You are on page 1of 8

Construction and Building Materials 132 (2017) 462–469

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage:

Effect of combined pre-cracking and corrosion on the method of repair of

concrete beams
Ali Siad a,⇑, Mohamed Bencheikh a, Luaay Hussein b
Geo-materials Development Laboratory (GMDL), M’Sila University, Algeria
Department of Civil Engineering, Ryerson University, Toronto, Canada

h i g h l i g h t s

 Effects of combined pre-cracking and corrosion on repair methods were studied.

 Two different methods were used to repair the damaged concrete beams.
 The properties of the repaired concretes were based on those of the strengthen system.
 The method of repair influenced considerably the results of the reinforced beams.
 Combining pre-cracking and corrosion influences the properties of the repaired beams.

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

Article history: This study investigated the combined effect of pre-cracking and corrosion on repair methods of rein-
Received 25 August 2016 forced concrete beams. The experimental program consisted of prismatic pre-cracked reinforced concrete
Received in revised form 28 November 2016 beam specimens, exposed to accelerated corrosion. The corrosion rate varied from 5% to 15% and the pre-
Accepted 7 December 2016
cracking was achieved by preloading the beams up to 60% of their ultimate loads. Two different methods
were used to repair the specimens. Carbon Fiber Reinforced Polymers plates (CFRP) were used to
strengthen the bottom portion of the specimens in the first method, and sheets containing carbon fibers
were added to the bottom portion, left and right sides of the specimens, in the second method. The test
results revealed that, the harmful effect of combined corrosion and pre-cracking was notable especially at
Repair 15% corrosion rate. The two methods of repair used can be considered as convenient. The ultimate
Damage strengths of the repaired beams were equivalent or significantly higher than those of the control speci-
Composite materials mens, although reduced deflection capacities were registered after repair.
Reinforced concrete Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction from corrosion gets destroyed [2]. Therefore, the embedded steel
reinforcement corrodes, which leads to the reduction in the
The corrosion of steel reinforcements can be considered as the cross-sectional area of the reinforcing steel [3], cracking, delamina-
primary cause of damage and early failure in concrete structures. tion and spalling of the concrete cover [4], and thus deteriorating
Indeed, it results in significant technical and financial problems the bond at steel-concrete interface [5,6].
in terms of maintenance and/or rehabilitation applications and The repair and rehabilitation of reinforced concrete structures is
re-construction or replacement of existing structures. For instance, only successful if the new material interact efficiently with the old
it is estimated that it costs around 1.8 trillion US$ to repair cor- concrete and form a durable barrier against the ingress of carbon
roded structures around the world annually [1]. dioxide and chlorides. A variety of methods have been developed
In normal conditions, the reinforcement steel is protected from to repair the corrosion damage of reinforced concrete (RC) beams,
corrosion because of the high alkalinity of concrete. When carbon- and strengthen them using plates and steel sheets, or by projecting
ation occurs, or when the concentration of chloride ions near the the damaged area with concrete containing fibers [7]. However,
steel reinforcement increases, the alkalinity of concrete is reduced. there are many disadvantages to these methods, including the
As a result, the passive layer that protects the steel reinforcement durability of the projected concrete as well as the heavy weight
and corrosion of the steel plates. In order to overcome these disad-
⇑ Corresponding author. vantages, fiber reinforced polymers (FRP) have been used in recent
E-mail address: (A. Siad). years to repair or strengthen the RC elements [8–11]. Among the
0950-0618/Ó 2016 Elsevier Ltd. All rights reserved.
A. Siad et al. / Construction and Building Materials 132 (2017) 462–469 463

common FRP techniques used to strengthen/repair the damaged Table 1

RC beams, externally bonded reinforcement (EBR) and near- Experimental parameters of tested beams.

surface mounted (NSM) have gained great interest and use [12– Beam Pre-cracking Corrosion % Mode of repair
15]. The EBR system involves the use of FRP sheets bonded in the PC0T1 No 0 No
external part of the element with an epoxy resin. While, in the PC0T2 No 0 No
NSM technique, FRP laminates are inserted into precut grooves in PC0T3 No 0 No
the concrete cover of the beams. Although, the use of FRP tech- PC5 No 5 No
PC5P Yes 5 No
niques has been investigated by several researchers, and proved PC5P-M1 Yes 5 M1
to be of high level of strengthening efficacy [16,17], it’s important PC5P-M2 Yes 5 M2
to note that the damage in the RC elements used in these studies PC10 No 10 No
was limited to corrosion or pre-cracking. PC10P Yes 10 No
PC10P-M1 Yes 10 M1
Recently, limited studies have been completed about the effect
PC10P-M2 Yes 10 M2
of the combination of two simultaneous factors on the repair PC15 No 15 No
method of concrete beams. Wu (2010) [18], experimentally and PC15P Yes 15 No
theoretically investigated the combined effect of pre-cracking PC15P-M1 Yes 15 M1
and cyclic loading on reinforced concrete short beams. The results PC15P-M2 Yes 15 M2

