Construction and Building Materials 123 (2016) 649–660

Contents lists available at ScienceDirect

Construction and Building Materials

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

Mechanical properties of corroded steel bars in pre-cracked concrete

suffering from chloride attack
Chunhua Lu a,b,⇑, Siqi Yuan a, Pu Cheng a, Ronggui Liu a
a
Department of Civil Engineering, Jiangsu University, Zhenjiang 212013, PR China
b
Department of Civil and Environmental Engineering, University of California at Berkeley, CA 94720, USA

h i g h l i g h t s

A 4-year chloride-induced corrosion of steel bar in cracked concrete is achieved.

Obvious pitting corrosion occurs on embedded bars due to the transverse crack.

Both nominal strengths and ductility decrease with increasing the corrosion degree.

The non-uniform coefficient can effectively assess the mechanical properties.

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

Article history: A 4-year natural corrosion experiment coupled with a cyclic wetting-drying process with 5% NaCl solu-
Received 8 April 2016 tion and an indoor natural corrosion process was carried out for pre-cracked reinforced concrete ele-
Received in revised form 11 July 2016 ments in this study. The diameter loss along the bar length which was greatly influenced by crack
Accepted 12 July 2016
effect was measured, and the degree of corrosion evaluated by average mass loss was calculated for all
Available online 21 July 2016
cleaned bar specimens. The tensile test results showed that both the nominal yield and ultimate
strengths and the percentage elongation decreased with the increase of mass loss, and that the effect
Keywords:
of corrosion on degradation of ductility was much greater than that of tensile strengths. Taking the
Reinforced concrete
Chloride
impact of pitting corrosion into account, a non-uniform coefficient, S, directly related to pit depths of
Transverse crack corroded bars was proposed and the degradation models of yield strength, ultimate strength and percent-
Pitting corrosion age elongation were derived based on S. It seems that this method might be a more effective way to
Mechanical property assess the mechanical properties of corroded bars with obvious pits. Finally, the reduced area at necking
Non-uniform coefficient and fracture surface morphology were compared between sound and corroded bars.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction spalling and delamination of concrete cover; ii) reduction of cross

Corrosion of reinforcing steel bars has been recognized as one of
the major causes of performance deterioration of reinforced con-
crete (RC) structures worldwide [1,2]. It is estimated that the
annual investments on corrosion related maintenance and repairs
for concrete infrastructure cost approximately $100 billion around
the world [3]. Corrosion of reinforcement in concrete can be usu-
ally induced by two major mechanisms: concrete carbonation
and chloride attack, and the latter, which is typically caused by
chloride ions from seawater or de-icing salt, is more predominant
than the former [4,5]. Steel corrosion in concrete usually brings the
following coupled consequences [6]: i) longitudinal cracking,
⇑ Corresponding author at: Department of Civil Engineering, Jiangsu University,
Zhenjiang 212013, PR China.
E-mail addresses: lch79@ujs.edu.cn, luch_ucb@berkeley.edu (C. Lu).
section and mechanical property of corroded steel and, iii) loss of
bond between steel and concrete and eventually degradation of
bearing capacity of RC element. This study will focus on the second
issue of steel corrosion in chloride contaminated environment.
Much research work has been conducted on the mechanical
properties of corroded rebars in recent years, as summarized in
Table 1. The tested specimens of corroded steel bars were mainly
obtained from following ways: i) bare bars directly subjected to
chloride salt spay [7–10] or accelerated corrosion with impressed
current [11,12]; ii) bars embedded in concrete subjected to accel-
erated corrosion with impressed current [5,11,13–17]; iii) bars
got from service concrete suffering from chloride attack [9,18] or
carbonation [15] and; iv) bars in experimental concrete element
exposed to cyclic wetting and drying condition [14] or artificial
chloride environment [10,19]. Some commonly recognized conclu-
sions can be made from above studies that the yield loads, ultimate

http://dx.doi.org/10.1016/j.conbuildmat.2016.07.032
0950-0618/Ó 2016 Elsevier Ltd. All rights reserved.
650 C. Lu et al. / Construction and Building Materials 123 (2016) 649–660

Table 1
Summary of recent research work on mechanical properties of corroded rebars.

No. Authors Specimens Corrosion Yield strength Ultimate strength Elongation/ Elastic modulus
condition Ductility
Nominal Residual area Nominal Residual area
area area
1 Almusallam et al. [5] Bars in concrete Electrical N/S N/S Decreased Marginally Decreased N/S
increased
2 Palsson and Mirza [18] Bars in old Service/chloride Not N/S Not N/S Decreased N/S
concrete affected affected
3 Cairns et al. [13] Bars in concrete Artificial pitting; N/S Not affected N/S Slightly Decreased N/S
Electrical increased
4 Du et al. [11] Bare bars; Electrical Decreased Decreased Decreased Decreased Decreased N/S
Bars in concrete
5 Apostolopoulos et al. Bare bars Chloride salt Decreased Decreased Decreased Decreased Decreased N/S
[7] spay
6 Apostolopoulos and Bare bars Chloride salt Decreased Not affected Decreased Decreased Decreased N/S
Papadakis [8] spay
7 Lee and Cho [14] Bars in concrete Electrical; Decreased N/S Decreased N/S Decreased Decreased
Salt + wetting
and drying
8 Papadopoulos et al. [9] Bare bars; Chloride salt Decreased N/S Decreased N/S Decreased N/S
Bars in old spay;
concrete Service
9 Zhang et al. [15] Bars in old Carbonation; N/S Decreased N/S Decreased Decreased N/S
concrete; Electrical
Bars in concrete
10 Francois et al. [19] Bars in concrete Chloride Decreased Not affected Decreased Increased Decreased Not affected
11 Xia et al. [12] Bare bars; Electrical Decreased N/S Decreased N/S Decreased N/S
Bars in concrete
12 Apostolopoulos et al. Bare bars; Chloride salt Decreased Decreased — — Decreased —
[10] Bars in concrete spay
12 Tang et al. [16] Bars in concrete Electrical N/S Marginally N/S Marginally Decreased Tensile stiffness
increased increased decreased
13 Zhang et al. [17] Bars in concrete Electrical N/S Not affected N/S Not affected Decreased Not affected

N/S = not stated.

