Construction and Building Materials 182 (2018) 637–645

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

Bond-slip behavior between pre-corroded rebar and steel fiber

reinforced concrete
Lijun Hou, Bingxuan Zhou, Shang Guo, Ning Zhuang, Da Chen ⇑
College of Harbour, Coastal and Offshore Engineering, Hohai University, Nanjing 210098, China

h i g h l i g h t s

The bond between pre-corroded rebar and SFRC simulated the retrofit condition.

A slip plateau was observed near the peak stress for specimens with c/d of 5.75.

The effect of corrosion ratio on bond behavior was related to c/d and failing mode.

A bond strength model was proposed considering coupling effect of corrosion and c/d.

a r t i c l e i n f o a b s t r a c t

Article history: For the corroded reinforced concrete (RC) members repaired using steel fiber reinforced concrete (SFRC),
Received 1 January 2018 bonding between corroded rebar and SFRC greatly affected the mechanical performance of repaired
Received in revised form 18 May 2018 members. In the present paper, the central pull-out tests were carried out to investigate the bond behav-
Accepted 14 June 2018
ior of SFRC and pre-corroded steel bar obtained in concrete specimens. The varied parameters included
corrosion ratio and cover thickness to rebar diameter ratio (c/d). The experimental results revealed that
the confinement and bridging action of fiber can improve the bond strength and transform the failure
Keywords:
mode from splitting to pull-out or combined splitting and pull-out. Bond strength decreased with
Bond
Pre-corrosion
increasing corrosion ratio and decreasing c/d, whereas the residual bond strength remained no less than
SFRC that of non-corroded rebar especially for specimens with a small c/d. Bond strength degradation, bond
Pull-out toughness variation and bond-slip descending behaviors with corrosion ratio were correlated to c/d
Strength and failing mode. A modified bond strength model considering the coupling effect of corrosion and c/d
Toughness was proposed, and the calculated strength agreed well with the tested.
Ó 2018 Published by Elsevier Ltd.

1. Introduction by corrosion effectively [4]. Moreover, the reverse loading greatly

Bonding between rebar and concrete is a foundation in rein-
forced concrete (RC) structures. For RC structures subjected to
long-term carbonation and chloride corrosion, bond of steel bar
and the surrounding concrete is first improved due to the filling
effect of corrosion product, and then reduces due to corrosion-
induced cover cracking [1,2]. For unconfined rebar, a rebar mass
loss of about 0.5% resulted in about 50–60% enhancement in bond
strength, whereas a mass loss of 3.9% led to a sharp reduction up to
54% [3]. In contrast, for confined rebar with around 4–6% corrosion
ratio, the bond strength was almost 30–65% larger than that of
unconfined rebar at a similar corrosion level, and the confinement
of transverse reinforcement can counteract the bond loss induced
⇑ Corresponding author.
E-mail address: chenda@hhu.edu.cn (D. Chen).
aggravated the corrosion-induced bond degradation, especially
within the first five cycles and for the unconfined rebar [5]. In addi-
tion, the corrosion-induced bond degradation decreased with
increasing strain rate, but strain rate effect was gradually weak-
ened with increasing corrosion level [6].
More severely, the structural performance of corroded RC mem-
bers at the service and ultimate limit states degraded apparently
due to the deteriorated bond as well as reduced rebar area, ductility
and strength of rebar, etc. And thus, the patch repair for
corroded RC members is usually applied to improve the load-
carrying capacity, stiffness and durability for extending the service
age. Steel fiber reinforced concrete (SFRC), showing good mechani-
cal properties in tension, flexure, shear, impact and fatigue, etc.
[7–11] and durability in freezing and thawing, permeability and
carbonation, etc. [11–13], has been used in retrofit and strengthening
of old and damaged structures as well as the construction of new

