Durability of Reinforced Concrete Beams Externally Strengthened with CFRP Laminates under Harsh Climatic Conditions

Forty typical-scale RC beams were cast. Fig. 1 depicts the specimen dimensions and reinforcement details. The length of each beam was 2 m, and the cross-sectional dimensions were 140 × 280 mm. The beam samples were designed as a tension-controlled section per ACI Committee 318 (ACI 2014) flexural design guidelines. An adequate number of stirrups were provided to avoid shear failure. This helps clarify the effect of the conditioning on the contribution of CFRP strengthening. The primary flexural reinforcement consisted of two bars that were 12 mm in diameter, and the top compression reinforcement was two bars that were 8 mm in diameter. Stirrups of 8 mm-diameter rebar were provided at a distance of 120 mm center-to-center in the shear spans, and the distance between the central two stirrups was 100 mm, as illustrated in Fig. 1. Fig. 1 also schematically presents the final layout of the strengthened specimens. The CFRP sheets were placed over a length of 1,600 mm, whereas 200 mm from either side was left uncovered. The width of CFRP laminates was 100 mm; hence, 20 mm of beam width from either side of the sheets was not strengthened. The thickness of the CFRP sheet was 0.17 mm, and the final thickness of the CFRP laminates (CFRP plus epoxy) was estimated to be 1.02 mm.

Table 1 presents the testing plan for the RC beam specimens. The samples were divided into three main categories, depending on the exposure type: laboratory conditions, outdoor sunlight, and immersion in saline water. Each group was subdivided into three further groups, depending on the exposure period (i.e., 6, 12, and 24 months). For each set of exposure type, two CFRP-strengthened and two control specimens were cast. This was to eliminate possible error in workmanship and material variability while obtaining reliable results. The number of samples was limited to two because these beams samples are not easy to handle and control in large numbers, which would be the case if the samples for each category were increased to three. In addition, two samples each were also tested at 28 days to establish a benchmark to study the effects of exposure and time on the strengthened RC beams. The specimens were designated according to their exposure type and time of testing. For instance, for the CB-S-180-1 specimen, CB stands for control beam, S denotes sunlight exposure, 180 represents the length of exposure in days, and 1 indicates that it is the first beam to be tested in this category. Moreover, C-360-W-2 represents a CFRP-strengthened beam exposed to saline water (W) for 360 days, which is the second beam to be tested in this category.

In this study, commercially available CFRP laminates (V-Wrap™ C200H 2014) were employed. As per the manufacturer, the CFPR laminates (sheet with epoxy) had a design thickness of 1.02 mm, tensile strength of 1,240 MPa, modulus of elasticity of 73.8 GPa, and strain at rupture of 1.7%. The CFRP sheets were attached to concrete surfaces using a two-component epoxy comprising Parts A (epoxy) and B (hardener) (V-Wrap™ 700S 2014). As per the instructions of the manufacturer, Part A was premixed for 2 min, and then the entire packaged quantities of both parts were mixed for 3 min at about 400 rpm. Because the epoxy resin has an immense influence on the durability properties of strengthened materials, it is important to mention the properties of the epoxy system. The epoxy had a modulus of elasticity of 3.2 GPa, tensile strength of 45 MPa, and elongation at rupture of 5.5%. The T_g of this epoxy adhesive was 76.6°C.

Soffits of the beams were sandblasted to have a roughened concrete surface for CRFP laminate application, as illustrated in Fig. 2. The wet-layup method was used, in which the CFRP sheets were first immersed in a prepared epoxy until saturation was achieved. First, a layer of epoxy adhesive was applied at the prepared concrete surface [Fig. 2(a)]. Then, the saturated laminates were carefully placed at the marked length of the beam soffit [Fig. 2(b)], followed by another epoxy layer applied at the top surface of the sheets [Fig. 2(c)]. Steel rollers were moved with gentle pressure over the sheets to remove all air voids. The strengthening procedure was conducted after curing the concrete beams for two weeks. After application, the epoxy adhesives cured for two more weeks for the specimens tested at 28 days.