loads and elongation (or ductility) of corroded steel bars will
reduce obviously with the increase of the degree of corrosion.
The effect of corrosion on the tensile strengths of corroded bars,
however, shows great difference and even contradiction in these
researches shown in Table 1, due to the facts that various speci-
mens with different non-uniform reduction in the cross-sectional
area along the length of the bars were examined, and that different
methods were used to assess the strength properties, such as nom-
inal (or apparent) strength and residual (or true) strength.
For the nominal yield and ultimate strengths of corroded rebar
calculated with nominal steel area, Apostolopoulos et al. [7], Apos-
tolopoulos and Papadakis [8], Papadopoulos et al. [9], Apos-
tolopoulos et al. [10], Du et al. [11], Xia et al. [12], Lee and Cho
[14] and Francois et al. [19] believed that they would decrease with
increasing the degree of corrosion of rebars based on their experi-
mental results. Only the investigation completed by Palsson and
Mirza [18] reported the nominal strengths of corroded bars got
from abandoned concrete bridge wouldn’t be affected by the level
of corrosion.
However, the results of residual strength determined based on
the effective cross-sectional area of corroded steel were diverse.
Although some experimental data obtained by Apostolopoulos
et al. [7], Du et al. [11] and Zhang et al. [15] showed that the resid-
ual yield strength would also decrease with increasing the corro-
sion degree, many researchers, such as Apostolopoulos and
Papadakis [8], Cairns et al. [13], Zhang et al. [17] and Francois
et al. [19] reported that the residual yield strength shouldn’t be
affected by corrosion, and the results got by Tang et al. [16] even
showed a marginal increase in yield strength of corroded bar.
The similar conclusions can also be found for corroded bars when
the residual ultimate strength was assessed, as shown in Table 1.
Except for the influence of bar type and diameter on residual
capacity of corroded rebar [11,17], the main reasons of discrepancy
in above studies may be ascribed to two aspects: i) corrosion types
of reinforcing steel bars corroded by impressed current corrosion is
very different from that under natural corrosion condition [19,20],
that is, different patterns varying from uniform corrosion to local-
ized (pitting) corrosion have different effect on performance degra-
dation of corroded bars and, ii) the critical residual steel area is
hard to determine accurately even with 3D laser scanner [16,17],
so many researchers [11,15–17] adopted average cross-sectional
area predicted by mass loss to analyze the effective tensile
strengths of corroded bars.
It should be recognized that these above efforts have made
great contributions in analyzing the mechanical properties of cor-
roded steel bars, however, experimental data on them under con-
dition of natural corrosion are still scarce [10,19] and a deeper
understanding of tensile behaviors of corroded bars under the
chloride-induced pitting corrosion in concrete is required. More-
over, from the viewpoint of life cycle cost strategies with mainte-
nance, repair and rehabilitation of deteriorated RC structures
[21,22], the initial corrosion process before or during longitudinal
cracking of cover and its influence on performance degradation
of the steel bar should be effectively analyzed and modelled in
advance. Aiming at this problem, in this paper, a natural corrosion
experiment coupled with a cyclic wetting-drying process with 5%
NaCl solution and an indoor natural corrosion process was carried
out for several pre-cracked RC elements. The total duration of the
experiment was 4 years. The corroded bars were extracted from
these elements and the non-uniform reduction in the cross-
sectional area along the length of the bars were measured. A tensile
test for these corroded bars was completed and the influence of
corrosion degree and pitting effect on mechanical properties of
corroded bars were studied and discussed.
 C. Lu et al. / Construction and Building Materials 123 (2016) 649–660 651

2. Experimental programs Table 3

Mix design and strength of concrete.

2.1. Preparation of specimens W/C Water Cement Fine Coarse aggregate 28d cubic
ratio (kg/m3) (kg/m3) aggregate (kg/m3) compressive
(kg/m3) strength (MPa)
In our experiment, a steel type of HRB400 which has been com-
monly used in China for decades was used and five reinforced con- 0.43 185 429 536 1250 32.1
crete specimens of 150  100  1500 mm were prepared. The
notation HRB400 refers to a hot-rolled ribbed rebar (HRB) with a
yield strength not less than 400 MPa. Two HRB400 rebars, 16 mm laboratory with natural indoor curing up to 28 days. The 28-day
in diameter, were placed in each specimen with cover thicknesses compressive strength of concrete, measured on standard 150 mm
of 20 mm and 40 mm respectively and stretched out for certain cube, was 32.1 MPa (see Table 3). The detailed crack parameters
length at each end in order to evaluate corrosion state with a for all specimens are listed in Table 4. In the second column of
half-cell potential method, as shown in Fig. 1. In the concrete mix- Table 4, the crack parameters are crack width, space and depth
ture, a Grade 42.5 ordinary Portland cement (OPC) which can be in sequence. For example, the notation of 02-150-40 means the
approximately classified as Type I cement based on the ASTM cracks in that specimen are 0.2 mm in width w (02), 150 mm in
C150 standard [23] was used. The mineral and chemical compo- space (150) and 40 mm in depth (40).
nents of cement are shown in Table 2. Crushed gravel and natural
river sand were used as coarse aggregate and fine aggregate, 2.2. Natural corrosion experiment
respectively. The maximum size of the gravel was 16 mm, and
the fineness modulus of sand was 1.62. On the basis of the cubic Just before natural corrosion experiment, the stretched steel
strength grade 30 MPa, the design mix of the concrete expressed sections at two ends of the specimen should be coated with epoxy
with the ratios to cement is Cement:Water:Gravel: resin after they were connected with cables. Then all specimens
Sand = 1:0.43:2.91:1.25. The detailed mix contents of concrete were placed into an indoor pool with 5% by weight NaCl solution
were given in Table 3. and the procedures of natural corrosion were divided into two
In order to simulate cracking induced by mechanical loading in stages:
RC beams, which is usually inevitable under service condition, and
accelerate the processes of chloride penetration and steel corro-
Stage I for 0–3 years: a mechanism of drying–wetting cycles
sion, except one reference member, another four specimens were with 7 days wetting and 7 days drying in one cycle was applied
artificially pre-cracked at the cover side with a piece of foil, which to all specimens in order to accelerate the corrosion process (see
has a thickness of 0.1 or 0.2 mm. Two crack spaces of 150 mm and Fig. 2a). The annual average temperature and RH in the lab were
250 mm were used. The widths and spaces of pre-cracks are in 16.3 °C and 76%, respectively. After nearly 2 years, a half-cell
accordance with the code requirements in chloride environment potential method with CANIN+ equipment was periodically
[24–26]. This work of pre-cracking in concrete was done after
the vibrating and flatting procedures were finished. Several pieces
Table 4
of foil with same thickness were slowly inserted into the cover
Crack parameters and steel numbers for all specimens.
concrete (perpendicular to the surface) up to the designed depth
(see Fig. 1) and then the area near the foil was flatted again. About No. of Crack No. of steel Cover thickness,
Specimens parameters bars c/mm
6 h later, these foils were slowly pulled out from concrete (also
perpendicular to the surface). Here, it should be stated that the 1 Uncracked 1-B20 20
1-B40 40
mechanism of pre-cracking in concrete is only an approach to
2 02-250-40 2-B20 20
accelerate steel corrosion, and that the quantitative effect of cracks 2-B40 40
on steel corrosion is beyond the scope of this project. Therefore, the 3 02-150-40 3-B20 20
reduce of crack width and depth caused by self-healing was not 3-B40 40
focused and considered. 4 01-150-40 4-B20 20
4-B40 40
All of the mixing and casting works were carried out in a stan-
5 02-150-20 5-B20 20
dard laboratory at 20 ± 2 °C and 50 ± 5% RH. After demolding, all 5-B40 40
specimens were moist-cured for 7 days and then placed in the

Fig. 1. Layout of reinforced concrete specimens (unit: mm).