revealed that the influence of repair under combined pre-cracking

and cyclic loading was insignificant. The effects of coupled corro-
sion and progressive loading on the FRP repair of reinforced con- mens were 960 mm long, 150 mm wide, and 150 mm high, as illus-
crete beams, were considered by Ali (2012) [19], through Monte- trated in Fig. 1. All of the beams were reinforced with four 10 mm
Carlo simulation technique, which is based on Neural Networks bottom bars (tensile reinforcement), two 10 mm top bars (com-
and finite element method. The results of this study clearly showed pression), and 6 mm stirrups spaced at 100 mm. The concrete
that the coupling of two variables has a particular significance, cover was 20 mm. Each bottom steel reinforcement was fitted with
which requires an extensive attention. In addition, regarding the an electric connection in order to facilitate the corrosion process, as
real field observations, that usually showed two or more internal illustrated in Fig. 1.
or external combined actions, the study of simultaneous degrada- Beams PC0T (0%) were control beams with no corrosion or pre-
tion factors on the mechanical and durability properties of con- cracking, while beams PC5, PC10, and PC15 were corroded at a 5%,
crete beams repair can be considered as more accurate than that 10%, and 15% corrosion rate, respectively, with no pre-cracking.
of dealing with a single factor. Beams PC5P, PC10P, and PC15P were pre-cracked to 60% of the ulti-
Cracks in RC elements may be expected, even under normal ser- mate load of the control beams at a 5%, 10%, and 15% corrosion rate,
vice conditions that include tensile strength of concrete, humidity, respectively. Beams PC5P-M1, PC10P-M1, and PC15P-M1 were pre-
temperature, shrinkage, and creep effects [20]. In addition, con- cracked and corroded at a 5%, 10%, and 15% corrosion rate respec-
crete cracking can be subject to development during the entire ser- tively, then repaired by external bonding of CFRP sheets to the bot-
vice life of RC members due to the extreme exposure environments tom of the beams. Beams PC5P-M2, PC10P-M2, and PC15P-M2
caused by either the overloading or the ingression of chemical were pre-cracked and corroded at 5%, 10%, and 15% corrosion rates
gases and liquids [21]. Although, cracks in RC elements do not cer- respectively, then repaired by external bonding of CFRP sheets to
tainly cause the failure of concrete structures, the presence of the bottom part, and carbon fibers tissue sheets in the bottom, left
cracking can cause an excessive carbonation and chlorides ingress, and right sides of the beams.
which aggravates and accelerates the corrosion and the damage of
concrete cover. When the cracking of the concrete cover occurs due 2.2. Material properties
to steel corrosion, the repair of RC members is essential to main-
tain or restore the serviceability of structure [22,23]. Despite the The concrete mixture used in this study was composed of water
importance of studying the repair of RC elements subjected to to cement (W/C) ratio of 0.63, according to the Dreux Gorisse
the combined effects of cracking and corrosion, no investigation method [24], in order to get slump of 7 cm and 28-day compressive
into this subject has so far been carried out. Hence, this study strength around 40 MPa. Ordinary Portland cement complying
investigated the effect of combined pre-cracking and corrosion with ASTM C150 Type I cement [25] was used as part of the cemen-
on the method of repair of RC beams. In order to examine the effect titious materials. Its principal properties provided by the manufac-
of these simultaneous factors on the structural behavior of pre- turer indicated: C3S and C3A contents of 60% and 7.5% respectively,
cracked RC beams, corrosion rates varying from 5% to 15% were specific gravity of 3.13, Blaine fineness of 380 m2/kg and 28-day
used. These corrosion percentages represent the loss in the cross- compressive strength of 42.5 MPa. Siliceous round sand and
sectional area of the steel reinforcement on the tension side. Two crushed silico-calcareous rocks with maximum grain sizes of
methods were used to repair the damaged reinforced concrete cor- 3 mm and 12.5 mm were used as fine and coarse aggregate in
roded beams. For the first method, Carbon Fiber Reinforced Poly- the production of the concrete. The yield strength of the reinforce-
mers (CFRP) were used to strengthen the bottom portion of the
specimens. And for the second method, sheets containing carbon
fibers were added to the bottom portion, left and right sides of
the specimens.

2. Materials and methods

2.1. Specimens details

A total of 15 RC beams were tested in this study as summarized

in Table 1. The test beams were categorized into four series,
according to the corrosion rate (0%, 5%, 10% and 15%). The speci- Fig. 1. Geometry of beams specimens.
464 A. Siad et al. / Construction and Building Materials 132 (2017) 462–469

Table 2
Composition of concrete mixtures.