After strengthening, the beams were exposed to sustained sunlight [as shown in Fig. 2(d)] and immersed under saline water with 3.5% NaCl [Figs. 2(e and f)]. In contrast, the laboratory conditioned specimens were placed inside a room with a controlled climate of temperature 23°C and relative humidity (RH) of about 40%. The exposure of the specimens was carried out in the facilities of the Qatar University in Doha, Qatar. It was an open area with no shelter for the sunlight-exposed samples. Saline water tanks were also directly exposed to sunlight, and the tanks were refilled regularly to keep the sample immersed for up to two years. The climate in Doha city is extreme, with temperatures between 18°C and 50°C and RH between 40% and 100%. The UV index varies from 6 in January to 12 in the months of June to October. From May to October, the temperature remains in the range of 40°C to 50°C, and the RH is between 80% and 100%.

Fig. 3 presents the instrumentation for the four-point bending tests performed using a servo-hydraulic test frame with a load capacity of 2,000 kN. Displacement control loading was applied at a rate of 1 mm/min. The machine support rollers provided the conditions of a simply supported beam. The tested span length was 1,690 mm between the two reaction supports, and the shear span was 561.5 mm, the distance between the reaction support and loading point. The loading points at the top of the beams were 567 mm apart. A strain gage was applied at each flexural reinforcement at the top surface of the beam and the bottom of the CFRP laminates, all at the midspan of the beam. Two L-shaped aluminum rods were pasted at either side of the beam at midspan, and linear variable displacement transducers were fixed to them to measure the deflections, as displayed in Fig. 3.

The loads versus deflections and strain curves were recorded through the data acquisition machine. Furthermore, the crack initiation and propagation in the beams and possible debonding or fiber rupture in the CFRP laminates were visually monitored.

The effects of the exposure types and their application times were assessed in terms of strength retention by the CFRP-strengthened RC beams. The parameters (degradation factors) used to analyze the results are elaborated in Eqs. (1)–(5). These strength reduction factors are developed based on the approach presented by Tatar and Hamilton (2016c). Fig. 4 schematically explains the concept behind these degradation factors. Measuring the loss in flexural strength and bond strength through ultimate loads after exposure is an indirect method to evaluate the environmental effects on the durability of the FRP–epoxy strengthening systems. The laboratory strength ratio (LSR) was used to quantify the gain in strength with a CFRP-strengthened system. It is the ratio between the ultimate load capacity of the strengthened laboratory specimen, P_nf, divided by the ultimate load of the control-laboratory specimen, P_nl, as shown in Eq. (1). The exposure strength ratio (ESR) is obtained by dividing the ultimate load of a strengthened exposed specimen, P_nfe, to the ultimate load of an unexposed strengthened laboratory specimen, P_nf. This factor covers the possible degradation of the concrete matrix, steel corrosion, damage to FRP materials, and degradation of epoxy adhesives under certain exposure types and times.

To assess the bond strength retention, the degradation originated from the concrete member (corrosion or concrete decay) must be subtracted, and only the degradation on the level of epoxy adhesives is considered. For this purpose, indices R_b and R_b28 are introduced [Eqs. (3) and (4)], where the original strength of the exposed concrete member, P_ne is subtracted from the total strength of the strengthened exposed CFRP beams, P_nfe, and the strength of the unexposed concrete member, P_nl, is subtracted from the total strength of the unexposed CFRP-strengthened member, P_nf. For further clarification, please refer to Fig. 4. For R_b, the difference (P_nfe − P_ne) was divided by P_nf − P_nl, which quantifies the bond degradation. Similarly, the bond degradation against 28-day samples was calculated by R_b28. The flexure strength retention factor, R_n, is elaborated in Eq. (5), and the ultimate strength of an exposed-strengthened specimen is divided by the ultimate strength of an unexposed strengthened specimen; this value is further divided by the ratio of strength between the exposed unstrengthened to unexposed unstrengthened values [Eq. (5)]:where P_nl, P_nf, P_ne, and P_nfe, = ultimate loads of the control laboratory specimen, strengthened laboratory specimen, exposed control specimen, and exposed strengthened specimen, respectively.

Two parameters were defined for the beam ductility until the yielding of the main reinforcement and after reaching the ultimate loads. Eqs. (6) and (7) elaborate these ductility indices, respectively:where δy, δu, and δf = deflection at the steel yielding, ultimate load, and failure loads, respectively.