Table 2
Mineral and chemical composition of cement.

Mineral composition Chemical composition

Type C3S C2S C3A C4AF CaO SiO2 Al2O3 Fe2O3 MgO SO3
Content (%) 61.4 11.6 8.7 10.2 64.6 20.9 5.43 3.37 1.58 1.85
652 C. Lu et al. / Construction and Building Materials 123 (2016) 649–660

(a) Drying-wetting cycle (b) Watering every week

Fig. 2. Specimens in natural corrosion processes: (a) drying-wetting cycle and (b) watering every week.

(a) Rust locations (b) One piece of bar 3-B20

Fig. 3. Samples of corroded steel bars: (a) rust locations and (b) one piece of bar 3-B20.

used to evaluate the corrosion state of steel bars [27]. The Table 5
experiment was diverted to next stage when the active corro- Cross section losses and tensile test results of steel bars.
sion was verified at the end of the third year. No. of xmax/ gavg/ xavg/ fyc/ fuc/ D/ W/

Stage II for the 4th year: all specimens were removed from the steel bar mm % mm MPa MPa mm kJ
solution pool and placed in a shady place (see Fig. 2b). The spec- Reference 0 0 0 471* 680* 38.2 5.34
imens were kept humid with watering every week until the 1-B20 0.418 2.43 0.195 482 665 28.8 4.13
experiment ended. 0.465 2.75 0.222 468 674 25.4 3.59
0.471 2.81 0.226 475 677 32.1 4.40
1-B40 0.267 2.30 0.185 412 576 37.2 4.86
The corroded steel bars were obtained by breaking all speci- 0.327 1.87 0.165 423 579 34.1 5.22
mens (see Fig. 3a). Based on the distribution of corrosion pits, each 0.526 2.82 0.227 417 572 24.7 4.61
corroded bar was cut into 3 pieces of average length 500 mm and 2-B20 0.487 3.77 0.305 461 664 34.2 5.26
total 30 samples were prepared. Then all steel samples were 0.826 3.78 0.306 457 668 31.6 5.37
0.306 3.55 0.286 453 665 34.5 4.75
cleaned according to the treatment method provided in ASTM
2-B40 0.626 5.06 0.410 422 556 25.0 3.77
G1-03 [28] in order to remove the rust. Some typical corroded bars 1.232 4.45 0.360 437 639 25.3 4.33
before and after cleaning are shown in Fig 3b. 1.231 4.53 0.366 430 645 28.7 5.16
3-B20 1.423 4.36 0.352 404 562 27.6 4.30
1.987 4.92 0.399 405 551 21.7 3.80
2.3. Measurements of diameter loss and corrosion degree 1.348 5.77 0.468 396 545 31.2 3.75
3-B40 0.731 4.78 0.387 434 641 32.5 4.94
0.510 5.24 0.425 440 637 33.7 5.27
The residual diameter, dres, of corroded steel bar was measured
1.245 4.54 0.367 426 638 31.0 5.26
with the help of a microcalliper after complete removal of the rust. 4-B20 0.719 3.71 0.299 454 662 35.4 5.88
As stated in literature [19], it was very difficult to assess the precise 0.689 3.54 0.285 452 658 29.9 4.65
diameter in this way since the surface shape of the corroded bar 0.916 3.49 0.282 461 681 31.6 4.85
varied substantially. Therefore, at every section, the measurement 4-B40 0.481 4.10 0.331 469 678 29.8 4.97
0.290 3.80 0.307 473 682 33.3 5.01
should be made at least 3 times and a mean value was taken as the
0.553 3.44 0.278 479 673 32.2 5.68
final result. Considering the original diameter of sound (uncor- 5-B20 1.176 4.12 0.333 406 561 33.5 3.69
roded) steel bar, the maximum pit depth, xmax, of each sample is 0.347 3.91 0.316 412 567 28.4 3.75
obtained and listed in Table 5. Here, it should be noted that the 0.917 4.10 0.332 410 560 24.6 3.63
diameter of the ribbed bar stated in the code is the nominal diam- 5-B40 0.493 4.55 0.368 408 565 36.6 5.03
0.630 3.69 0.298 414 567 33.5 4.35
eter, dnom, which takes the effect of ribs into account and is greater 0.373 4.06 0.328 405 562 33.0 4.56
than the effective inner diameter, dinn. For example, the dnom value
*
of the bar used here is 16 mm, however its dinn value is 15.8 mm. fy0 = 471 MPa, fu0 = 680 MPa.
 C. Lu et al. / Construction and Building Materials 123 (2016) 649–660
The degree of corrosion of the bars was quantified in this study
by average mass loss, defined as the ratio of the mass difference of
the bar before and after corrosion to its original mass. This defini-
tion method of corrosion level may not be able to distinguish
the difference between uniform corrosion and pitting corrosion,
however, many researchers [10–15] still believed it a practical
approach to assess the effect of corrosion on the tensile behaviors
of steel bars. The main reasons may be ascribed to two following
facts: i) in engineering practice it is very hard to determine the
location of the critical corrosion pit and calculate the minimum
cross-sectional area of corroded bars [15]; ii) the average mass loss
of corroded bars can be easier predicted by Faraday’s law with the
measured corrosion current density in service structures [11,12].
All cleaned bars were weighed using an electronic scale with a pre-
cision of 0.01 g. The corrosion degree, i.e., average mass loss gavg
(%), of the corroded bars can be calculated from Eq. (1) by consid-
ering the difference of bars’ lengths. Based on the value of gavg, the
average diameter loss of corroded bars, xavg, can be assumed with
Eq. (2). The total results of gavg and xavg are given in Table 5.
m0 =l0  mc =lc
gavg ¼  100%
m0 =l0
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