https://doi.org/10.1016/j.conbuildmat.2018.06.116
0950-0618/Ó 2018 Published by Elsevier Ltd.
638 L. Hou et al. / Construction and Building Materials 182 (2018) 637–645
structures [14]. Note that the property of SFRC subjected to chloride
corrosion is of great significance for the engineering application in
marine environment. The review on SFRC [15] confirms that the
steel fiber in uncracked SFRC had a superior chloride corrosion
resistance compared with rebar in concrete mainly due to the
discontinuity and the uniform steel surface of fibers along with
the dense and uniform fiber-matrix interface. For cracked SFRC
with crack width below the critical width (i.e., 0.2 mm), the crack
and damaged fiber-matrix interface can heal eventually [16,17].
In the meantime, fiber-matrix interface bond was improved due
to the increasing fiber roughness, healing fiber-matrix interface
and accumulated chemical binding products [18]. The study con-
ducted by Granju and Balouch [19] indicated that the flexural
strength of pre-cracking SFRC specimens was enhanced under chlo-
ride corrosion for one year. The fibers across the crack within a 2–3
mm rim from the specimen surface suffered severe corrosion,
whereas the fibers located out of this rim exhibited a light corrosion
without section loss. Balouch et al. [20] further revealed that the
minimum cover of 0.1 mm or 0.2 mm can prevent the surface
corrosion of SFRC with w/c no larger than 0.5. Therefore, SFRC
may be used in the repair of marine engineering structures if a
certain thickness of cover can be provided for fiber.
When corroded RC members were repaired using SFRC, the
cracking cover, damaged and chloride-rich concrete surrounding
the corroded rebar and corrosion products bonded on rebar must
be removed before repair [21,22]. Accordingly, bonding between
corroded rebar and SFRC is of great importance for the mechanical
property of repaired members. However, the related studies con-
cerning the bond between rebar and SFRC mainly concentrated
on the non-corroded rebar and the effect of fiber fraction and
parameters. The experimental results reported by Yazici et al.
[23] revealed that bond strength was enhanced with increasing
fraction and fiber length to diameter ratio (lf/df). In detail, for steel
fiber with lf/df of 40–60, the addition of 60 kg/m3 fiber resulted in
6–9% bond strength enhancement [23]. The bond experiment con-
ducted by Campione et al. [24] indicated that the use of steel fiber
in volume fraction within 0.5–2% produced 7–38% improvement in
bond strength. For pre-stressing strands, the bond strength
increased by 30% when using 15 kg/m3 steel fiber, and the optimal
fiber fraction was related to the strand diameter and fiber param-
eters [25]. In contrast, the experimental study conducted by Gao
et al. [26] indicated that the addition of steel fiber within 2% vol-
ume fraction merely resulted in a 3% increase in bond strength.
However, the bond behavior of corroded rebar embedded in
SFRC remains unclear, especially for the corroded rebar on which
corrosion products bonded are removed under the retrofit condi-
tion. In the present paper, the pre-corroded rebar, obtained in con-
crete specimens through accelerated corrosion test, was cleaned
and used for pull-out specimens to simulate the actual repair con-
dition. The effect of corrosion ratio and cover-rebar diameter ratio
on bond behaviors was discussed concerning failure modes, bond-
slip curves, bond strength and bond toughness, etc. The experi-
mental results can provide a reference for the retrofit of corroded
RC members using SFRC.
2. Experimental program
2.1. Pre-corrosion of rebar
For simulating the retrofitting condition of corroded RC mem-
bers, a total of 27 steel bars used for bond test were pre-
corroded in concrete cylinders. The pre-corrosion region was taken
as 150 mm in the middle of concrete cylinder, and concrete cover
thickness was 25 mm. The non-corrosion region of rebar in con-
crete was covered with epoxy resin and protected using PVC
Fig. 1. Details of the pre-corroded specimen.
sleeve. Fig. 1 shows the detailed configuration of pre-corroded
specimens. The variable of pre-corrosion test included rebar diam-
eter and corrosion ratio. The deformed rebar with 12, 16 and 20
mm in diameter was used and, the corresponding center distance
between crescent ribs was 8.48, 9.94 and 10.05 mm for these three
types of rebar. The target corrosion ratio was designed as 5%, 10%
and 15%. Thus these 27 specimens can be divided into 9 groups
with 3 specimens each group.
The accelerated corrosion test using applied direct-current was
employed for pre-corrosion. Concrete cylinders were immersed
into 3.5% NaCl solution covering the whole pre-corrosion region
of rebar. The reinforcement was connected with the anode of
power, and stainless steel bars with a diameter of 8 mm were con-
nected with the cathode. The corrosion current density applied was
300 lA/cm2, and corrosion age can be estimated according to
Faraday’s law,
zF Dm
t¼ ð1Þ
MiS
where, F is Faraday’s constant (96,500 C/mol), z is the ionic charge
number (z = 2) of iron, Dm is the mass loss of corroded reinforce-
ment (g), M is the molar mass of iron (56 g/mol), i is the corrosion
current density (A/cm2), and S is the superficial area of rebar within
the corrosion region (cm2). After the pre-corrosion, the corroded
rebar was removed though splitting cylinders. Further, corrosion
product and concrete bonded on the rebar were cleaned up using
steel wire brush to obtain a similar rebar surface to that in an actual
repair condition.
2.2. Pull-out specimens
The 150 mm cube with a centrally placed rebar was used as the
pull-out specimen. A total of 45 pull-out specimens were cast
including 9 concrete comparison specimens and 36 SFRC samples.
The variable parameters included corrosion ratio (q) and cover
thickness to rebar diameter ratio (c/d). In this test, corrosion ratio
was taken as 0% and the actual pre-corrosion ratio. c/d was
designed as 5.75, 4.19 and 3.25 with increasing rebar diameter
from 12 to 20 mm. Note that the actual pre-corrosion ratio was
measured after pull-out tests, and the details were presented in
the following section.
Fig. 2 shows the details of pull-out specimen. The bond length-
rebar diameter ratio (l/d) was taken as 5.0 with bond length of 60,
80 and 100 mm for 12, 16 and 20 mm rebar respectively. A pair of
PVC sleeves was used at two ends to obtain a precise bond length.
And the gap between PVC pipe and rebar was sealed using glue to
prevent the infiltration of fresh cement matrix.
 L. Hou et al. / Construction and Building Materials 182 (2018) 637–645
Fig. 2. Configuration of central pull-out specimens.
2.3. Materials
Plain concrete and SFRC were used for pull-out specimens. The
mixture proportion of concrete by weight was cement:water:
sand:coarse aggregate = 1:0.45:1.73:2.00. The matrix of SFRC was
same as the concrete used. The fraction of hooked steel fiber added
was 0.75% in volume. P.O. 42.5 cement, coarse aggregate with a
maximum diameter of 10 mm, and the medium river sand were
used to prepare concrete and SFRC. The hooked steel fiber had a
length of 35 mm, a length to diameter ratio of 65, an elastic mod-
ulus of 200 GPa, and a tensile strength of 1345 MPa.
Pull-out specimens were cast in plywood molds. Meanwhile, two
groups of 100 mm cube were left to measure the compressive
strength of concrete and SFRC. All the specimens were cured outdoors
for 28 days though covering sand and plastic sheet and watering 1–2
times every day to remain specimen wet. The cube compressive
strength fcu of concrete and SFRC was 59.57 and 52.52 MPa. In addi-
tion, the tested yield strength fy and ultimate tensile strength fu were
460 and 594 MPa for 12 mm rebar, 456 and 586 MPa for 16 mm
rebar, and 439 and 578 MPa for 20 mm rebar, respectively.
2.4. Test apparatus and procedure
Pull-out test was carried out on a 300 kN electrohydraulic servo-
controlled testing machine. The load was applied using a displace-
ment controlled mode at a speed of 0.5 mm/min. Fig. 3 shows a
Fig. 3. Test apparatus and loading set-up used in this test.
639
custom-built loading set-up used in this test. In this set-up, a pair
of spherical hinges was arranged at up and lower end plates of load-
ing frame to ensure a central tensile force. A 100 kN load cell was used
to measure the pull-out force. Two pairs of linear variable difference
transformers (LVDTs) were symmetrically arranged at the loading
and free ends of pull-out specimen to obtain the relative slip between
rebar and matrix. All the test data were collected continuously by a
quasi-static data acquisition system connected with a computer.
2.5. Actual corrosion ratio and corrosion appearance
After the pull-out tests, the corroded rebar was taken out
through splitting the specimens. Then, the matrix bonded on the
rebar was removed using steel wire brush. At last, the bond region
of corroded rebar was cut out to measure the mass loss. The ratio of
tested mass loss Dm to initial mass m0 (Dm/m0) was used to repre-
sent the actual corrosion ratio qc. The calculated corrosion ratio is
summarized in Table 1. Note that the actual corrosion had a differ-
ence from the target pre-corrosion ratio. Here, if the difference in
actual corrosion ratio was within 1%, the corresponding specimens
were considered as one series and the average corrosion ratio was
used to denote the corrosion level of this series. In addition, con-
crete and SFRC were simplified as C and SC, and the reinforcement
was denoted by D12, D16 and D20 according to rebar diameter.
The Arab number 1, 2, 3 and 4 was utilized to represent different
sample in the same series. For example, specimen D16-10.4-SC2
represents the second SFRC specimen of the series with 10.4%
actual average corrosion ratio and 16 mm diameter of rebar.
Fig. 4 shows the typical appearance of corroded rebar in the bond
region. As shown in Fig. 4, the corrosion at the side close to NaCl
solution was far severe compared with the back side. The rib damage
was aggravated with increasing corrosion ratio and decreasing rebar
diameter. At corrosion below about 5%, a slight corrosion pit
occurred along rebar, and then corrosion pit and rib damage aggra-
vated, and even the ribs at the heavy corrosion side have disap-
peared at corrosion ratio beyond about 12% for D12 and D16 series.
3. Experimental results and discussion
In this section, the effect of corrosion ratio and cover thickness
to rebar diameter ratio on bond behavior was discussed. Table 1
summarizes the experimental results including nominal bond
strength (su), average nominal bond strength for the same series
(su,av), and failure mode, etc.
The nominal average bond stress s along the bond length is
generally used to represent the local bond approximately, as
expressed by Eq. (2),
P
s¼ ð2Þ
pdl
where, P is pull-out force, l is bond length, and d is nominal rebar
diameter. However, a non-uniform corrosion was presented along
longitudinal length especially for rebar with large corrosion ratio.
And thus it is difficult to measure the diameter after corrosion pre-
cisely. In addition, the nominal diameter of non-corroded rebar was
always used to compute the bond strength without considering the
variation of rebar section in post-corrosion condition where bond
was measured after corrosion without removing corrosion products
[3–6]. Therefore, for simplification, the initial nominal diameter was
still used for calculating the average bond stress in the present
pre-corrosion condition. Bond strength (su) can be taken as the
nominal average bond stress at the peak point. Average slip
between steel bar and matrix is usually used to represent the slip
behavior [27], as expressed by Eq. (3)
sl þ sf
s¼ ð3Þ
2
640 L. Hou et al. / Construction and Building Materials 182 (2018) 637–645