Fig. 5 lists the failure modes in CB and CFRP-strengthened specimens at 28 and 180 days, and Fig. 6 lists the failures after 360 and 730 days of laboratory, sunlight, and saline water exposure. One sample from each category was selected to demonstrate the effect of conditioning on the type of failures in the CFRP beams. However, the failure modes of each specimen are presented in Table 2. In the case of CFRP-strengthened beam specimens, up to four types of failure modes were observed: cohesive, adhesive, interfacial, and rupture of laminates. Cohesive failure is primarily in the concrete matrix near the tensile face, where a large chunk of concrete is delaminated from the bulk, and the bond between the concrete layer and laminates is not broken. Interfacial failure is a mix with the debonding of laminates from some part of the concrete surface with failure also passing through the concrete cover. Adhesive failure is on the level of the bond at the interface between the composites and concrete substrate. Only the CFRP laminates debond from the soffit of the beams with some traces of concrete on the surface. The rupture of CFRP laminates at failure is considerably rare in the case of concrete application. These failure modes have been observed and reported in several studies in the literature (Grace and Singh 2005; Choi et al. 2012, Hawileh et al. 2014; Tatar and Hamilton 2016c; Choobbor et al. 2019). In this study, the first three types of failure were observed in the beams tested at 28 days and up to 730 days of applied exposure (Figs. 5 and 6).

The failure in CBs was a conventional one under all exposure types. At first, flexural cracks appeared, then the yielding of the flexural reinforcement initiated, followed by concrete crushing under compression [Fig. 5(a)]. In CFRP-strengthened specimens, the failure was initiated with the appearance of flexural cracks, followed by steel yielding, shear cracks, debonding or delamination, and then concrete crushing at the top compression face. First, flexural cracks appeared and moved along the epoxy at the interface between the CFRP laminates and concrete surface, which resulted in either cover separation or sheet debonding. The case in which epoxy adhesives resisted higher loads, higher stress concentrations occurred at the laminate ends, which the concrete could not withstand, resulting in forming horizontal shear cracks and concrete delamination (concrete cover separation) causing cohesive failure.

From two samples tested at 28 days, one showed cohesive failure and the other exhibited interfacial failure with a large chunk of concrete cover attached to the CFRP laminate [Fig. 5(b)]. One sample tested at 180 days after laboratory exposure exhibited adhesive failure and the other exhibited interfacial failure. Both specimens that were exposed to saline water for 180 days had adhesive failures [Fig. 5(d)] with no concrete cover detached from the beam. Both samples that were exposed to the sunlight for 180 days had cohesive failures with the concrete cover separating from the level of the main reinforcement [Fig. 5(e)]. At 360 days, one cohesive [Fig. 6(a)] and one adhesive failure were observed in the laboratory specimens. All specimens exposed to the sunlight and saline water demonstrated adhesive failure [Figs. 6(b and c)]. At 730 days of exposure, the laboratory samples again exhibited the cohesive mode of failures, as it was the case for specimens tested at 28 days [Fig. 6(d)]. The sunlight-exposed specimens demonstrated cohesive and interfacial failure. Both samples exposed to saline water had adhesive failure with relatively concrete-free CFRP laminates. Although no clear pattern of failure mode with the time of exposure was found, the exposure type revealed that saline water transformed the failure mode from cohesive to completely adhesive starting from 180 days to 730 days. Laboratory exposure resulted in no change in failure mode, and most samples after 180, 360, and 730 days exhibited either cohesive or interfacial modes of failure. Sunlight exposure was also not as detrimental as the saline water exposure. This indicates that hydrothermal exposure is more severe than indoor and dry heat exposure. Pan et al. (2015) also reported a change in failure mode from cohesive to adhesive after only two weeks of water immersion on CFRP-strengthened single-lap shear samples. In contrast, Grace and Singh (2005) reported no change in failure mode for RC beams strengthened with CFRP fabrics and CFRP plates after 10,000 h of 100% humidity, in saltwater solution, and dry heat exposure. The observed failure mode was delamination or debonding of strengthening laminates.

The onset of debonding or concrete delamination primarily coincides with the ultimate load capacity of the beams, which means both occurred at the same loads, illustrating that the failures were initiated due to the debonding of the laminates or delamination of the concrete. Until failure, the concrete crushing at the compression zone was not initiated. However, as the sheet delaminated, concrete crushing immediately followed. This observation was made in all samples exposed for up to two years. This verifies that the CFRP laminates retain the integrity of the beam section until debonding and govern the failure load of the RC beam. The CFRP laminates or the individual carbon fibers were not broken or ruptured at failure.