xavg ¼ dnom 1  1  gavg =100
where m0 and mc are the masses of the sound and corroded bars and
l0 and lc are the lengths of the sound and corroded bars, respec-
tively; dnom is the nominal diameter of original sound steel bar.
2.4. Tensile tests
Tensile tests of all sound and corroded steel bars were carried
out using a CSS-44300 electronic universal testing machine, which
was produced by Changchun Research Institute for Mechanical
Science (see Fig. 4). Before testing, on the bar surface marks with
an initial distance of ten times the nominal diameter of bars, i.e.,
10dnom = 160 mm, were made to determine the original gauge
length, L0. The tensile process was controlled by displacement with
2 mm/min. During the tensile test, the applied load and deforma-
tion data were recorded using a computerized data acquisition sys-
tem. The data obtained were utilized to plot load-displacement
curve, determine the yield load and ultimate load, and calculate
(a) CSS-44300 machine
Fig. 4. Tensile test setup: (a) CSS-44300 machine and (b) bar’s anchorage.
653
the elongation for each tested bar. Besides, the energy consumed
in the tensile process up to failure of the sample was also recorded.
Here, based on the national standard GB/T 228.1-2010 [29],
which referred to international standard of ISO 6892-1: 2009, the
percentage elongation after fracture was evaluated as the ratio of
the extension, D, between two marks nearest to the fracture area
to the initial length, which can be expressed as:
D Lu  L0
A¼ ¼  100% ð3Þ
L0 L0
where A (%) is the percentage elongation with original gauge length
pffiffiffiffiffi
L0 ¼ 11:3 S0 ¼ 10dnom [29], S0 is the original cross-sectional area of
steel bar; Lu is the final gauge length of the bar after fracture.
3. Results and discussions
3.1. Loss of diameter for corroded bars
Taking the influence of cover thickness and crack parameters
into account, Fig. 5 shows the measurements of diameter loss of
ð1Þ part corroded bars in a comparative pattern. The vertical dash lines
in the figures denote the locations of cracks along the length. It can
be clearly seen from Fig. 5 that the diameter losses of the corroded
ð2Þ bars in concrete are greatly affected by cover thicknesses and crack
parameters with width, space and depth. Specially, the obvious
diameter losses can be found at the locations where cracks exist.
Under the same condition of crack parameters, increasing cover
thickness can reduce the loss of diameter, as shown in Fig. 5a. From
Fig. 5b, c and d, it can be concluded that shortening crack space,
increasing crack width and depth may effectively cause deeper
losses of bars’ diameter when other conditions keep same. Besides,
the area of localized corrosion at crack zone will be augmented
when the crack runs through the steel bar, which may extend pit-
ting corrosion to uniform corrosion gradually. The reason can be
ascribed to the fact that all aggressive mediums (Cl, O2, H2O,
etc) will be speeded to penetrate into inner steel-concrete interface
through crack surfaces [30,31].
In conclusion, the corrosion of steel bars in cracked concrete can
be regarded as a typical localized corrosion, although corrosion
area usually be extended by crack effect. As a result, the character-
istics of non-uniform corrosion of steel bar along its longitudinal
length should be taken into account when the mechanical properties
are considered.
(b) Bar’s anchorage
654 C. Lu et al. / Construction and Building Materials 123 (2016) 649–660

16 15 30 45 60 75 90 105 120 135

16

Residual diameter (mm)

Residual diameter (mm)

15.5
15.5
15
15
14.5
14.5
14 14
13.5 13.5
3-B20 3-B40 1-B40 2-B40 3-B40
13 13
0 15 30 45 60 75 90 105 120 135 150 0 25 50 75 100 125 150
Length of the bar (cm) Length of the bar (cm)
(a) Different cover thickness (w=0.2mm) (b) Different crack spaces (c=40mm)

16 16
Residual diameter (mm)

15.5

Residual diameter (mm)

15.5

15 15

14.5 14.5

14 14

13.5 13.5
1-B20 3-B20 4-B20 1-B20 3-B20 5-B20
13 13
0 15 30 45 60 75 90 105 120 135 150 0 15 30 45 60 75 90 105 120 135 150
Length of the bar (cm) Length of the bar (cm)
(c) Different crack widths (c＝20mm) (d) Different crack depths (c＝20mm)
Fig. 5. Distribution of residual diameters of corroded steel bars: (a) different cover thickness (w = 0.2 mm), (b) different crack spaces (c = 40 mm), (c) different crack widths
(c = 20 mm) and (d) different crack depths (c = 20 mm).

3.2. Yield and ultimate strengths As stated in the introduction, the tensile strengths of corroded
bars from the measured forces can be calculated based on the
3.2.1. Test results nominal steel area [7–10,12–14,19,32], the average residual steel
The tensile test results of the load-deformation curves of typical area [11,15–17] and the residual minimum steel area [16,17,19].
corroded rebars with different corrosion degrees are shown in Until now, there is no consensus among researchers regarding
Fig. 6. It can be seen that the maximum deformation decreased which method is more appropriate than the others [32]. Besides,
gradually with the increasing of corrosion degree. The results of according the valid standards, there is no special consideration
yield and ultimate loads of corroded bars presented a slightly for the reduction of the nominal diameter of the reinforcing bars,
decreasing trend when the degrees of corrosion varied within 6%. so the reduction of steel area due to corrosion should be neglected
In general, these observations are in accordance with previous in this way [9,10]. Therefore, in this study, the yield and ultimate
experimental studies [7–17,19]. strengths of corroded bars were calculated with the nominal
cross-sectional area of original steel bar. The corresponding results
of nominal yield strength, fyc, and nominal ultimate strength, fuc, of
160 corroded bars are given in Table 5. Besides, the nominal yield and
ηavg=0% ultimate strengths of sound bar are listed in the second row of
140
Table 5 with gavg = 0.
Tensile load (kN)

120 As given in Table 5, almost all nominal yield and ultimate

100 ηavg=2.06% strengths of corroded bars with gavg less than 6% can satisfy
80 ηavg=5.77% the strength minimums of fy = 400 MPa and fu = 540 MPa required
by design code [25] for HRB400 bars, respectively. The correlation
60 ηavg=3.69% of fuc/fyc ratio and average mass loss was presented in Fig. 7. It
40 can be found that it is difficult to obtain a direct relationship
20 between them because of the large scatter observed. However,
0 there is a tendency for the fuc/fyc ratio to slightly decrease as
0 10 20 30 40 50 60 70 the average mass loss gavg increases up to 6%. According the code
Axial deformation (mm) requirement in seismic design that the fuc/fyc ratio of deformed
steel bars should be more than 1.25 [25] (the same requirement
Fig. 6. Load–displacement curves of typical corroded bars. can be found in ACI 318 code [26]), it can be concluded from
 C. Lu et al. / Construction and Building Materials 123 (2016) 649–660 655

1.6
1.0
1.5
0.8 10% scatter
1.4 range
fuc/fyc ratio

fyc/fy0 ratio
0.6
1.3 fyc/fy0 =1-0.0123ηavg[32]
0.4 fyc/fy0 =1-0.0198ηavg[14]
1.2 Limitation of 1.25 fyc/fy0 =1-0.0303ηavg[9]
fyc/fy0 =1-0.0748ηavg[10]
0.2 fyc/fy0 =1-0.0195ηavg[this study]
1.1

1.0 0.0
0 1 2 3 4 5 6 7
0 1 2 3 4 5 6 7
Average mass loss, ηavg (%)
Average mass loss, ηavg (%)
(a) Nominal yield strength
Fig. 7. Relationship between fuc/fyc ratio and gavg.