Components Composition (kg/m3)

Cement 375
Gravel 960
Sand 512
Water 235
NaCl 0.019

ment bars was 410 and 420 MPa for the 6 and 10 mm bars, respec- Fig. 3. Accelerated corrosion process.
tively, with a modulus of elasticity of 200 GPa. Table 2 shows the
composition of the concrete mixtures. Chloride in the form of NaCl
with a concentration of 5% by weight of cement was added during Table 3
the mixing process to quickly surpass the critical corrosion- Accelerated corrosion parameters.

inducing limit and initiate the reinforcement corrosion [26–28]. The rate of corrosion: % 5 10 15
Mass of rebar exposed to 1094.1 1094.1 1094.1
corrosion (g)
2.3. Preparation of test specimens Length of rebar exposed to 920 920 920
corrosion (mm)
Current density used (i) 432 432 432
Wooden prismatic forms were used to cast the test specimens.
The reinforcements were placed inside the forms, providing a min- Applied current (I) (mA) 250 250 250
imum clear concrete cover of 20 mm. After casting, the specimens Surface of rebar exposed to 57,805 57,805 57,805
were cured for 28 days in a climatic room at 20 ± 2 °C and 95 ± 5% corrosion (A) (mm2)
Molar mass of steel (M) 55.9 55.9 55.9
humidity. All beams were tested after 28 days of casting.
Pre-cracking was achieved by preloading the beams to 60% of Number of load of valence 2 2 2
the ultimate load to induce maximum crack openings of 0.3 mm, Fez+ (z)
in accordance with the suggestions of Schiessl and Raupach Constant of Faraday (F) 96,500  103 96,500  103 96,500  103
(1997) [29]. During the pre-cracking process, one cycle of load- (mA/s)
The mass loss per day (g) 3.61 3.61 3.61
ing/unloading was applied with an initial preloading force of
The total mass loss (g) 54.705 109.41 164.11
1 kN and standard increments rate of 0.005 kN/s. When 60% of The exposure time under 15 30 45
the ultimate load was achieved (49 kN), the loading was stopped electric current (days)
and the unloading proceeded with the same loading speed. The
pre-cracking method was similar for all the beams. All of the
beams were loaded using four-point loading according to ASTM order to ensure electric conductivity [26]. The electrolyte solution
C1341-13 [30] recommendations for more reliable flexural was changed every week in order to maintain a steady concentra-
strength and extended cracking along the specimen surface. The tion. The exposure time for the beams in the electric field was cal-
load was applied using a 300 kN hydraulic actuator. Linear variable culated using Faraday’s law. Table 3 summarizes all the necessary
displacement transducers (LVDT) with a 0.01 mm resolution were data for this stage.
installed at the mid span of each beam in order to measure the
deflection. Fig. 2 shows the transversal microscopic shear cracks m iM
that formed during the pre-cracking stage. During the pre- ¼ ð1Þ
tA zF
loading of the beams, the first cracks appeared and were localised
beneath the two loads. When the preloading was increased, differ- I
ent micro-cracks between the two applied loads appeared. i¼ ð2Þ
Where, m: mass of the rebar exposed to corrosion, t: corrosion
2.4. The process of accelerated corrosion duration, i: current density, M: Molar mass of steel, A: Surface of
the rebar exposed to corrosion area, z: Number of electrons/ atom
In order to obtain the desired rate of corrosion within a reason- generated or consumed, F: Faraday’s constant, and I: Applied
able time frame, an electrochemical method based on the applica- current.
tion of a constant electric current was carried out [8,26,31,32]. This
involves imposing a current density between the steel reinforce- 2.5. Repair methods
ment bars (anode), and a counter-electrode made of a stainless
plate (cathode) below the beams (Fig. 3). The beams were partially To evaluate the effectiveness of the repair method on the flexu-
immersed in a tank containing 5% NaCl (electrolyte solution) in ral capacity of pre-cracked-corroded RC beams, two different

Appearance of cracks in the bottom part Appearance of cracks in the side

Fig. 2. Microscopic cracks during the pre-cracking stage.

A. Siad et al. / Construction and Building Materials 132 (2017) 462–469 465

Fig. 4. Repair methods of tested beams.

Table 4
Properties of composite sheets.

Materials Width (mm) Length (mm) Thickness (mm) Density Fiber percentage (%) Elasticity modulus (MPa)
SikaCarboDur 100 700 1.2 1.6 >68 165,000
SikaWrap 450 700 0.331 1.81 96 242,000

Table 5
Properties of epoxy resins.

Resins Mixture density Shrinkage (%) Tensile strength (MPa) Bond strength (MPa)
Sikadur 30 1,65 ± 0,1 0.04 65–75(3 days at + 10 °C) >4
Sikadur 330 1.3 ± 0.1 / 30(7 days at 23 oC) >4