In the RC beams, the cracks appear once the flexure stress crosses the modulus of rupture of concrete at the tension side of the beam. Because CFRP sheets provide confinement at the soffit of the beam and act as an external flexural reinforcement, resisting the applied moments, they increase the beam stiffness and its modulus of rupture to higher loads. In both CB and CFRP, vertical cracks appeared in the tension side in the constant moment area after a certain load level. The number of cracks increased and elongated in the compression zone. As loads increased, the cracks were also initiated in the shear zones on both sides. Once entered in the compression zone, all cracks became flexural-shear cracks and started to incline toward the loading points. This behavior was observed in all beams after all exposure types and times.

In the case of the CB, the cracks started appearing at between 21 and 33 kN loads, whereas in CFRP-strengthened specimens, they appeared at between 35 and 53 kN. That is a 66% to 152% increase in the load at the first crack. The exposure to sunlight and saline water for up to two years exhibited no effects on the crack loads. The exposure conditions did not reduce the ability of the CFRP-strengthened systems to delay crack initiation. The delay in the cracking load is an essential aspect of the serviceability of RC structures because the cracks facilitate the ingress of harmful agents, such as chlorides and carbon dioxide to the steel surface. By minimizing the cracks at the service loads, the corrosion initiation can be retarded, which could positively affect the overall durability of the concrete structural members strengthened with CFRP.

The exposure conditions reduced the total number of cracks formed before failure in both the control and strengthened specimens. For example, the numbers of cracks at failure in CB-28-L- and CB-28-L-2 were 30 and 32, respectively. Moreover, in CB-360-S-1 and CB-360-S-2, the cracks numbered 24 and 18, respectively. In the strengthened specimens, such as C-28-L-1 and C-28-L-2, the numbers of cracks were 70 and 80, respectively. In the CFRP-strengthened specimens conditioned at the laboratory, the numbers of cracks in the two-beam specimens (Specimens 1 and 2) were 45 and 47 at 180 days, 48 and 40 at 360 days, and 47 and 46 after 730 days. For two beams under saline water, the numbers of cracks were 43 and 35, 31 and 33, and 41 and 31 after 180, 360, and 730 days of conditioning, respectively. After sunlight exposure, the numbers of cracks were 34 and 36, 33 and 34, and 49 and 47 for two samples tested at 180, 360, and 730 days, respectively. The decrease in cracks was higher for saline exposure than sunlight exposure. The laboratory specimens also exhibited a drop in cracks over time but not as much as in the other two outdoor exposure types. The CB specimens also demonstrated that the number of formed cracks at failure reduced from 32 to 21. The reduced number of cracks in CB could be due to the hydration of the unhydrated concrete, which might have enhanced the modulus of rupture of the concrete. In CFRP specimens, because the strength is increased over time in concrete and strengthening composites, the cracks reduced and the width of the cracks increased.

Fig. 7 illustrates the midspan load versus deflection curves for CB and CFRP-strengthened beam specimens tested at 28 days. Fig. 8 presents the curves for the samples tested at 180 and 360 days, and Fig. 9 presents the curves for the specimens tested after 730 days for all three conditions. The performance of CFRP-strengthened specimens was evaluated against the CB at 28 days, laboratory specimens at the same age, and CBs conditioned for the same periods. From the load versus deflection graphs, the beam stiffness, yield loads, ultimate loads, and deflections at steel yielding and beam failure were recorded. Table 2 presents their average values for the two specimens after 28 days of curing and after 180, 360, and 730 days of exposure.

Figs. 7–9 reveal that the load–deflection curves were bilinear for CB and CFRP-strengthened specimens until failure. The initial low values due to the support and loading arm adjustments before the application of bending loads on the beams were filtered from all the curves. All beams exhibited similar stiffness until crack initiation because the whole uncracked cross-sectional area for all specimens was available for moment resistance. Once the cracks were initiated, the CFRP-strengthened beams exhibited higher stiffness compared with CBs (refer to Figs. 7–9). For illustration, the CFRP-strengthened specimens demonstrated 42% higher stiffness during this loading stage at 28 days, and 52% in specimens exposed to saline water for 360 days [Fig. 8(f)]. The second phase of the load–deflection curves was from steel yielding until the failure of the beam specimens. The stiffness of the CFRP beams in this region was also up to 40% higher than their CBs. This range of increased stiffness after crack initiation was observed in all exposed samples compared with their counterpart CBs.