Table 1
Details and experimental results of test specimens.

Specimen l/mm c/d qc/% qc,av/% su/MPa su,av/MPa su,cal/MPa failure mode
D12-0-C1 60 5.75 0.00 0.00 19.09 18.82 16.84 S
D12-0-C2 60 5.75 0.00 20.00 16.84 S
D12-0-C3 60 5.75 0.00 17.36 16.84 S
D12-0-SC1 60 5.75 0.00 0.00 14.33* 20.85 18.52 PO
D12-0-SC2 60 5.75 0.00 21.31 18.52 PO
D12-0-SC3 60 5.75 0.00 20.39 18.52 PO
D12-4.4-SC1 60 5.75 4.81 4.44 17.45 16.67 16.63 PO
D12-4.4-SC2 60 5.75 4.07 15.89 16.92 PO
D12-5.6-SC1 60 5.75 5.54 5.57 16.69 17.73 16.34 PO
D12-5.6-SC2 60 5.75 5.95 17.37 16.18 PO
D12-5.6-SC3 60 5.75 5.23 19.11 16.46 PO
D12-8.3-SC1 60 5.75 8.06 8.26 15.28 13.33 15.35 PO
D12-8.3-SC2 60 5.75 8.45 11.37 15.19 PO
D12-12.3-SC1 60 5.75 12.30 12.30 12.24 12.24 13.68 PO
D12-16.9-SC1 60 5.75 16.91 16.91 12.63 12.63 11.86 PO
D16-0-C1 80 4.19 0.00 0.00 15.76 14.80 13.64 S
D16-0-C2 80 4.19 0.00 13.79 13.64 S
D16-0-C3 80 4.19 0.00 14.84 13.64 S
D16-0-SC1 80 4.19 0.00 0.00 14.80 14.54 14.99 S-PO
D16-0-SC2 80 4.19 0.00 15.49 14.99 S-PO
D16-0-SC3 80 4.19 0.00 13.32 14.99 S-PO
D16-3.4-SC1 80 4.19 3.50 3.39 15.12 14.13 14.12 S-PO
D16-3.4-SC2 80 4.19 3.28 13.14 14.17 S-PO
D16-5.1-SC1 80 4.19 5.09 5.09 13.82 13.82 13.73 S-PO
D16-9.2-SC1 80 4.19 9.18 9.18 14.18 14.18 12.71 S-PO
D16-10.4-SC1 80 4.19 10.56 10.38 12.06 12.15 12.37 S-PO
D16-10.4-SC2 80 4.19 10.20 12.24 12.46 S-PO
D16-12.9-SC1 80 4.19 12.58 12.88 13.37 13.37 11.86 S-PO
D16-12.9-SC1 80 4.19 13.17 12.29 11.72 S-PO
D16-14.1-SC2 80 4.19 14.08 14.08 11.62 11.62 11.49 S-PO
D20-0-C1 100 3.25 0.00 0.00 10.13 9.44 11.52 S
D20-0-C2 100 3.25 0.00 8.88 11.52 S
D20-0-C3 100 3.25 0.00 9.33 11.52 S
D20-0-SC1 100 3.25 0.00 0.00 13.54 12.90 12.66 S-PO
D20-0-SC2 100 3.25 0.00 12.62 12.66 S-PO
D20-0-SC3 100 3.25 0.00 12.54 12.66 S-PO
D20-4.9-SC1 100 3.25 4.98% 4.86% 11.86 11.76 11.79 S-PO
D20-4.9-SC2 100 3.25 4.73% 12.45 11.84 S-PO
D20-4.9-SC3 100 3.25 4.38% 11.45 11.90 S-PO
D20-4.9-SC4 100 3.25 5.34% 11.29 11.73 S-PO
D20-6.7-SC1 100 3.25 6.20% 6.71% 11.10 10.54 11.58 S-PO
D20-6.7-SC2 100 3.25 7.21% 9.98 11.40 S-PO
D20-8.6-SC1 100 3.25 8.15% 8.56% 10.37 10.82 11.24 S-PO
D20-8.6-SC2 100 3.25 8.96% 11.28 11.10 S-PO
D20-11.0-SC1 100 3.25 10.98% 10.98% 10.71 10.71 10.75 S-PO
Mean 1.01
C.V. 0.11

Note: * this bond strength was not used to calculate su,av due to a difference from su,av beyond 15%; qc,av is the average corrosion ratio.