Table 6
strengthening techniques with CFRP sheets were used. Those two
Experimental results of tested beams.
methods were chosen because of their easier and quicker installa-
tion practices, in addition to their proved reliability in the rein- Beams Ultimate deflection Ultimate load Stiffness k
(mm) (kN) kN/mm
forcement of corroded or pre-cracked RC beams, with increased
stiffness, ductility and load carrying capacity [33]. In the first PC0T 10.31 79.21 25.40
PC5 6.37 82.67 22.76
method (M1), Carbon Fibers Reinforced Polymers (CFRP) were used
PC5P 6.2 82.71 21.39
to strengthen the bottom portion of the beams, as illustrated in PC5P-M1 2.83 88.16 45.89
Fig. 4. The CFRP were Sika CarboDur sheets embedded in an epoxy PC5P-M2 3.11 135.83 56.28
matrix [34], presented in the form of prefabricated plates that can PC10 6.23 80.71 17.65
be used to reinforce concrete and masonry structures. These plates PC10P 6.53 78.51 17.20
PC10P-M1 2.42 89.53 50.91
are attached to the beams using double component epoxy resin PC10P-M2 2.75 135.32 53.24
Sikadur30 (Table 5) [35], ensuring perfect bonding between the PC15 5.77 72.32 18.15
concrete and CFRP, this method is often used when the lower por- PC15P 5.45 69.62 15.01
tion of the beam is visible. The second method (M2) is similar to PC15P-M1 1.59 76.57 56.69
PC15P-M2 2.47 112.81 67.9
the first (Fig. 4), except in this case, a Sika Wrap 600C [36] type tis-
sue was used below the CFRP sheet and on both sides of the beams.
The type of CFRP used is a stitched unidirectional carbon fiber fab-
ric, oriented in the longitudinal direction of the sheet. It is com- 3.1. Behavior of corroded beams without external reinforcement
posed of 96% mid-range strength black carbon fibers, (PC0T, PC5, PC10, PC15)
characterised by sheet thickness of 0.331 mm and density of
1.81. An epoxy resin Sikadur330 (Table 5) [37] was used in the sec- Fig. 5 shows the load-deflection curves at the mid-span for the
ond method to ensure a good anchorage between the sheets and corroded beams PC5, PC10, and PC15. The behavior of the beams
the beams. Tables 4 and 5 show the properties of all the repair can be divided into two phases. The first phase is linear and the
materials according to the manufacturer’s specifications. force is proportional to the deflection until the loading threshold
(the elastic domain interval). The second phase is non-linear and
characterized by a slight increase in the load compared to the rapid
3. Results and discussion evolution of the deflection. In this phase, vertical micro-cracks
appeared at the bottom of the beams, which corresponds to the
Table 6 summarizes all the experimental results for the control, ultimate failure of the beams.
sound/corroded and pre-cracked/corroded beams without external It is clear from the test results that corrosion adversely affects
reinforcement, and for the pre-cracked/corroded beams that were the flexural capacity and the stiffness of reinforced concrete beams.
reinforced using methods M1 and M2. PC0T represents the average The results of the stiffness of the beams presented in Table 6 were
results of the control beams PC0T1, PC0T2 and PC0T3 tested with calculated from the load-deflection diagrams, where the material
no corrosion or pre-cracking. For comparison purposes, the men- was in phase 1. It can be noted from this table that the stiffness
tioned results are discussed in depth in the following sections of the beams decreased from 25.40 kN/mm for beam PC0T to
according to the state of tested specimens. 22.76 kN/mm for beam PC5T. This slight reduction in stiffness
466 A. Siad et al. / Construction and Building Materials 132 (2017) 462–469

Fig. 5. Behavior of corroded beams.