The yielding of the main reinforcement was initiated at much higher loads in CFRP-strengthened specimens than the CBs (Table 2). The CB specimens tested at 28 days had an average yield and failure load of 89 and 90 kN, respectively. The CFRP-strengthened beams exhibited an average yield and failure load of 124 and 150 kN, respectively, which indicates an increase of about 40% in the yield load and 67% in the ultimate load-carrying capacity.

At 180 days of laboratory, saline water, and sunlight conditioning, the average yield loads for CB were 86, 89, and 92 kN, respectively, and they failed at 99, 102, and 103 kN, respectively. Moreover, the CFRP-strengthened specimens had an average yield load of 120, 117, and 115 kN, respectively. This shows an average increase of 40%, 32%, and 25% from their respective CB samples exposed to the laboratory and saline water, respectively. The ultimate loads were 150, 134, and 1,146 kN at 180 days for the laboratory, saline water, and sunlight exposure, respectively. This is an average increase of 53%, 32%, and 41% from their counterpart CBs, respectively. A decrease in the percentage enhancement in yield and failure loads occurred at 180 days compared to that of 28 days. This could be due to the loss of bond strength of the epoxy during the first six months. Moreover, CB specimens at 180 days exhibited higher ultimate strength compared with that at 28 days for the control specimens due to the continuous hydration of unhydrated cement, which also lowered the percentage of the strength enhancement for CFRP-strengthened specimens.

Figs. 8(d–f) shows the effects of three exposure at 360 days. The average yield loads of CBs were 89, 91, and 91 kN, whereas the failure loads were 101, 99, and 99 kN for the laboratory, saline water, and sunlight exposure, respectively. The CFRP specimens began yielding at 122, 120, and 119 kN and at ultimate loads of 150, 141, and 147 kN under laboratory, saline water, and sunlight exposure, respectively. Thus, increases occurred of 38%, 33%, and 30% in the yield loads over their CBs, laboratory, saline water, and sunlight exposure, respectively. In addition, the average increases in the ultimate load-carrying capacity were 49%, 42%, and 48%, respectively. Fig. 9 illustrates the behavior of the specimens tested after 730 days (two years) of exposure. The yielding of steel for CBs initiated at 99, 96, and 97 kN, whereas for CFRP beams, it was initiated at 121, 137, and 129 kN, respectively, for laboratory, saline water, and sunlight exposure. The ultimate load capacity in the CBs after 730 days was 109, 96, and 103 kN, respectively, for the laboratory, saline water, and sunlight exposure. For the CFRP, the average ultimate loads were 148, 164, and 156 kN for the laboratory, saline water, and sunlight exposure, respectively. The increase in the yield load in the CFRP-strengthened specimens was 22%, 42%, and 32% for the laboratory, saline water, and sunlight exposure, and the ultimate load was increased by 36%, 71%, and 51% compared with the CBs exposed to similar conditions. Thus, the sunlight and saline water exposure had no significant effect on the ultimate strength of the CFRP-strengthened RC beam, until 730 days. The load–deflection curves demonstrated no loss in beam stiffness, yield loads, or ultimate loads. In addition, these curves followed a behavior similar to those tested at 28 days or to those tested under laboratory conditions.

A decline in strength of the CFRP-strengthened specimens occurred at 180 days under saline water and sunlight exposure. Laboratory exposure demonstrated no loss in strength over the two years. The strength loss under outdoor exposure at 180 days was regained at 360 and 730 days of exposure. Such behavior has already been documented, where an initial loss in strength occurs, and then the strength of the CFRP-based specimens was enhanced (El-Hawary et al. 1998; Choi et al. 2012; Tatar and Hamilton 2016c). El-Hawary et al. (2000) observed an initial drop in the strength of the cylindrical samples strengthened with three different commercial epoxies and exposed to the harsh outdoor climate of the Arabian Gulf coastal region. The strength of the specimens was regained six months after the previous testing, which followed another drop in strength after six months. This suggests that epoxy-based materials are affected by the seasonal environment and fluctuations in temperature and humidity. The strength could be affected by the weather of the previous six months to one year before the testing dates. In this study, the samples were strengthened in April when the temperature is quite moderate in the Gulf region. The 180-day testing was conducted in the month of October. During this exposure period, the specimens were exposed to the weather of extreme temperatures of (≥50°C) and 80% to 100% RH, which is encountered in the months of June to September under sunlight and saline water exposure. The temperature on the concrete surface could have reached above T_g, which softens the epoxy. The temperatures remained in the proximity of 40°C to 50°C until October. This does not allow sufficient time for the softened epoxy to cure before the testing. Hence, lower strengths were observed in the samples tested at 180 days, and most of the failures were adhesive (i.e., the debonding of sheets with traces of concrete). In the case of specimen tests after 360 days (one year) of exposure, the specimens had already weathered the hot temperature cycle, where the softening of epoxy causes more reaction sites and cross-linking inside the epoxy matrix. Then, the winter temperatures allowed proper curing and hardening again. This increases the bond strength between the epoxy and concrete substrate. Hence, a regain in strength in all samples was observed.