5.0 8.96 11.0 8.96 failure. The failure mode of tested specimens is listed in Table 1.
3.5 10.2
The failure mode was mainly dependent on c/d of pull-out speci-
14.1 10.2
mens and tension property of matrix material surrounding the
5.5 8.5 16.9 8.5 rebar. The concrete comparison specimens all failed in splitting
even if a large c/d of 5.75 was used. In contrast, SFRC specimens
with c/d of 5.75 failed in pull-out, whereas the rest SFRC specimens
failed in combined splitting and pull-out. This can be mainly attrib-
uted to the improved tension capacity and ductility due to steel
fibers distributed in SFRC. The splitting cracks were restrained
effectively depending on the bridging action of fibers across cracks.
3.5%~5.5% 8.5%~10.2% 11%~16.9% back Fig. 5 shows the typical crack pattern for different failure
modes. For splitting failure, the specimen was separated into two
Fig. 4. Appearance of corroded rebar for D16 series.
parts at the peak load. For pull-out failure, no radial and longitudi-
nal cracking occurred during pull-out of rebar. For combined split-
where, s is the average slip, sl and sf are the slip at the loading and ting and pull-out failure, the splitting crack formed at the loading
free ends. end of specimens when approaching the peak load and then grad-
ually widened and developed to the free end with increasing slip.
3.1. Failure mode As shown in Fig. 5, rebar corrosion within about 15% of mass loss
had a negligible effect on the crack and failure pattern. However,
Three failure modes were showed including splitting (S) failure, the crack width increased apparently with decreasing c/d regard-
pull-out (PO) failure and combined splitting and pull-out (S-PO) less of matrix material. For concrete samples, D12 series with c/d
 L. Hou et al. / Construction and Building Materials 182 (2018) 637–645 641

interlocking action between rebar and SFRC was weakened due

D12-0-C D16-0-C D20-0-C
to damage of SFRC teeth between the rebar ribs. However, for
D16 and D20 series, the confinement provided by surrounding
matrix was weakened because of splitting cracking near the peak
load. Consequently, SFRC teeth between ribs may be subjected to
a heavy damage when experiencing a slight slip, which resulted
in a sharp stress reduction at the peak point especially for D20 ser-
(a) ies with a smaller c/d.
As shown in Fig. 6, the residual bond strength increased with
D12-0-SC D16-0-SC D20-0-SC increasing c/d as the peak bond stress did. Here, the residual bond
strength sr is assumed to be the bond stress at the slip of 1.5cr, in
which cr is the center distance between rebar ribs. It is found that
sr varied from 7.48 MPa for c/d of 5.75 to 1.65 MPa for c/d of 3.25.
Note that the residual bond strength was mainly provided by the
friction resistance between rebar and surrounding matrix. The
D16-10.4-SC D16-14.1-SC D20-11.0-SC increasing c/d greatly improved the integrity of specimens during
pull-out of embedded rebar depending on declined splitting crack
width or even no splitting cracking.
Fig. 7 shows the bond-slip curves of SFRC specimens. As shown
in Fig. 7(a)–(c), rebar corrosion greatly affected the bond-slip
behavior. For the ascending branch, both bond stiffness and peak
(b) bond stress of corroded rebar became lower than that of non-
corroded rebar, and an increase in c/d aggravated the bond loss.
Fig. 5. Typical failure patterns of test specimens: (a) concrete specimens; (b) SFRC For the descending branch, the degradation gradient of bond stress
specimens with different corrosion ratios.
decreased with increasing corrosion ratio, especially for D12 series
failing in pull-out. In addition, the drop of bond stress at the peak
point was mitigated for D16 and D20 series with corroded rebar.
of 5.75 showed a slight splitting crack propagation compared with
The deteriorated mechanical interlocking action due to corroded
D16 and D20 series. For SFRC specimens, the splitting crack width
rebar ribs was mainly responsible for the bond strength loss.
increased from 0 mm for c/d of 5.75 to 0.28–0.80 mm for c/d of 4.19
Meanwhile, the rib corrosion damage was relatively severe for
and to 0.72–1.34 mm for c/d of 3.25.
rebar with small diameter at the same corrosion level, resulting
in a more apparent bond stiffness reduction for D12 series com-
pared with other series.
3.2. Bond-slip curves
Note that the bond stress at the slip of 1.5cr may be larger than
the stress at the lower end point of descending branch, such as
Fig. 6 shows bond-slip curves of non-corroded rebar embedded
specimen D16-14.1-SC1 shown in Fig. 7(b). Therefore, the residual
in concrete and SFRC. As shown in Fig. 6, the concrete specimens
bond strength was further taken as the minimum value of the bond
merely had an ascending branch, whereas the SFRC specimens all
stress at the slip of 1.5cr and that at the lower end point of
showed typical three-branch curves including ascending, descend-
descending branch. It is found from Fig. 7 that the residual bond
ing and residual branches.
strength of corroded rebar was not lower than that of non-
Moreover, c/d had a great effect on the bond-slip curves of SFRC
corroded rebar, especially for rebar with corrosion ratio beyond
specimens, especially for the bond behavior around the peak load.
about 6% except for specimen D12-16.9-SC1. This mainly can be
D12 series presented an apparent nonlinear slip behavior around
attributed to two aspects. One is the reduced stress concentration
the peak bond stress. In contrast, D16 and D20 series showed a
on matrix surrounding the damaged ribs, which resulted in rela-
sharp reduction in bond stress at the peak point, and this drop
tively light damage in matrix and thus improved confinement.
was aggravated with decreasing c/d. This phenomenon mainly
The other is the non-uniform corrosion appearance along steel
arose from different failure modes. For D12 series failing in pull-
bar. It may lead to varied contact profiles between rebar and
out, the embedded rebar was still subjected to a strong confine-
matrix within bond length, and thus the friction resistance can
ment from the surrounding SFRC without splitting cracks. The peak
be improved during pull-out of rebar. In contrast, with respect to
bond stress can be maintained within a certain slip until the
the relative residual bond strength (sr/su), it was gradually
enhanced with increasing corrosion ratio regardless of c/d. In
25 detail, sr/su increased from 0.32 to 0.65 for D12 series, from 0.21
D12-0-SC3
to 0.50 for D16 series and from 0.15 to 0.39 for D20 series.
20
3.3. Bond strength
D12-0-C1
15
/MPa