1.0
Fig. 7 that all sound and corroded bars got from our experiment
can satisfy this limitation.
0.8
15% scatter

fuc/fu0 ratio
3.2.2. Effect of degree of corrosion on tensile strength of corroded bar range
0.6
In terms of the degree of corrosion, gavg, many researchers
[12–14,32] proposed empirical linear correlations between the fuc/fu0 =1-0.0115ηavg [32]
0.4
nominal yield (or ultimate) strength of corroded bar and that of fuc/fu0 =1-0.0157ηavg [14]
sound bar from their experimental results. Such relationships fuc/fu0 =1-0.0284ηavg [9]
may be represented by the expressions given in Eqs. (4) and (5). 0.2 fuc/fu0 =1-0.0231ηavg [this study]

f yc ¼ ð1:0  ay gavg Þf y0 ð4Þ 0.0

f uc ¼ ð1:0  au gavg Þf u0
where fyc and fuc are the yield and ultimate strengths of corroded
bars, which are calculated based on the original nominal cross sec-
tional area; fy0 and fu0 are the yield and ultimate strengths of sound
bars, respectively; ay and au are regressive constants got from
experimental results. For steel bars corroded in concrete, these
two constants were collected from previous studies in Table 6.
Based on the nominal yield and ultimate strengths of sound bar,
that is 471 MPa and 680 MPa respectively, the normalized ratios of
yield strength and ultimate strength of corroded bars, that is fyc/fyo
and fuc/fuo, were calculated based on the experimental results
shown in the fifth and sixth columns of Table 5, as given in
Fig. 8. Linear regressive analysis of these normalized ratios was
performed based on Eqs. (4) and (5), and the outcomes were added
in Table 6 and Fig. 8. Besides, in order to get a direct comparison
with previous studies, some empirical models [9,10,14,32] got
0 1 2 3 4 5 6 7
ð5Þ Average mass loss, ηavg(%)
(b) Nominal ultimate strength
Fig. 8. Comparison of normalized tensile strengths between predictions and
experimental results: (a) Nominal yield strength and (b) nominal ultimate strength.
from natural or artificial chloride conditions in Table 6 were also
plotted in Fig. 8. It can be seen that the normalized ratios of nom-
inal yield and ultimate strengths of corroded bars decrease gradu-
ally with the increase of average mass loss. Comparing the test data
with the empirical curves given by other authors [9,10,14,32], it
can be found that our experimental results are consistent with
the predictions in these studies, although the bar diameter and
the degree of corrosion exist certain difference in all studies.
Besides, it is noticed that the predictive curve plotted from Apos-
tolopoulos et al.’s study [10] shows a fast fall of nominal yield

Table 6
Regressive constants of ay and au for steel bars corroded in concrete.

Authors Bar type Diameter/mm Corrosion condition gavg/% ay au

Cairns et al. [13] Plain 16 Electrical 0–4 0.0120 0.0110
Deformed 20 (10–50 lA/cm2)
Lee and Cho [14] Deformed/SD 295A, 345 13 Electrical 0–35 0.0124 0.0107
Xia et al. [12] Deformed 16, 20 Electrical 0–15 0.0210 0.0210
(200 lA/cm2)
Ou et al. [32] Deformed 13, 29 Electrical 0–31 0.0127 0.0116
(600 lA/cm2)
Lee and Cho [14] Deformed/SD 295A 10, 13 Salt + wetting and drying 0–35 0.0198 0.0157
Papadopoulos et al. [9] Deformed/B St I 8, 10 Natural chloride 8–70 0.0072* 0.0086*
(100 years)
Deformed/B St III 8, 10 Natural chloride 4–16 0.0303* 0.0284*
(30 years)
Ou et al. [32] Deformed 13, 16, 19 Natural chloride (built in the 1970 s) 6–82 0.0123 0.0115
Apostolopoulos et al. [10] Deformed 8 Artificial chloride salt spay (1 year) 0–4 0.0748* N/S
This study Deformed 16 Artificial chloride solution (4 years) 0–6 0.0195 0.0231
*
Data got from Ou et al. [32]; N/S = not stated.
656 C. Lu et al. / Construction and Building Materials 123 (2016) 649–660

2.5

Maximum pit depth, xmax (mm)

2.0

1.5
At steel body
1.0

0.5

0.0
0 1 2 3 4 5 6 7
Average mass loss, ηavg (%)
Near longitudinal rib
Fig. 10. Relationship between xmax and gavg.

two parameters can be found. Tang et al. [16] proposed a

Near transverse rib
Fig. 9. Various pitting locations at corroded bars.
strength ratio with increasing corrosion degree. One possible rea-
son is that the regressive constant of ay = 0.0748 established by
another authors [32] was obtained based on minor experimental
data (only ten).
Obviously, the two regressive constants of ay and au proposed
in this study are within the range of those in Table 6 got from nat-
ural or artificial chloride conditions. Comparing the values of the
regressive constants from natural or artificial chloride conditions
and from applied electrical conditions (in the second to fifth rows
of Table 6), it can be observed that the former values are greater
than the latter values, which means that the effect of natural cor-
rosion on degradation of tensile behaviors of corroded bars is
higher than that caused by accelerated corrosion with impressed
current [14,15]. It is in accordance with the fact the pitting corro-
sion of bars embedded in concrete is more dominant under natural
corrosion condition caused by chloride attack (see Fig. 9). Further-
more, this fact is still an important factor which induces a 10–15%
scatter range between our predictions and experimental data (see
Fig. 8) when the average mass loss is used. From Fig. 9, it can be
still found that the mass losses caused by the corrosion occurred
at ribs are inevitably included in the calculations of corrosion
degree. These corrosion losses, however, would scarcely affect
the tensile behaviors of corroded bars, and as a result there exists
cross-sectional spatial heterogeneity factor, R, given in Eq. (6),
and used this factor as an index of cross section variation, which
was also discussed by Zhang et al. [17].
R ¼ Aavg =Amin ð6Þ
where Aavg is the average cross-sectional area of corroded steel
specimen; Amin is the minimum cross-sectional area of the same
corroded steel specimen.
Assuming the average mass loss approximately equal to the
average cross-sectional area loss of corroded bars, the factor R
can be expressed with gavg and gmax, as shown in Eq. (7). Here, gmax
is the maximum cross-sectional area loss of corroded bar. Based on
Eqs. (6) and (7), the calculations of heterogeneity factor, R, were
gained from the experimental results given in these studies
[16,17], and the correlation of R and gavg are plotted in Fig. 11. It
can be seen that for steel bars corroded in concrete by impressed
current, the heterogeneity factor R is mainly in the range of 1.0–
1.3 when gavg is less than 30%. Therefore, it can be found from
two studies [16,17] that this spatial effect of corrosion distribution
on tensile strengths of corroded bars was limited by comparing the
two strength results calculated with gavg and gmax, respectively.
The main reason for this consequence may be ascribed to the fact
that the corrosion in concrete caused by impressed current is a typ-
ical uniform corrosion, which may induce relatively uniform
reduction in bar area.
100  gavg
R¼ ð7Þ
100  gmax
1.6
Data from Tang et al. [16]
Heterogeneity factor, R