No clear patterns were found for the ultimate strength differences between the CFRP-strengthened and CB specimens, and either no loss or a slight increase in the ultimate load capacity after exposure to sunlight and saline water was observed. For example, 71% and 51% increases in the load capacity of CFRP beams after 730 days occurred for saline water and sunlight exposure, respectively, compared with the counterpart CBs, whereas the increases were only 32% and 41% at 180 days, respectively. The increase in strength over prolonged exposure could be due to polymeric cross-linking in epoxy composites. This phenomenon is facilitated at higher temperatures and forms complex interactions in the epoxy matrix. Hence, it improves the interfacial bonding between the CFRP laminate and concrete (Choi et al. 2012; Al-Tamimi et al. 2015). A similar observation of enhancement in bond strength between the CFRP and concrete prisms, when exposed to the sunlight and saline water, was reported by Al-Tamimi et al. (2015). Cromwell et al. (2011) reported similar results in which, under hygrothermal exposure, the strength of the beams strengthened in flexure was enhanced. It was argued that additional post-curing at higher temperatures regenerates the reaction site; hence, adhesive bonds could increase. In addition, some particles of cement always remain unhydrated at the initial curing, and the hydration is accelerated based on the availability of moisture over time, enhancing the concrete strength. Hence, an increase in concrete strength could also be a reason for the increase in the ultimate strength of the strengthened beam.

Table 2 presents the ductility indices for CB and CFRP-strengthened specimens. The CB specimens had much higher ductility under all exposure conditions and testing durations. However, the CFRP specimens lost ductility at failure due to higher attained loads with lesser deflections. The μ_Δy values for CB over all exposure types were between 1.23 and 3.95. The lower value of 1.23 is only observed in samples where loading was stopped after yielding. This was to keep the beam integrity intact for another ongoing strengthening project. The specimens that were loaded until concrete crushing had values of between 2.06 and 3.93. That is an increase of up to 300% of deflection at failure when compared with that at yielding. For the CFRP-strengthened specimens, the μ_Δy values were between 1.52 and 1.95. Thus, the ductility loss was up to 159% in the CFRP specimens compared with CBs. Again, no clear pattern in the μ_Δy values was observed for the strengthened specimens after exposure to three different environments over two years. The samples tested after one year under laboratory, saline water, and sunlight exposure had an average value of 1.76, 1.71, and 1.76, respectively. After two years, the values were 1.75, 1.65, and 1.57, respectively. The highest value of 1.76 was observed for the sunlight-exposed samples. The slight decrease in the ductility in the samples at 730 days compared with 360 days was due to the gain in the ultimate load capacity.

The μ_Δu values for CB were from 1.03 to 1.70, which means up to 70% higher deflection at the failure load than at the ultimate load was observed. Lower values of μ_Δu for CB specimens are due to the flat load–deflection curves after yielding until failure. Hence, the ultimate loads are quasi-similar to the failure loads. The CFRP specimens had μ_Δu values quite close to 1 because the ultimate and failure loads occurred almost at the same deflections as for those specimens. This loss in ductility in strengthened samples remained in the same range for all exposure times. The exposure to harsh climatic conditions exhibited no visible trend (Table 2).

Fig. 10 depicts the load versus strains of flexural reinforcement and concrete in compression at the midspan beams tested at 28 days. Figs. 11–13 list the strains for the specimens conditioned under laboratory, saline water, and sunlight exposure after 180, 360, and 730 days, respectively. The curve for one sample from each exposure and time is presented.