Bond strength su and its average value su,av for all the series are
10 D16-0-C3 summarized in Table 1. For specimens with non-corroded rebar,
bond strength of SFRC specimens was larger than that of concrete
D16-0-SC1 specimens for D12 and D20 series, whereas the bond strength of
5 D20-0-C3 SFRC and concrete specimens was close to each other for D16
D20-0-SC2 series. Note that the compressive strength of SFRC was about
0
0 3 6 9 12 15 18 6.71 MPa lower than that of concrete. If the bond strength is nom-
cu , su,av/fcu of series D16-0-SC and D16-0-C was 2.06
inalized to f0.5 0.5
s /mm
and 1.97 respectively. In other words, the addition of fiber was
Fig. 6. Bond-slip curves of non-corroded rebar in concrete and SFRC. helpful to improve the bond strength due to the fiber bridging
642 L. Hou et al. / Construction and Building Materials 182 (2018) 637–645

25 between pullout and beam tests was inadequate so far. de Almeida

Filho et al. [28] compared the bond between reinforcement and
D12-0-SC3 concrete using RILEM pullout and beam tests, where the same
20
c/d but different l/d were used. The experimental results indicated
that the bond strength obtained from both types of tests was close
15 D12-5.6-SC2
D12-8.3-SC1 to each other although the bond length was 5d and 10d for pullout
/MPa

D12-16.9-SC1 and beam specimens respectively. In general, the bond strength

10 gradually decreases with increasing bond length. In other words,
the bond strength obtained from beam test is expected to be supe-
5 rior to that from pullout test if using a same bond length. However,
D12-12.3-SC1 D12-4.4-SC1 the further comparison study on the pullout and beam tests needs
to be performed to understand the relationship of bond behavior
0
0 3 6 9 12 15 18 obtained from these two types of tests.
s /mm Fig. 8 shows the variation of bond strength with corrosion ratio
for SFRC specimens. As shown in Fig. 8, the bond strength was
(a) correlated to both corrosion ratio and c/d. The bond strength
almost varied with corrosion ratio linearly from 20.85 to 12.63
18 MPa for D12 series, from 14.54 to 11.96 MPa for D16 series, and
D16-0-SC1
from 12.9 to 10.71 MPa for D20 series. The bond strength of
15 D16-3.4-SC1 deformed rebar mainly depended on three aspects: adhesion,
mechanical interlock and friction. And the mechanical interlocking
12 action played a key role because the adhesion failed early and the
D16-5.1-SC1 D16-14.1-SC1 friction mainly determined the residual bond strength. The dam-
/MPa

9 aged ribs due to corrosion weakened the mechanical interlocking

D16-12.9-SC1
effect and resulted in reduction in bond strength. Moreover, D12
6 D16-10.4-SC2
series obtained a relatively large reduction in bond strength com-
3 pared with D16 and D20 series. Bond strength loss was approxi-
D16-9.2-SC1 mate 39%, 20% and 17% at the corrosion ratio of about 17% for
0 D12 series, 14% for D16 series and 11% for D20 series respectively.
0 3 6 9 12 15 18 For D12 series failing in pull-out, the interlocking action of ribs can
s /mm be fully exerted during pull-out, and thus the damage in ribs is
expected to greatly deteriorate the bond strength. In contrast,
(b) D16 and D20 series failed in combined splitting and pull-out, and
the mechanical interlocking action was discounted due to splitting
15 cracking of the surrounding SFRC. Consequently, the bond perfor-
D20-0-SC2
mance of specimens with a large c/d may be more sensitive to
12 the corrosion damage.

9 D20-6.7-SC1 3.4. Bond toughness

/MPa

D20-8.6-SC2
D20-11.0-SC1 Bond toughness was evaluated by the area under bond-slip
6
curves referring to the flexural toughness evaluation method ASTM
C 1609 [29], and can be obtained by
3 Z si
D20-4.9-SC1
Ai ¼ sds ð4Þ
0 0
0 3 6 9 12 15 18
where, Ai is area under bond-slip curve up to the ith target slip si,.
s /mm Note that the target slips were extended to cover the whole

(c)
24
Fig. 7. Bond-slip curves of corroded rebar embedded in SFRC: (a) D12 series; (b) c/d=5.75
c/d=4.19
D16 series; (c) D20 series. 20
c/d=3.25

action. Moreover, bond strength of SFRC and concrete specimens 16

with non-corroded rebar was enhanced with increasing c/d. As
/MPa

12
shown in Table 1, bond strength of SFRC specimens varied from
20.85 MPa for c/d of 5.75 to 12.9 MPa for c/d of 3.25. This can be
u

8
attributed to the improved confinement effect of the surrounding
matrix with an increase in c/d.
4
However, the pullout test was applied rather than beam test in
this study. During pullout testing, the matrix near the loading end 0
is subjected to compression action of loading plate, which is differ- 0 3 6 9 12 15 18
ent from the actual stress state of structural members. In contrast,
c
/%
the flexure test may be close to the actual stress state in structural
members. However, the comparison study on bond behavior Fig. 8. Bond strength between SFRC and rebar with different ratios.
 L. Hou et al. / Construction and Building Materials 182 (2018) 637–645 643

descending and residual branches, including the slip sp at the peak improved compared with that of non-corroded rebar. This mainly
point, and 3, 5, 10 and 15 mm. arose from the relatively small bond strength loss but improved
Fig. 9 shows the bond toughness of specimens with corrosion residual bond strength and decreased bond stress drop at the peak
ratio and c/d. It is revealed that the variation of bond toughness point especially for the cases with a large corrosion ratio.
with corrosion ratio is correlated with c/d. For D12 series shown
in Fig. 9(a), the bond toughness of corroded rebar first remained
close to that of non-corroded rebar at corrosion ratio below 8.3%, 4. Bond strength model and discussion
and then reduced at corrosion ratio beyond 12.3%. Although the
ratio of residual bond strength to su (sr/su) of corroded series For the non-corroded specimens, a group of existed bond
D12-12.3-SC and D12-16.9-SC were far larger than that of non- strength models, used for plain concrete proposed by Aslani and
corroded series D12-0-SC, it hardly compensated the toughness Nejadi [30] and for SFRC proposed by Harajli et al. [31], were
loss due to the severe reduction of bond strength. In contrast, for applied to evaluate the bond strength, as expressed by Eq. (5),
D16 and D20 series, the bond toughness of corroded rebar was 8h  0:6  i 0 0:55
>
< 0:679 dc þ 3:88 dl f c for plain concrete
su0 ¼ qffiffiffiffiffiffi  qffiffiffiffiffiffi ð5Þ
>
: 0:78 f cþK c
0 2=3 0
c d
6 2:57 f c for SFRC
200 Ap A3 A5 A10 A15
where, f0c is cylinder compressive strength, taken as 80% of cube
160 compressive strength fcu (0.8fcu) [32]; Kc is fiber confinement factor,
given as K c ¼ 0:45cV f Lf =df , in which Vf, lf and df are the fiber fraction
A /(MPa mm)