1.5
Data from Zhang et al. [17]
certain error in the calculated average mass loss. Therefore, except
1.4
for the corrosion degree, the effect of the non-uniform distribution
of pits on strength degradation of corrosion bars should not be 1.3
ignored, and it will be discussed in the next section. 1.2
1.1
3.2.3. Effect of non-uniform corrosion property on tensile strengths
Considering the fact that the non-uniform distribution of corro- 1
sion pits or the non-uniform reduction in the cross-sectional area 0.9
along the length of the bars, the specimens with same or adjacent
0.8
average mass loss may appear great difference in the amounts and 0 5 10 15 20 25 30
depths of corrosion pits. Fig. 10 shows the relationship between Average mass loss, ηavg (%)
the maximum pit depth, xmax, and average mass loss, gavg, of our
test specimens. As expected, no direct correlation between above Fig. 11. Relationships between R and gavg obtained from previous studies [16,17].
 C. Lu et al. / Construction and Building Materials 123 (2016) 649–660 657

In our experiment in this study, however, the corrosion of steel
bars in pre-cracked concrete suffering from chloride attack has
been testified to be a typical pitting corrosion from visual observa-
tions and testing results. Moreover, taking the influence of stress
concentration at pit location on the tensile fracture of corroded
bars into account, a new non-uniform coefficient, S, of longitudinal
corrosion distribution was proposed in this study, and it was estab-
lished directly based on diameter loss, not on area loss, of corroded
bars. In order to eliminate the effect of bar’s diameter, the defini-
tion of S would be described as the ratio of the maximum percent-
age diameter loss to the average percentage diameter loss of
corroded rebar specimens, which can be expressed as
qmax
S¼
qavg
where, qmax is the measured maximum percentage diameter loss
(%), qmax = xmax/dnom  100%; qavg is the average percentage diame-
ter loss (%), qavg = xavg/dnom  100%.
Based on the values of xmax and xavg listed in Table 5, the max-
imum and average percentage diameter losses of qmax and qavg in
our experiment were 1.68–12.42% and 1.03–2.93%, respectively. It
can be found that the former is greater than the latter. Accordingly,
the calculated value of non-uniform coefficient S varied between 1
and 5. Fig. 12a and b show the relationships between S and qavg,
and between S and qmax, respectively. It can be observed that there
is no obvious correlation of S and qavg because the parameter qavg
can’t reflect the non-uniform distribution of pitting corrosion. It is
found that, however, there is a distinct linear relationship between
6 several severe corrosion locations of RC element.
Non-uniform coefficient, S
S and qmax, and their regressive expression can be given in Eq. (9)
and shown in Fig. 12b.
S ¼ 0:351qmax þ 0:681 ð1:67% 6 qmax 6 12:42%Þ ð9Þ
It can be reasonably realized that the parameter S can be used to
describe the level of non-uniform corrosion along the length of cor-
roded rebar. The larger the value of S is, the greater the effect of
non-uniform reduction of bar’s diameter on degradation of tensile
strength might become. As a result, the strength ratios of fyc/fyo and
fuc/fuo of corroded bars may reduce not only with increasing the
average mass loss gavg, but also with increasing the non-uniform
coefficient S, as shown in Fig. 13. Similar to Eqs. (4) and (5), two
regressive linear relationships between the tensile strengths and
the non-uniform coefficient can be proposed as:
ð8Þ
f yc ¼ ð1:0  by SÞf y0 ð10Þ
f uc ¼ ð1:0  bu SÞf u0 ð11Þ
where by and bu are the regressive constants and by ¼ 0:030,
bu ¼ 0:035 were obtained for nominal yield and ultimate strengths,
respectively.
Comparing Fig. 13 with Fig. 8, it can be concluded that the pre-
cision of predictions from Eqs. (10) and (11) is better than that
from Eqs. (4) and (5). So this method might be an effective way
to evaluate the mechanical properties of pitting rebars. In order
to testify the applicability of Eqs. (10) and (11) and use them fur-
ther, the value of non-uniform coefficient S is advised to be got
from two following ways under different conditions. For experi-
mental studies, it can be calculated with Eq. (8) based on measured
pitting depths of bar specimens. For field investigations, however,
it can be predicted with Eq. (9) based on the inspective results at

4 1.0

3 0.8
fyc/fy0 ratio

0.6 fyc/fy0=1-0.030×S
2
R2=0.8236
1 0.4

0 0.2
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
0.0
Average percentage diameter loss, ρavg ( ) 0 1 2 3 4 5 6
(a) S and ρavg Non-uniform coefficient, S
(a) Nominal yield strength
6
Non-uniform coefficient, S

5 1.0

4 0.8
fuc/fu0 ratio

3 0.6 fuc/fu0=1-0.035×S
R2=0.7517
2 S=0.351×ρmax+0.681 0.4
R2=0.8573
1 0.2

0 0.0
0 2 4 6 8 10 12 14 0 1 2 3 4 5 6
Maximum percentage diameter loss, ρmax (%) Non-uniform coefficient, S
(b) S and ρmax (b) Nominal ultimate strength
Fig. 12. Relationships between non-uniform coefficient and percentage diameter Fig. 13. Relationships between normalized strength ratio and non-uniform coef-
loss: (a) S and qavg and (b) S and qmax. ficient: (a) Nominal yield strength and (b) nominal ultimate strength.
658 C. Lu et al. / Construction and Building Materials 123 (2016) 649–660

3.3. Elongation and ductility
3.3.1. Test results
The measured elongations, D, after fracture (original gauge
length L0 = 160 mm) were recorded in the seventh column of
Table 5. Based on Eq. (3), the percentage elongations after fracture,
Ac, which can be used to evaluate the ductility property of corroded
bars, were obtained and the results were shown in Fig. 14. It can be
seen that the percentage elongations of corroded bars decline obvi-
ously as the average mass loss gavg increases up to 6%. According
the limit requirement given in the code [29] that the percentage
elongation of HRB400 steel bars should be more than 16% for
HRB400 rebars, it can be found from Fig. 14 that the percentage
elongations drop gradually close to the limitation as gavg increases
and even few tested specimens have a percentage elongation
below the minimum value of 16%. This outcome indicates that at
the same corrosion degree the degradation of ductility of corroded
bars is much greater than that of tensile strengths, which had been
observed and testified by many researchers [7–19], as shown in
Table 1.
The energy, W, consumed in the tensile-fracture process is
another indice to reflect the ductility property of reinforcing steel
bars [8,16]. Fig. 15 presents the relation between recorded energies
and the degrees of corrosion of corroded bars. As expected, a nota-
ble decline of W can be observed when the mass loss gavg increases
up to 6%, which is similar to the phenomenon of elongations found
in Fig. 14. This outcome is consistent with the result found by other
researchers [16].
3.3.2. Effect of non-uniform corrosion property on ductility
Based on the percentage elongation of sound bar, that is
A0 = 23.88%, the normalized ratio of percentage elongation of cor-
roded bars, Ac/Ao, was calculated and its correlation with average
mass loss gavg was plotted in Fig. 16a. A linear empirical model pro-
posed by Lee and Cho [14] under the condition of artificial natural
corrosion was compared with our experimental results in Fig. 16a.
It can be seen that the descent speed of experimental data is
greater than the predictive curve given by Lee and Cho [14] as
the average mass loss increases. Besides, some researchers
[16,17] had advised to adopt exponential function to describe the
effect of corrosion degree on ductility. In order to compare with
above studies, the linear and exponential regressive analysis was
tried to conducted in these data, although the large scatter was
visually observed. The corresponding outcomes were added in
Fig. 16a. It can be observed that two linear and exponential func-
tions have similar regression results for our experimental data
and that both of them have a nearly 25% scatter range between
predictions and experimental data.
Similar to strength analysis, the effect of non-uniform coeffi-
cient S on the ductility degradation of corroded bar was also dis-
cussed in this study. Based on the calculated values of S with Eq.
(9), the relationship between Ac/Ao and S was given in Fig. 16b.
Comparing the data distributions in Fig. 16a and b, it can be seen
that the tendency for Ac/Ao ratios to decline is more obvious as
the non-uniform coefficient S increases. At same time, a smaller
scatter of data is also found in Fig. 16b. A linear regressive expres-
sion between the Ac/Ao ratio and S was proposed as Eq. (12) and the
outcome was also shown in Fig. 16b.