can be attributed to the loss of bond between reinforcing steel and

concrete due to corrosion. In addition, it can be noted that the stiff- Fig. 6. Longitudinal cracks appeared on the beam PC15P.
ness decreases as the corrosion rate increases. From Fig. 5 and
Table 6, the ultimate deflection of beams PC5, PC10, and PC15
was 6.23 mm, 6.2 mm, and 5.77 mm, respectively, while for the
control beams (PC0T), it was 10.31 mm, on average. The differences
between PC0T and PC5, PC10, and PC15 were 38.22%, 39.57%, and
44.03% respectively. It is clear from these results that corrosion
adversely affects the deflection capacity and the stiffness of rein-
forced concrete beams. In addition, beams PC5 and PC10 achieved
a high ultimate load compared to PC0T, while for PC15, the maxi-
mum load was 8.7% smaller than that of PC0T. This can be attribu-
ted to the reduction in the cross-sectional area of the steel
reinforcement, and the presence of corrosion products that affect
the adherence between concrete and steel. The results also
revealed that the ultimate load for the PC5 beam (82.67 kN) was
almost similar to that of PC10 (80.71 kN). This finding was not pre-
dicted because increases in corrosion are usually accompanied by a
decrease in ultimate strength. However, other researchers have
also observed similar behavior [8,38], and they concluded that an Fig. 7. Behavior of pre-cracked/corroded beams.
enhancement of anchorage capacity, believed to be associated
principally with increased radial stresses at the steel-concrete 8.7% drop registered for PC15 (corroded only), around 6.3% extra
interface, was able to offset the loss of the bar section. Although, decrease was generated from the combined effect of pre-cracking
the ultimate load of corroded PC5 and PC10 beams was higher than and corrosion. In terms of mid-span beam deflection capacities,
that of the control beam PC0T, this finding was not true for PC15. results decreased by 39.86% for PC5P, 36.66% for PC10P, and
Increasing the corrosion rate up to 15% caused around 8.7% 47.13% for PC15P, compared to PC0T, which shows that the com-
decrease in the ultimate strength of PC15 compared to PC0T. This bined pre-cracking and corrosion didn’t influence the ultimate
can explain that the corrosion rate can be considered harmful to deflexion results considerably, compared to the first series without
the ultimate load, especially when it equals or exceeds 15%. pre-cracking.
However, for a deflection value of 2 mm, the loads were 78 kN
3.2. Behavior of pre-cracked/corroded beams without external for PC5P and 61 kN for PC10P, compared to 68 kN and 54 kN for
reinforcements (PC0T, PC5P, PC10P, PC15P) PC5 and PC10, respectively. These small differences between cor-
roded and pre-cracked/corroded beams may be attributed to the
When the pre-cracking and corrosion were combined, two initial pre-cracking that made these beams more ductile by delay-
types of cracks were noticed; transversal cracks, due to the pre- ing the load of plasticity. This delay increases the plasticity stage of
loading, and longitudinal cracks (Fig. 6) resulted from the applied the beams (PC5P and PC10P) up to a certain corrosion rate.
corrosion rate. Fig. 7 shows the load-deflection curves at the At a higher corrosion level (15%), the pre-cracking has a detri-
mid-span for the pre-cracked and corroded beams PC5P, PC10P, mental effect on the behavior of the beams, accelerating the plas-
and PC15P. These results reveal that the ultimate load for beams ticisation stage. This is demonstrated by the values of the applied
PC5P and PC10P was almost similar to that of the beams in series loads at a 2 mm deflection for beams PC15 and PC15P (60 kN and
one (PC5 and PC10), with slight decrements. For instance, com- 50 kN, respectively). This reduction may be explained by the pres-
pared to PC5 and PC10 results (82.67 kN and 80.71 kN respec- ence of the transverse crack from the initial pre-cracking, which
tively), PC5P and PC510 showed ultimate loads of 82.31 kN for facilitated the transport of chloride [39], leading to the augmenta-
PC5P and 78.51 kN for PC10P. Which explain that even after pre- tion in the contact pressure at the steel-concrete interface, and in
cracking, the application of 5% and 10% corrosion rate did not influ- the tensile stresses in the concrete member parallel to the rein-
ence the maximum flexure strength supported by the beams. How- forcements. This causes the formation of longitudinal cracks, and
ever, with 15% corrosion rate, the ultimate load of PC15P decreased acceleration in the corrosion rate, which evidently decreases the
by 12.11% compared to the control beam PC0T. If considering the ultimate load.
A. Siad et al. / Construction and Building Materials 132 (2017) 462–469 467

Fig. 8. Behavior of pre-cracked/corroded beams reinforced using the M1 method.

Fig. 10. Behavior of pre cracked/corroded beams repaired with M2 method.

3.3. Behavior of pre-cracked /corroded beams reinforced using the M1 3.4. Behavior of pre-cracked/corroded beams reinforced using the M2
method (PC0T, PC5P-M1, PC10P-M1, PC15P-M1) method (PC0T, PC5P-M2, PC10P-M2, PC15P-M2)