120 in volume, fiber length and diameter respectively. If the fiber vol-
ume fraction is equal to 0, then the bond strength model for SFRC
80 degrads into a model for plain concrete.
Based on the plain concrete bond model in Eq. (5), the calcu-
40 lated bond strength ranged from 17.53 to 21.96 MPa for concrete
series and from 16.36 to 20.63 MPa for SFRC series without consid-
0 ering the fiber contribution with increasing c/d from 3.25 to 5.75.
0 4 8 12 16 20 Compared with the experimental results, the calculated strength
c
/% had the maximum overestimation of 17% for D12 series, but an
average overestimation of approximate 27% for D16 and D20 series
(a)
except for D20-0-C series with a deviation up to 86%. In contrast,
150 the predicted bond strength using the SFRC model in Eq. (5) ranged
Ap A3 A5 A10 A15 from 11.52 to 16.84 MPa for concrete specimens with an underes-
timation below 10% for D12 and D16 series but an overestimation
120
of 22% for D20 series. And for SFRC specimens, the calculated bond
strength was 16.66, 14.99 and 12.66 MPa for D12, D16 and D20
90
A /(MPa mm)

series respectively. It indicates a good prediction for D16 and

D20 series failing in S-PO but a 20% underestimation for D12 series
60 failing in PO when using SFRC model in Eq. (5). The comparison
above reveals that the SFRC model proposed by Harajli et al. [31]
is more applicable to predict the bond strength of non-corroded
30
specimens regardless of concrete and SFRC. And thus it was further
applied in the present study.
0 Moreover, if further ignoring the upper limit of the Harajli
0 3 6 9 12 15 model in Eq. (5), the ratio of predicted to tested strength of D12-
c
/% 0-SC series became 0.89 almost satisfying the requirement in engi-
(b) neering field. Also, Campione et al. [24] studied the bond behavior
of rebar embedded in SFRC with c/d up to 6.58, and the experimen-
125 tal details are summarized in Table 2. As the Table shows, the bond
Ap A3 A5 A10 A15
strength calculated based on the Harajli model was 0.64–0.78su
100 but became 0.83–0.91su if ignoring the upper limit requirement
of 2.57f0c 0.5. It is pointed out that the upper limit of bond strength
in the Harajli model was determined based on the specimens with
A /(MPa mm)