30
Percentarge elongation, Ac (%)

1.0
25
0.8
20
Ac/A0 ratio

0.6 25% scatter

15 Ac/A0=exp(-0.059×ηavg) range
Limitation of 16%
10 0.4
Ac/A0=1-0.0259×ηavg[14]

5 0.2 Ac/A0=1-0.046×ηavg[this study]

0 0.0
0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7
Average mass loss, ηavg (%)
Average mass loss, ηavg (%)
Fig. 14. Relationship between Ac and gavg.
(a) Ac/A0 ratio and ηavg

1.0
7.0
0.8
6.0
Ac/A0 ratio

0.6
Energy, W (kJ)

5.0

4.0 0.4 Ac/A0=1-0.078×S

R2=0.6513
3.0
0.2
2.0
0.0
1.0 0 1 2 3 4 5 6
Non-uniform coefficient, S
0.0
0 1 2 3 4 5 6 7 (b) Ac/A0 ratio and S
Average mass loss, ηavg (%)
Fig. 16. Effect of corrosion properties on Ac/A0 ratios: (a) Ac/A0 ratio and gavg and (b)
Fig. 15. Relationship between consumed energy and average mass loss. Ac/A0 ratio and S.
 C. Lu et al. / Construction and Building Materials 123 (2016) 649–660
Fibrous
zone
(a) Sound steel bar
Fig. 17. Fracture surface morphologies for sound and corroded steel bars: (a) Sound steel bar and (b) corroded steel bar.
80
y = 68.295 -3.507x
70
R² = 0.9545
Necking area (%)
60
50
40
30
20
10
0 watering for one year. On the basis of the appearance inspection
0% 2.06% 3.69% 5.77%
Average mass loss, ηavg (%)
Fig. 18. Percentage necking area of typical steel bars.
Ac ¼ ð1:0  bA SÞA0
where Ac and A0 are the percentage elongations for corroded and
sound bars, respectively; bA is the regressive constant with
bA ¼ 0:078 in this study.
3.4. Reduced area at necking and fracture surface morphology
The reduced area at necking after the failure of steel bars is
another way to demonstrate the loss of ductility. Here, the surface
morphology of corroded bars was measured to determine the final
cross-sectional area at the fracture point (see Fig. 17). The percent-
age necking area was calculated as the ratio of the difference
between the initial cross-section and the final cross-section to
the initial cross-section. The results of few typical corroded rebars
with different corrosion degrees were shown in Fig. 18. It can be
observed that the reduced area at necking decreases linearly with
the increase of average mass loss. This similar conclusion can be
found in other studies [19].
In Fig. 17, the representative fracture surface morphologies of
sound and corroded steel bars were shown and compared. For
the fracture surface of sound bar shown in Fig. 17a, an obvious
necking behavior can be observed and the shear surface with about
45° angle (described with circle line with dashes) can also be seen
at the edge of the fracture surface. Its remainder surface has a
fibrous zone in the color of grey. For the fracture surface of cor-
roded bar given in Fig. 17b, however, the necking behavior is influ-
enced by the corrosion pit at the fracture surface and no shear
surface is found. The fracture originated at the corrosion pit
(described with circle line with dashes) and propagated to the
remaining cross section, which was indicated by its radiation area
659
Corrosion pit
Fracture
origin
Shear
surface Radiation area with
fan-shaped ridges
(b) Corroded steel bar
with fan-shaped ridges. All similar phenomena were also found
and recorded by other investigators [16].
4. Conclusions
In this paper a 4-year natural corrosion experiment was carried
out for several pre-cracked reinforced concrete elements with dif-
ferent crack parameters. The processes of reinforcement corrosion
in concrete were divided into two stages: (i) nearly 3-year drying-
wetting cycles of 5% NaCl solution; and (ii) natural corrosion with
and tensile testing, the following conclusions can be drawn from
this study:
(1) From the measured results of the non-uniform reduction in
the cross-sectional area along the length of the corroded
ð12Þ bars, it can be found that the diameter losses are greatly
affected by cover thicknesses and pre-crack parameters
(width, space and depth). Specially, the obvious diameter
losses can be found at the locations where cracks exist. As
a result, a typical localized corrosion (pitting corrosion)
was achieved on the bar specimens, although corrosion area
usually be extended by crack effect.
(2) In our experiment, the degree of corrosion of the corroded
bar samples evaluated by average mass loss, gavg, was in a
range from 1.87% to 5.77%. Although the average corrosion
level of the corroded bars was limited for our natural corro-
sion condition, the maximum percentage diameter loss,
qmax, expressed with the ratio of maximum pit depth to
nominal bar diameter was greater than mass loss due to pit-
ting effect and it reached up to 12.42%.
(3) Based on the tensile test results, it is found that the nominal
yield and ultimate strengths drop linearly with the increase
of corrosion degree. Linear regressive analysis of strength
degradation was completed and the results were compared
with previous studies. It can be concluded that the effect
of natural pitting corrosion on degradation of tensile behav-
iors of corroded bars is higher than that of uniform corrosion
caused by impressed current. Besides, the ratio of ultimate
strength to yield strength of corroded HRB400 bars in this
study can nicely satisfy the minimum code requirement of
1.25.
(4) Both the percentage elongation and total energy reduce
obviously with the increase of average mass loss. Specially,
some data of percentage elongation of corroded HRB400
bars were found to be less than the minimum code value
of 16%. By comparison, it can be concluded that the effect
660 C. Lu et al. / Construction and Building Materials 123 (2016) 649–660
of corrosion on the degradation of ductility is much greater
than that of tensile strengths.
(5) A new non-uniform coefficient, S, which was established
directly based on diameter loss, was proposed to describe
the mechanical properties of corroded bars. The value of S
is in a range of 1–5 in this study. It is found that there is a
linear relationship between S and qmax. The degradation
models of yield strength, ultimate strength and percentage
elongation were derived based on the coefficient, S. It can
