural Engineering, 119(12), 3590-
CSA, (2012). “Design and construction of building 3610.
structures with fibre reinforced polymers (CSA S806- Yamaya, A., Nakamura, H. and Nagai, T., (1999). “Shear
12).” Toronto: Canadian Standards Association. behavior analyses of RC beams using rotating crack
El Refai, A., Ammar, M.-A. and Masmoudi R., (2015). model.” Doboku Gakkai Ronbunshu, 620/V-43, 187-
“Bond performance of basalt fiber reinforced polymer 199. (in Japanese)
bars to concrete.” Journal of Composites for