75
c/d ranged from 2.30 to 3.18 [33]. However, in the cases with a
large c/d, i.e., 5.75 of D12 series in the present test and 6.58 of
50 c/d used in Campione et al.’s tests [24], the bond strength may be
underestimated using this bond upper limit model due to a supe-
25 rior confinement of SFRC cover. Based on this point, the upper limit
of 2.57f0c 0.5 is suggested to be cancelled within c/d of 6.58 in the
0 current model, but the further investigation needs to be conducted
0 2 4 6 8 10 12 concerning the effect of c/d on the bond strength upper limit of
c
/% SFRC.
As aforementioned in experimental results, the rebar corrosion
(c)
had an apparent effect on the bond strength. Bond strength almost
Fig. 9. Bond toughness of SFRC specimens with corrosion ratio: (a) D12 series; (b) decreased with corrosion ratio in a linear relationship, and the
D16 series; (c) D20 series. degradation gradient was greatly related to c/d. Accordingly,
644 L. Hou et al. / Construction and Building Materials 182 (2018) 637–645
Table 2
Experimental details and results reported by Campione et al. [24].
specimen ld lf/df Vf c d c/d
Squat 5 60 0 79 12 6.58
Squat 5 60 0.5 79 12 6.58
Squat 5 60 1 79 12 6.58
Squat 5 60 2 79 12 6.58
a general function considering the coupling effect of corrosion ratio
and c/d is introduced to predict the bond strength,
 a description of the cross-sectional variation along the bars. But
c
su ¼ su0 - SFRC 1 þ Aqc þ Bqc
d the related test apparatus in our laboratory.
where, su0-SFRC denotes the bond strength model for SFRC in Eq. (5);
A and B are the regression coefficients based on tested bond
strength. The following equation is obtained through multi-
parameter regression analysis,
qffiffiffiffiffiffi 2=3 
c þ Kc c
su ¼ 0:78 f 0 c 1  0:004qc  0:003qc
d d
The calculated bond strength (su,cal) based on Eq. (7) is summa-
rized in Table 1. The mean and coefficient of variation (C.V.) of the
ratio between calculated and tested bond strength (su/su,cal) were
1.01 and 0.11 respectively. The calculated strength agreed well
with the tested value, and the proposed equation can be used to
predict the bond strength of pre-corroded rebar embedded in SFRC
for corroded RC members repaired using SFRC. However, these
equations may be limited to the SFRC reinforced using hooked steel
fiber due to only this type fiber used in the present study and Hara-
jli et al.’s studies [31,34]. In addition, owing to a small experimen-
tal data in the present study, the Eq. (7) needs to be further verified
in future experimental study before used in the practice.
Note that the corrosion effect in Eq. (7) was obtained based on
the present experimental data for fiber volume fraction of 0.75%,
and may have a limitation for other fiber fractions. In fact, the SFRC
model in Eq. (5) proposed by Harajli et al. [31] was obtained based
on the bond data for SFRC with fiber volume fraction in the range
of 0.5–2.0%. On the other hand, the fiber fraction plays a key role in
the bond between rebar and SFRC when remaining the same
parameters of c/d, l/d and compressive strength. If considering a
similar influence of fiber fraction on pre-corroded rebar to that of
non-corroded rebar, then the proposed Eq. (7) may be used for
other fiber fractions. It is well known that the improved bond
under the addition of fiber mainly lies in the enhancement in
mechanical interlocking action between rebar and the surrounding
concrete due to fiber bridging action across splitting cracks. The
experimental results revealed that the bond strength was about
1.79f0.5 0.5 0.5
cu , 1.65fcu and 1.52fcu at the corrosion ratio of about 17%
for D12 series, 14% for D16 series and 11% for D20 series respec-
tively, which was far larger than 0.23–0.25f0.5cu for plain round rebar
embedded in concrete with c/d of 2.91–5.75 [35]. This means that a
good mechanical interlocking action can still be obtained
especially for D16 and D20 series although the rebar ribs were
subjected to corrosion damage. As a result, it seems to be feasible
to consider a same fiber fraction influence on bond between non-
corroded and pre-corroded reinforcement within obtained
corrosion ratio in this test. However, the further experimental
study is still needed to evaluate the fiber fraction effect on bond
of pre-corroded rebar and SFRC.
In Eqs. (6) and (7), the corrosion ratio was used rather than a
parameter reflecting the change in roughness of reinforcement
after corrosion. In fact, it is difficult to determine and measure
the surface profile of the corroded reinforcement precisely due to
a non-uniform corrosion distribution along the longitudinal length
fc kc su su-cal 2.57f0c 0.5 su-cal/su 2.57f0c 0.5/su
22.00 0.00 15.53 12.85 12.05 0.83 0.78
26.20 10.67 18.44 15.20 13.10 0.83 0.71
27.50 21.33 19.00 16.85 13.48 0.89 0.71
29.00 42.66 21.52 19.68 13.84 0.91 0.64
and on the cross section. Berrocal et al. [36] employed a 3D scan-
ning technique to determine the local corrosion level, and obtain
ð6Þ this high-precision measure is difficult to achieve due to lack of
Moreover, in practice, the corroded rebar cannot be cut out to
measure the local corrosion level using 3D scanning technique. In
general, it is the corrosion depth that is measured to simply repre-
sent the corrosion level and the change in rebar surface profile.
Correspondingly, the corrosion ratio can be obtained based on
the measured corrosion depth, the relationship between pit and
ð7Þ average corrosion depth, and correlation of cross-sectional loss
with pit and average corrosion, as presented in Campione et al.’s
study [37]. In other words, the corrosion depth and corrosion ratio
can be represented with each other. Moreover, compared with the
corrosion depth measure, the measure of corrosion ratio repre-
sented by cross-section loss (equivalent to mass loss) was of high
precision and simple operation. In addition, the corrosion ratio is
the most-common parameter to represent the rebar corrosion
degree and to be used for evaluating the residual mechanical per-
formance of corroded RC members. Therefore, the parameter of
corrosion ratio was used in the present test and in Eqs. (6) and
(7). In summary, the bond strength of corroded rebar and SFRC
can be obtained based on Eq. (7) and the transformation relation-
ship between corrosion depth and corrosion ratio.
5. Conclusions
In this paper, bond-slip behavior between SFRC and pre-
corroded rebar obtained in a concrete environment was investi-
gated through central pull-out tests. The following conclusions
can be drawn:
(1) Compared with concrete comparison specimens failing in
splitting, SFRC specimens all showed either pull-out failure
with no splitting crack or combined splitting and pull-out
failure with longitudinal splitting cracks. Besides, the split-
ting crack propagation was mainly affected by c/d rather
than rebar corrosion ratio, and the crack width reduced
gradually with increasing c/d from 0.72 to 1.34 mm for c/d
of 3.25 to 0 mm for c/d of 5.75.
(2) All the SFRC specimens showed full three-branch bond-slip
curves whereas concrete specimens had the ascending
branch merely. For non-corroded rebar embedded in SFRC,
an apparent slip plateau was presented around the peak
bond stress for specimens with c/d of 5.75, whereas a sharp
reduction in bond stress occurred at the peak point for spec-
imens with c/d below 4.19. Besides, the residual bond
strength increased with increasing c/d due to improved con-
finement of surrounding matrix.
(3) For the bond-slip curves of corroded rebar in SFRC, bond
stiffness gradually reduced with increasing corrosion ratio
especially for specimens with c/d of 5.75. Meanwhile, the
residual bond strength of corroded rebar was no lower than
that of non-corroded rebar. Further, the ratio of residual
bond strength to bond strength (sr/su) almost increased with
corrosion ratio linearly.
 L. Hou et al. / Construction and Building Materials 182 (2018) 637–645
(4) The addition of steel fiber can improve the bond strength