Fig. 8 shows the load-deflection curves at the mid-span for the Fig. 10 shows the load-deflection curves at the mid-span of the
corroded beams PC5P-M1, PC10P-M1, and PC15P-M1. A linear ini- corroded beams PC5P-M2, PC10P-M2, and PC15P-M2. Those curves
tial phase was observed for all the curves followed by a sudden revealed a significant increase in the ultimate strength of the
increase in deflection, especially for PC5P-M1 and PC10P-M1. The repaired beams, compared to the control beams (PC0T). The repair
linear behavior was without any appearance of cracks up to by using M2 method resulted in decrements of about 171.48% for
approximately 70–90 kN. At this range, visual cracking appeared PC5P-M2, 146.84% for PC10P-M2, and 142.42% for PC15P-M2, com-
and the initial de-bonding of the composite sheets was noticed. pared to the ultimate strength of PC0T. However, those significant
The ultimate load of PC15P-M1 was 96.66% of that of the control improvements were accompanied by a remarkable decrease in the
beams PC0T. This indicates that PC15P-M1 was almost back to its mid-span deflection capacities. For instance, ultimate deflections
initial state. For PC5P-M1 and PC10P-M1, the ultimate load of PC5P-M2, PC10P-M2 and PC15P-M2 were 69.83%, 73.32%, and
exceeded the original state by 11.11% and 13%, respectively. In 76%, respectively, lower than that of PC0T. The improvement in
addition, from the results presented in Table 6, the stiffness of the mechanical properties between damaged/repaired (with CFRP
the beams repaired with M1 method increased significantly com- reinforcement) and control beams was already observed by Ben-
pared to the control specimen, which explains the effectiveness jeddou et al. (2007) [40]. They explained that for any damage
and the improved performance of the repaired beams by using degree, the overall result came from the high strength CFRP lami-
the M1 method. nate that enhanced the mechanical performances of the damaged/
However, with regards to the ultimate deflections, significant repaired beams. The elastoplastic behavior of beams changed to
decrements of about 72.55%, 76.52%, and 84.57% were calculated elastic behavior, which explains the lower ductility performance
for PC5P-M1, PC10P-M1 and PC15P-M1, respectively, compared of the repaired beams.
to PC0T. These results can be explained by the fact that the com-
posite sheet increased the stiffness of the beams, preventing them
from deforming. The addition of the composite plate changed the 3.5. Comparison of repair methods M1 and M2
mode of failure from ductile to brittle, as illustrated in Fig. 9.
For both methods M1 and M2, it is clear from the test results
that using the composite material to repair the damaged beams
significantly enhanced and improved their flexural capacity. The
high strength of the CFRP sheets compensated the negative effect
of corrosion and pre-cracking. A comparison of results from these
two methods (Fig. 11) revealed that, compared to M1, the ultimate
loads in method M2 increased by 60.18%, 33.84%, and 45.76% for
corrosion rates of 5%, 10% and 15%, respectively. In addition, the
ultimate deflections increased by 2.72%, 3.2%, and 8.57% for corro-
sion rates of 5%, 10% and 15%, respectively. Also it can be observed
that the stiffness of the beams increased from method M1 to
method M2.
The mechanical behavior of the repaired beams was depending
upon the performance of the strengthening system. According to
Benjeddou et al. (2007) [40], the characteristics of CFRP sheets is
the only reason that can influence the rigidity of the repaired
beams. For the first method M1, the bonding of the SikaCarboDur
CFRP plates in the tension zone significantly enhanced the rigidity
of reinforced beams. Therefore, the failure mode changed from
ductile to brittle. For the M2 method, the addition of SikaWrap
CFRP sheets to the bottom region, left and right sides (tension
and shear regions) advantageously enhanced the stiffness, the ulti-
Fig. 9. The detachment of the composite sheets. mate load, and the ductility of the repaired beams, probably
468 A. Siad et al. / Construction and Building Materials 132 (2017) 462–469

Fig. 11. Comparison of the repair methods M1 and M2.

because of the properties of the SikaWrap sheets compared to Sika- ucts can accumulate at the bottom of the crack, and prevent the
CarboDur CFRP plates. In addition, if considering that de-bonding further ingress of chlorides, water, and oxygen into the concrete
of the CFRP reinforcement was proved to be the main reason of [42,43]. The second phenomenon is maybe due to self-healing,
the flexural failure of beams reinforced with CFRP [41], de- which depends on the width of the cracks and the cement hydra-
bonding of CFRP in M1 was in tension zone only (bonding zone), tion. Indeed, the ability of cementitious materials to hydrate the
whereas the de-bonding of CFRP in M2 occurred in almost all of additional cement particles can help the pre-cracked concretes to
the beams strengthened in shear and tension zones. These can recover some of their physical and mechanical properties [44,45].
explain that method M2 increased the bond capacity of CFRP For the other levels of corrosion (10% and 15%), the results
sheets compared to method M1, which can also be the cause of revealed differences of 2.78% and 3.41% between PC10 and PC10P
enhanced properties of pre-cracked/corroded beams repaired by (for a corrosion rate of 10%), and between PC15 and PC15P (for a
using method M2. corrosion rate of 15%) respectively. These decrements in the ulti-
mate strength of pre-cracked/corroded beams compared to those
3.6. Relationship between the pre-cracking and rate of corrosion of corroded beams can be explained by the formation of higher
amounts of corrosion products due to concrete pre-cracking and
A comparison of the test results of corroded beams (without increased level of corrosion. The accumulation of corrosion prod-
pre-cracking) and pre-cracked/corroded beams (Fig. 12) revealed ucts at the steel-concrete interface induced more contact pressure,
only slight differences, in spite of the presence of pre-cracking which in turn increased the corrosion crack width, especially in the
during corrosion exposure. For example, the ultimate load for the longitudinal direction parallel to the tensile reinforcement (Fig. 6).
corroded beam PC5 was 82.67 kN at a corrosion rate of 5%, while This phenomenon was expected to highly decrease the ultimate
the ultimate load for the pre-cracked/corroded beam at the same load between the corroded beams and the pre-cracked/corroded
corrosion rate was 82.71kN. These results can be explained by beams. The small decrements recorded for 10% and 15% corrosion
two phenomena. The first phenomenon is the reduction in the pen- rates, proved the explanation mentioned above about the forma-
etration rate of chlorides within the beam because corrosion prod- tion of self-healing products or a rust layer in concrete cracks,

Fig. 12. Relationship between the pre-cracking and rate of corrosion.