The steel yielding strain as per the manufacturer was 0.0027 mm/mm. It is evident from Figs. 10 to 13 that the steel yielded in the CB at a load of between 86 and 92 kN and reached an ultimate strain of 0.0027. The CB specimens failed in flexure by steel yielding, followed by concrete crushing at the ultimate strains of up to 0.003 mm/mm. In some specimens, the strain gauges malfunctioned, and the complete behavior until failure was not observed at those locations. After yielding, the strain gauges recorded smaller values than those observed at yielding. For example, the steel strain gauges of the CB-360-L-1 and CB 360-S-1 specimens indicated unexpected behavior during loading. In the CFRP specimens, the slopes of the load versus the steel strain curves were higher than those of the CBs. After steel yielding, the crushing of concrete was not initiated. Instead, concrete kept taking strain; however, the slope of the curves was reduced (Figs. 10–13). Then, at the failure load, the concrete was crushed immediately due to the debonding of the CFRP sheet. Thus, the failure strains of the concrete in the strengthened specimens were also reduced compared with those of the CBs. This behavior remained similar in all specimens under the three exposure types and times.

Fig. 11 demonstrates the effects of laboratory exposure on the strain behavior of the CFRPs and CBs over time. From the samples placed in the laboratory, the steel strain value at 180 days was smaller than those at 360 and 730 days of exposure, respectively. The slope of the curve (stiffness) was slightly decreased for the 360-day specimens compared with the 180-day specimens. However, the specimens tested at 730 days again had a higher slope than the 180-day samples. Fig. 12 reveals that saline water exposure had no influence on the strain of flexure reinforcement over time (up to two years). The curve for the CFRP samples at 730 days exhibited higher stiffness than did those samples tested at 180 and 360 days. This is another indication that CFRP sheets and epoxy adhesives were not affected by saline water exposure. Instead, it had a positive effect on the durability of the strengthened concrete beams. The load versus the strain curves of the concrete in compression also exhibited similar behavior at up to 730 days for samples placed in the laboratory. The yield and ultimate crushing values were in close range for all samples at up to two years of exposure. In the case of saline water exposure, the concrete behavior was slightly different, according to the exposure time. Higher strains were achieved in the beams tested at 360 and 730 days compared with 180 days, before the crushing of the concrete. Fig. 13 presents the effects of sunlight exposure over 180, 360, and 730 days of exposure. Similarly, the exposure does not affect the concrete strains and beam stiffness. Instead, slightly higher stiffness with time was observed in the beams tested at 360 and 730 days compared with those at 180 days.

Fig. 14 compares the strains in the primary reinforcement and CFRP laminates in the specimens test at 28 days and the specimens tested under laboratory exposure at 180, 360, and 730 days. Figs. 15 and 16 present the values for specimens under the saline water and sunlight exposure for 180, 360, and 730 days, respectively. Strains in steel and CFRP laminates were similar until the first crack appeared in the strengthened beams. After cracks appeared, the strain in the CFRP laminates increased compared with the steel strain. This is attributed to the stress transfer to these laminates, as the concrete in the tensile zone seized to take any further stress after cracking. The strain in the laminates was also much higher compared with the steel strain, which is because these laminates are placed farther in depth from the neutral axis than the steel rebar, which generates higher tensile moments in the laminates. Up to 3 to 4 times higher strain values were endured by the CFRP laminates than the steel rebar at failure. Figs. 15 and 16 reveal that this behavior was similar in all the samples tested after each exposure period. For example, the exposure for two years had a negligible effect on the behavior of the CFRP laminates, strengthened beams, and bond between the concrete and laminates.

The degradation indices for the CFRP-strengthened beams after each exposure are graphically presented in Fig. 17. Table 3 lists the average values of the environmental degradation indices over two CFRP-strengthened specimens for each category. The severe climatic conditions could affect the CFRP-strengthened RC beams on three levels. The first is the degradation of RC beams caused by the corrosion of the steel reinforcement or by the chemical attacks on the concrete matrix. The second is the bond degradation between the CFRP laminates and concrete due to the damage caused to the epoxy matrix, and the third is damage to the CFRP sheets and its fibers.