between rebar and concrete, especially for specimens with
c/d of 3.25. For non-corroded rebar, bond strength of SFRC
and concrete specimens was improved with an increase in
c/d.
(5) Bond strength almost linearly reduced with increasing
corrosion ratio. Bond strength loss due to rebar corrosion
was related to c/d and bond failure mode of specimens.
D12 series failing in pull-out showed bond strength loss high
up to about 39% at corrosion ratio within 17%, whereas other
series failing in combined splitting and pull-out had 20% and
17% loss at corrosion ratio within 14% for c/d of 4.19 and 11%
for c/d of 3.25.
(6) The variation of bond toughness with corrosion ratio was
affected by c/d. Bond toughness of series with c/d below
4.19 was improved with increasing corrosion ratio, whereas
that of series with c/d of 5.75 first remained almost
unchanged at corrosion below 8.3% and then gradually
reduced.
(7) A modified bond strength model was proposed considering
the coupling effect of rebar corrosion and c/d, and a good
agreement between tested and calculated bond strength
was obtained.
Conflict of interest
The authors declare that they have no conflict of interest.
Acknowledgements
This study was financially supported by the program of the
National Natural Science Foundation of China (51408186,
51679080) and the Natural Science Foundation of Jiangsu Province
(BK20140853). The authors wish to express their gratitude to M.S.
H. Liu and undergraduate innovation training group headed by J.
Qu for their contribution to part of the experimental work.
References
[1] M. Maslehuddin, I.M. Allam, G.J. Al-Sulaimani, A.I. Al-Mana, S.N. Abduljauwad,
Effect of rusting of reinforcing steel on its mechanical properties and bond
with concrete, ACI Mater. J. 87 (5) (1990) 496–502.
[2] H.S. Lee, T. Noguchi, F. Tomosawa, Evaluation of the bond properties between
concrete and reinforcement as a function of the degree of reinforcement
corrosion, Cem. Concr. Res. 32 (8) (2002) 1313–1318.
[3] S. Coccia, S. Imperatore, Z. Rinaldi, Influence of corrosion on the bond strength
of steel rebars in concrete, Mater. Struct. 49 (1–2) (2016) 537–551.
[4] C. Fang, K. Lundgren, M. Plos, K. Gylltoft, Bond behaviour of corroded
reinforcing steel bars in concrete, Cem. Concr. Res. 36 (10) (2006) 1931–1938.
[5] C. Fang, K. Gylltoft, K. Lundgren, M. Plos, Effect of corrosion on bond in
reinforced concrete under cyclic loading, Cem. Concr. Res. 36 (3) (2006) 548–
555.
[6] W.P. Zhang, H. Chen, X.L. Gu, Bond behaviour between corroded steel bars and
concrete under different strain rates, Mag. Concr. Res. 68 (7) (2015) 1–15.
_
[7] Sß. Yazici, G. Inan, V. Tabak, Effect of aspect ratio and volume fraction of steel
fiber on the mechanical properties of SFRC, Constr. Build. Mater. 21 (6) (2007)
1250–1253.
[8] H.H. Dinh, G.J. Parra-Montesinos, J.K. Wight, Shear behavior of steel fiber-
reinforced concrete beams without stirrup reinforcement, ACI Struct. J. 107 (5)
(2010) 597–606.
645
[9] M. Nili, V. Afroughsabet, Combined effect of silica fume and steel fibers on the
impact resistance and mechanical properties of concrete, Int. J. Impact Eng. 37
(8) (2010) 879–886.
[10] F. Germano, G. Tiberti, G. Plizzari, Post-peak fatigue performance of steel fiber
reinforced concrete under flexure, Mater. Struct. 49 (10) (2016) 4229–4245.
[11] G. Zhao, C. Huang, Properties and application of steel fiber reinforced concrete,
Ind. Constr. 10 (1989) 3–10 (in Chinese).
[12] C. Yi, H. Xie, W. Gao, Effects of steel fiber on the restraining for propagation of
concrete crack and on the permeability of concrete, China, Civil. Eng. J. 37 (3)
(2004) 19–25 (in Chinese).
[13] X. Xie, D. Gao, J. Zhao, Tentative study on the mechanical property under the
action of freeze-thaw and carbonation of steel fiber reinforced concrete, J.
Xi’an Univ. Archit. Technol. (Nat. Sci. Ed.) 38 (4) (2006) 514–517 (in Chinese).
[14] S. Iqbal, A. Ali, K. Holschemacher, et al., Strengthening of RC beams using steel
fiber reinforced high strength lightweight self-compacting concrete (SHLSCC)
and their strength predictions, Mater. Des. 100 (2016) 37–46.
[15] V. Marcos-Meson, A. Michel, A. Solgaard, et al., Corrosion resistance of steel
fibre reinforced concrete – a literature review, Cem. Concr. Res. 103 (2018) 1–
20.
[16] S. Abbas, Structural and Durability Performance of Precast Segmental Tunnel
Linings, University of Western Ontario, Canada, 2014.
[17] D. Homma, H. Mihashi, T. Nishiwaki, Self-healing capability of fibre reinforced
cementitious composites, J. Adv. Concr. Technol. 7 (2009) 217–228.
[18] C. Frazão, J. Barros, A. Camões, A.C. Alves, L. Rocha, Corrosion effects on pullout
behavior of hooked steel fibers in self-compacting concrete, Cem. Concr. Res.
79 (2016) 112–122.
[19] J.-L. Granju, S.U. Balouch, Corrosion of steel fibre reinforced concrete from the
cracks, Cem. Concr. Res. 35 (3) (2005) 572–577.
[20] S.U. Balouch, J.P. Forth, J.-L. Granju, Surface corrosion of steel fibre reinforced
concrete, Cem. Concr. Res. 40 (3) (2010) 410–414.
[21] Code of China, Technical code for repair and strengthening of harbor and
marine structures, JTS 311-2011, Beijing, China, 2011. (in Chinese).
[22] G. Malumbela, M. Alexander, P. Moyo, Variation of steel loss and its effect on
the ultimate flexural capacity of RC beams corroded and repaired under load,
Constr. Build. Mater. 24 (6) (2010) 1051–1059.
ß . Yazici, H.S
[23] S ß. Arel, The effect of steel fiber on the bond between concrete and
deformed steel bar in SFRCs, Constr. Build. Mater. 40 (2013) 299–305.
[24] G. Campione, C. Cucchiara, L. La Mendola, M. Papia, Steel-concrete bond in
lightweight fiber reinforced concrete under monotonic and cyclic actions, Eng.
Struct. 27 (6) (2005) 881–890.
[25] E. Baran, T. Akis, S. Yesilmen, Pull-out behavior of prestressing strands in steel
fiber reinforced concrete, Constr. Build. Mater. 28 (1) (2012) 362–371.
[26] D. Gao, L. Xie, H. Zhu, Experimental research on bond property between steel
fiber reinforced high-strength concrete and rebars (in Chinese), Build. Sci. 20
(3) (2004) 56–60.
[27] F. Xu, Z. Wu, J. Zheng, Y. Hu, Q. Li, Experimental study on the bond behavior of
reinforcing bars embedded in concrete subjected to lateral pressure, J. Mater.
Civil Eng. 24 (1) (2012) 125–133.
[28] F.M. de Almeida Filho, M.K.E.I. Debs, A.L.H.C.E.I. Debs, Bond-slip behavior of
self-compacting concrete and vibrated concrete using pull-out and beam tests,
Mater. Struct. 41 (6) (2008) 1073–1089.
[29] ASTM, Standard Test Method for Flexural Performance of Fiber-reinforced
Concrete (Using Beam with Third-point Loading), ASTM C 1609, ASTM
International, West Conshohocken, PA, USA, 2005.
[30] F. Aslani, S. Nejadi, Bond behavior of reinforcement in conventional and self-
compacting concrete, Adv. Struct. Eng. 15 (12) (2012) 2033–2051.
[31] M.H. Harajli, B.S. Hamad, A.A. Rteil, Effect of confinement of bond strength
between steel bars and concrete, ACI Struct. J. 101 (5) (2004) 595–603.
[32] Z. Guo, Strength and Constitutive Relationship of Concrete: Principle and
Application, China Architecture & Building Press, Beijing, 2004 (in Chinese).
[33] M.H. Harajli, M. Hout, W. Jalkh, Local bond stress-slip behavior of reinforcing
bars embedded in plain and fiber concrete, ACI Mater. J. 92 (4) (1995) 343–
354.
[34] M. Harajli, B. Hamad, K. Karam, Bond-slip response of reinforcing bars
embedded in plain and fiber concrete, J. Mater. Civil Eng. 14 (14) (2002)
503–511.
[35] Z. Wu, X. Zhang, J. Zheng, Y. Hu, Q.B. Li, Bond behavior of plain round bars
embedded in concrete subjected to biaxial lateral tensile-compressive
stresses, J. Struct. Eng. 140 (4) (2013). 04013089-1-11.
[36] C.G. Berrocal, I. Löfgren, K. Lundgren, The effect of fibres on steel bar corrosion
and flexural behaviour of corroded RC beams, Eng. Struct. 163 (2018) 409–425.
[37] G. Campione, F. Cannella, L. Cavaleri, Shear and flexural strength prediction of
corroded R.C. beams, Constr. Build. Mater. 149 (2017) 395–405.