A. Siad et al. / Construction and Building Materials 132 (2017) 462–469 469

which in turn reduced the corrosion activity and limited the effect [13] A. Bilotta, F. Ceroni, M. Di Ludovico, E. Nigro, M. Pecce, G. Manfredi, Bond
efficiency of EBR and NSM FRP systems for strengthening concrete members, J.
of pre-cracking. These results also agreed with those of El Maad-
Compos. Constr. 5 (2011) 757–772.
dawy and Topper (2005) [46], who showed that the presence of [14] B. Almassri, F. Al Mahmoud, R. Francois, Behaviour of corroded Reinforced
flexural cracks during corrosion exposure had almost no effect on Concrete beams repaired with NSM CFRP rods, experimental and finite
the flexural strength of beams, especially after long term corrosion element study, Compos. B 92 (2016) 477–488.
[15] B. Almassri, Strengthening of corroded reinforced concrete (RC) beams with
exposure. near surface mounted (NSM) technique using carbon fiber polymer (CFRP)
rods, an experimental and finite element (FE) modelling study (Ph.D. thesis),
INSA of Toulouse, France, 2015.
4. Conclusions [16] M.Z. Jumaat, A.A. Shukri, M. Obaydullah, M.N. Huda, M.A. Hosen, N. Hoque,
Strengthening of RC beams using externally bonded reinforcement combined
with near-surface mounted technique, Polymers 8 (2016) 261.
The experimental tests carried out during this study revealed a [17] J.M. Sena-Cruz, J.A. Barros, M.R. Coelho, L.F. Silva, Efficiency of different
significant increase in the ultimate load of all beams reinforced, techniques in flexural strengthening of RC beams under monotonic and fatigue
using composite materials containing carbon fibers (up to loading, Constr. Build. Mater. 29 (2012) 175–182.
[18] Z.Y. Wu, Experimental study of the behavior of short concrete beams pre-
171.48% for PC5P-M2, 146.84% for PC10P-M2, and 142.42% for cracked and reinforced by composite materials under static loading and
PC15P-M2 compared to PC0T). However, there was a significant fatigue (Ph.D. thesis), LCPC Paris, France, 2010.
reduction in the ultimate deflection (up to 69.83%, for PC5P-M2, [19] O.M.M. Ali, Time-dependent reliability of FRP strengthened reinforced
concrete beams under coupled corrosion and changing loading effects (Ph.D.
73.32% for PC10P-M2, and 76% for PC15P-M2 compared to PC0T). thesis), Angers University, France, 2012.
The following conclusions were drawn from the experimental [20] D. Zhang, S. Shijun, Y. Zhao, J. Weiliang, U. Tamon, Cracking behavior of CFRP
results: laminate-strengthened RC beams with premechanical and postmechanical
environmental damage, J. Compos. Constr. 4 (2014) 04014066.
[21] M. Ekenel, J.J. Myers, Durability performance of RC beams strengthened with
 The different gains in the ultimate load of the repaired beams epoxy injection and CFRP fabrics, Constr. Build. Mater. 6 (2007) 1182–1190.
were depending on the performance of the strengthen system. [22] G. Malumbela, M. Alexander, P. Moyo, Serviceability of corrosion-affected RC
beams after patch repairs and FRPs under load, Mater. Struct. 44 (2011) 331–
The high strength of the composite reinforcement increased
the rigidity of the beams. [23] O. Rio, C. Andrade, D. Izquierdo, C. Alonso, Behaviour of patch repaired
 The method of repair influenced considerably the results of the concrete structural elements under increasing static loads to flexural failure, J.
reinforced beams. Mater. Civ. Eng. 17 (2005) 168–177.
[24] G. Dreux, J. Festa, A New Guide to Concrete and Its Constituents, France, Paris,
 The addition of CFRP sheets to the bottom, left and right sides 1998.
prevented the rapid de-bonding of the composite from the [25] American Society for Testing and Materials, Standard Specification for
beams, which in turn increased the mechanical performance Portland Cement ASTM C150-12, 2012.
[26] V. Bouteiller, J. Dauthuile, A. Deman, Feasibility of Accelerated Corrosion on
of the strengthen system. For instance, the ultimate load of Reinforced Concrete Beams APPLET L2-3-2, National Research Agency, France,
the beams reinforced using M2 increased by 60.18%, for corro- 2008.
sion rate of 5%, compared to M1 result. [27] F. Vallot, E. Cailleux, V. Pollet, Synthesis Report of Aging APPLET L3-2-2,
National Research Agency, France, 2009.
 The presence of pre-cracking during corrosion exposure influ- [28] J. Cairns, Y. Du, D. Law, Structural performance of corrosion-damaged concrete
ences the ultimate load capacity of reinforced concrete beams, beams, Mag. Concr. Res. 60 (2008) 359–370.