be preliminarily concluded from our project that this
method is an effective way to evaluate the mechanical prop-
erties of pitting rebars.
Acknowledgements
The financial supports from the National Natural Science Foun-
dation of PR China (Grant Nos. 51278230, 51378241 and
51578267), Jiangsu Overseas Research & Training Program for
University Prominent Young & Middle-aged Teachers and Presi-
dents (2015) and the ‘six talent peaks’ project in Jiangsu Province
(Grant No. 2015-JZ-008) are greatly acknowledged. Besides, Prof.
Paulo J.M. Monteiro at University of California at Berkeley is also
appreciated for providing advices to this paper.
References
[1] P.K. Mehta, Durability of Concrete – Fifty Years of Progress? Proceeding of 2nd
International Conference on Durability, Montreal, Canada, 1991. pp. 1–31.
[2] O.E. Gjørv, Durability Design of Concrete Structures in Severe Environments,
second ed., CRC Press, Taylor & Francis Group, 2014.
[3] Z.J. Chen, Effect of Reinforcement Corrosion on the Serviceability of Reinforced
Concrete Structures, Master’s thesis, University of Dundee, UK, 2004.
[4] Y.P. Liu, Modeling the Time-to-Corrosion Cracking of the Cover Concrete in
Chloride Contaminated Reinforced Concrete Structures, Doctor’s thesis,
Virginia Polytechnic Institute and State University, USA, 1996.
[5] A.A. Almusallam, Effect of degree of corrosion on the properties of reinforcing
steel bars, Constr. Build. Mater. 15 (8) (2001) 361–368.
[6] C.H. Lu, W.L. Jin, R.G. Liu, Reinforcement corrosion-induced cover cracking and
its time prediction for reinforced concrete structures, Corros. Sci. 53 (4) (2011)
1337–1347.
[7] C.A. Apostolopoulos, M.P. Papadopoulos, S.G. Pantelakis, Tensile behavior of
corroded reinforcing steel bars BSt 500s, Constr. Build. Mater. 20 (9) (2006)
782–789.
[8] C.A. Apostolopoulos, V.G. Papadakis, Consequences of steel corrosion on the
ductility properties of reinforcement bar, Constr. Build. Mater. 22 (12) (2008)
2316–2324.
[9] M.P. Papadopoulos, C.A. Apostolopoulos, A.D. Zervaki, G.N. Haidemenopoulos,
Corrosion of exposed rebars, associated mechanical degradation and
correlation with accelerated corrosion tests, Constr. Build. Mater. 25 (8)
(2011) 3367–3374.
[10] C.A. Apostolopoulos, S. Demis, V.G. Papadakis, Chloride-induced corrosion of
steel reinforcement – mechanical performance and pit depth analysis, Constr.
Build. Mater. 38 (2013) 139–146.
[11] Y.G. Du, L.A. Clark, A.H.C. Chan, Residual capacity of corroded reinforcing bars,
Mag. Concr. Res. 57 (3) (2005) 135–147.
[12] X. Jin, W.L. Jin, Y.X. Zhao, L.Y. Li, Mechanical performance of corroded steel bars
in concrete, Struct. Build. 166 (2013) 235–246.
[13] J. Cairns, G.A. Plizzari, Y.G. Du, D.W. Law, C. Franzoni, Mechanical properties of
corrosion-damaged reinforcement, ACI Mater. J. 102 (4) (2005) 256–264.
[14] H.S. Lee, Y.S. Cho, Evaluation of the mechanical properties of steel
reinforcement embedded in concrete specimen as a function of the degree
of reinforcement corrosion, Int. J. Fract. 157 (2009) 81–88.
[15] W.P. Zhang, X.B. Song, X.L. Gu, S.B. Li, Tensile and fatigue behavior of corroded
rebars, Constr. Build. Mater. 34 (2012) 409–417.
[16] F.J. Tang, Z.B. Lin, G.D. Chen, W.J. Yi, Three-dimensional corrosion pit
measurement and statistical mechanical degradation analysis of deformed
steel bars subjected to accelerated corrosion, Constr. Build. Mater. 70 (2014)
104–117.
[17] W.P. Zhang, H. Chen, X.L. Gu, Tensile behaviour of corroded steel bars under
different strain rates, Mag. Concr. Res. 68 (3) (2016) 127–140.
[18] R. Palsson, M.S. Mirza, Mechanical response of corroded steel reinforcement of
abandoned concrete bridge, ACI Struct. J. 99 (2) (2002) 157–162.
[19] R. Francois, I. Khan, V.H. Dang, Impact of corrosion on mechanical properties of
steel embedded in 27-year-old corroded reinforced concrete beams, Mater.
Struct. 46 (2013) 899–910.
[20] Y. Yuan, Y. Ji, S.P. Shah, Comparison of two accelerated corrosion techniques
for concrete structures, ACI Struct. J. 104 (3) (2007) 344–347.
[21] D.M. Frangopol, D. Saydam, S. Kim, Maintenance, management, life-cycle
design and performance of structures and infrastructures: a brief review,
Struct. Infrastruct. Eng. 8 (2012) 1–25.
[22] G. Barone, D.M. Frangopol, M. Soliman, Optimization of life-cycle maintenance
of deteriorating bridges with respect to expected annual system failure rate
and expected cumulative cost, J. Struct. Eng. 140 (2) (2014) 04013043.
[23] American Society for Testing and Materials (ASTM), Standard specification
for Portland cement, C150/C150M-12, West Conshohocken, PA, 2012.
[24] British Standards Institution (BSI), Eurocode2 (EN1992): Part 1-1: general
rules and rules for building, London, 2004.
[25] National Standard of the People’s Republic of China, Code for design of
concrete structures, GB50010-2010, Beijing, 2010.
[26] American Concrete Institute (ACI) Committee 318, Building Code
Requirements for Structural Concrete (ACI 318-11) and Commentary,
Farmington Hills, MI, 2011.
[27] American Society for Testing and Materials (ASTM), Standard Test Method for
Half-Cell Potentials of Uncoated Reinforcing Steel in Concrete, C876-09, West
Conshohocken, PA, 2009.
[28] American Society for Testing and Materials (ASTM), Standard Practice for
Preparing, Cleaning, and Evaluating Corrosion Test Specimens, G1-03, West
Conshohocken, PA, 2011.
[29] National Standard of the People’s Republic of China, Metallic materials-Tensile
testing-Part 1: Method of test at room temperature (ISO 6892-1: 2009, MOD),
GB/T 228.1-2010, Beijing, 2010.
[30] T.U. Mohammed, N. Otsuki, M. Hisada, T. Shibata, Effect of crack width and bar
types on corrosion of steel in concrete, J. Mater. Civ. Eng. 13 (3) (2001) 194–
201.
[31] P.P. Win, M. Watanabe, A. Machida, Penetration profile of chloride ion in
cracked reinforced concrete, Cem. Concr. Res. 34 (7) (2004) 1073–1079.
[32] Y.C. Ou, Y.T.T. Susanto, H. Roh, Tensile behavior of naturally and artificially
corroded steel bars, Constr. Build. Mater. 103 (2016) 93–104.