In the cases mentioned, the residual strength of the CFRP-strengthened beams is reduced compared with the unexposed samples. The values of ESR, and the ratios, R_n, R_b, and R_b28 should be less than 1.0. The ESR values were 0.96, 0.98, and 1.05 for sunlight exposure and 0.88, 0.94, and 1.11 for saline water exposure after 6, 12, and 24 months, respectively. The flexural strength reduction factor R_n for saline water exposure was a minimum of 0.86 at 180 days, whereas after 730 days, it increased to 1.25. For sunlight exposure, R_n was 0.92, 0.99, and 1.11 after 180, 360, and 730 days, respectively. The bond strength reduction factor R_b was 0.61, 0.84, and 1.71 for saline water exposure at 180, 360, and 730 days, respectively. Under sunlight exposure, R_b was 0.81, 0.97, and 1.34, respectively. For the first 180 days, the strength was reduced compared with 28 days; however, it was recovered after 360 and 730 days (Fig. 17). The specimens tested at 730 days indicated higher strength compared with the samples tested at 28 days. Most degradation index values were closer to 1, indicating a minor loss or some gain in strength.

This reveals that, with age, the epoxy-based strengthening system matures by forming more cross-links in the epoxy matrix; hence, a slightly higher mechanical performance and durability is possible for the strengthened specimens. This observation has also been made by other researchers (Grace and Singh 2005; Choi et al. 2012; Al-Tamimi et al. 2015). When the climate temperature reaches close to T_g, the epoxy materials soften, which reactivates the cross-linking in the matrix. If allowed to cure and hardened again, the material could achieve higher strength (Nogueira et al. 2001; Choi and Douglas 2010).

Fig. 17 indicates that some loss in strength occurs in the samples exposed to both exposure types for 180 days. However, the strength was regained after 360 and 730 days, especially in the saline water exposure. This behavior of the epoxy-based strengthening systems has already been reported under brackish water and sunlight exposure (Choi et al. 2012; Al-Tamimi et al. 2015; Tatar and Hamilton 2016c). The barnacles formed due to the lengthy immersion under brackish water also contribute to the strength of the whole member. Moreover, they protect the concrete and strengthening systems from deterioration by forming a protective layer (Choi et al. 2012; Al-Tamimi et al. 2015; Al Nuaimi et al. 2020). In addition, the type of epoxy, mixing ratio of its parts, and applied amount significantly affect the durability and strength performance of strengthened concrete specimens (Toutanji and Gómez 1997; Choi et al. 2012; Al-Tamimi et al. 2015). For detailed chemistry involved in the cross-linking and creation of activation sites, readers are referred to the works of Liu et al. (2019), Choi et al. (2012), and Al-Tamimi et al. (2015).

The strength-reduction factors (C_E), which are employed during the analysis and design of CFRP-strengthened RC beams, are presented in Table 4 in comparison with those recommended by ACI 440.2R-17 (ACI 2017), which provides C_E values for interior and exterior exposure for conventional CFRP-strengthened members at 0.95 and 0.85, respectively. Based on the results of this study, we present a comparison between the recommended C_E for the CFRP-strengthened RC beams under saline water and sunlight exposure conditions. The reduction factors for bond strength R_b and flexural strength R_n were 0.60 and 0.85 for saline water, respectively, and 0.80 and 0.95 for sunlight exposure, respectively. Grace and Singh (2005) found similar values for CFRP sheets under saline and 100% humidity exposure. In contrast, Tatar and Hamilton (2016a) recommended a bond durability factor of 0.60 after applying accelerated hydrothermal conditioning to the CFRP-strengthened systems.

Except for one odd value of R_b28 of 0.53 for saline water exposure, based on the results of this study, the C_E recommended values by the ACI 440.2R-17 (ACI 2017) guidelines are reasonable and logical. These bond degradation factors are estimated based on the strengthening system without any anchorage of CFRP laminates to the web of the beam specimens. Environmental reduction factors recommended in this study could be important for the design and analysis of RC beams strengthened with CFRP, especially when the durability of strengthening systems under harsh climates is under consideration.

Because the behavior of the two tested specimens at each exposure type was very similar, the effects of the applied environment were properly covered. However, because the number of samples for each category was limited to only two, a complete statistical analysis to enhance the confidence in the durability behavior of the CFRP systems was not possible. Furthermore, the study was limited to two years of exposure, which is a short period considering the service life of a repaired RC structure. That is why a prolonged study exceeding two years is recommended in which a comprehensive study on cracking, stiffness, and failure modes could be conducted to increase the database on the durability aspect of strengthening systems.