especially for corrosion rates of 10% and 15%. However, the for- [29] P. Schiessl, M. Raupach, Laboratory studies and calculations on the influence of
crack width on chloride-induced corrosion of steel in concrete, ACI Mater. J. 94
mation of self-healing products or a rust layer in concrete cracks
(1997) 56–62.
limited the combined effect of pre-cracking and corrosion. [30] American Society for Testing and Materials, Standard Test Method for Flexural
Properties of Continuous Fiber-Reinforced Advanced Ceramic Composites
ASTM C1341-13, 2013.
[31] A. Hassan, Bond of reinforcement in concrete with different types of corroder
References bars (MASc thesis), Ryerson University, Canada, 2003.
[32] R. Smith, The effects of corrosion on the performance of reinforced concrete
beams (MASc thesis), Ryerson University, Canada, 2007.
[1] G. Schmitt, Global Needs for Knowledge Dissemination, Research, and
[33] F. Shamsoon, Behaviour of reinforced concrete beams strengthened by CFRP
Development in Materials Deterioration and Corrosion Control, World
wraps with and without end anchorages, Procedia Eng. 77 (2014) 123–130.
Corrosion Organization, New York, 2009.
[34] SikaÒ CarboDurÒ Plates, Pultruded Carbon Fiber Plates for Structural
[2] J.P. Olivier, A. Vichot, Concrete Durability Chapter 9, Press of National School of
Bridges and Roads, France, 2008.
[35] SikadurÒ-30, Structural Adhesive for Bonding Reinforcement.
[3] A.K. Azad, S. Ahmad, S.A. Azher, Residual strength of corrosion-damaged
[36] SikaWrapÒ-600 C, Woven Unidirectional Carbon Fibre Fabric, Designed for
reinforced concrete beams, ACI Mater. J. 104 (1) (2007) 40–47.
[4] C. Li, J. Zheng, W. Lawanwisut, R.E. Melchers, Concrete delamination caused by
[37] SikadurÒ-330, Resin Epoxy of Impregnation for Tissue of Strengthening.
steel reinforcement corrosion, J. Mater. Civ. Eng. 19 (2007) 591–600.
[38] J. Cairns, G. Plizzari, Y. Du, D. Law, C. Franzoni, Mechanical properties of
[5] X. Wang, X. Liu, Bond strength modelling for corroded reinforcements, Constr.
corrosion damaged reinforcement, ACI Mater. J. 102 (2005) 256–264.
Build. Mater. 20 (2006) 177–186.
[39] T. Vidal, Requalification of structures degraded by corrosion of reinforcements
[6] L. Berto, P. Simioni, A. Saetta, Numerical modelling of bond behaviour in RC
(Ph.D. thesis), Toulouse University, France, 2003.
structures affected by reinforcement corrosion, Eng. Struct. 30 (2008) 1375–
[40] O. Benjeddou, M. Ben Ouezdou, A. Bedday, Damaged RC beams repaired by
bonding of CFRP laminates, Constr. Build. Mater. 21 (2007) 1301–1310.
[7] French Association of Civil Engineering (FACE), Rehabilitation of Reinforced
[41] A.M. Reza, C. Czaderski, M. Motavalli, Debonding failure modes of flexural FRP-
Concrete Degraded by Corrosion France, 2003.
strengthened RC beams, Compos. B 5 (2008) 826–841.
[8] A.H. Al-Saidy, A.S. Al-Harthy, K.S. Al-Jabri, M. Abdul-Halim, N.M. Al-Shidi,
[42] P. François, Influence of concrete cracking on reinforcement corrosion (Ph.D.
Structural performance off corroded RC beams repaired with CFRP sheets,
thesis), Laval University, Canada, 2009.
Compos. Struct. 92 (2010) 1931–1938.
[43] T. Lorentz, C. French, Corrosion of reinforcing steel in concrete: effects of
[9] C. Francesca, C. Edoardo, G. Manfredi, P. Marisa, Durability issues of FRP rebars
materials, mix composition and cracking, ACI Mater. J. 92 (1995) 181–190.
in reinforced concrete members, Cem. Concr. Compos. 28 (2006) 857–868.
[44] H. Haoliang, G. Ye, D. Damidot, Characterization and quantification of self-
[10] H. Wang, A. Belarbi, Ductility characteristics of fiber-reinforced-concrete
healing behaviors of microcracks due to further hydration in cement paste,
beams reinforced with FRP rebars, Constr. Build. Mater. 25 (2011) 2391–2401.
Cem. Concr. Res. 52 (2013) 71–81.
[11] A. Mosallam, S. Banerjee, Enhancement in in-plane shear capacity of
[45] V.C. Li, E.H. Yang, Self Healing in Concrete Materials, Self Healing Materials,
unreinforced masonry (URM) walls strengthened with fiber reinforced
Springer, Netherlands, 2007, pp. 161–193.
polymer composites, Compos. B 42 (2011) 1657–1670.
[46] T. El Maaddawy, T. Topper, Long term performance of corrosion-damaged
[12] M.Z. Jumaat, M.A. Alam, Strengthening of RC beams using externally bonded
reinforced concrete beams, ACI Struct. J. 102 (2005) 649–656.
plates and anchorages, Aust. J. Basic Appl. Sci. 3 (2008) 2207–2211.