Studies [14] [3] [4] [15] showed that temperatures lower than Tg  may have a limited effect on the bond strength in most cases. In other cases, it may result in a useful post-cure effect depending on the heating rate. Nonetheless, temperatures that approach or exceed the class transition temperature can be harmful on the bond strength.  
Kabir et al. [16] studied the impact of temperature lower than the Tg on the bond strength of CFRP sheet bonded with epoxy resin using wet lay-up technique. In this study, single shear test was implemented to explore the effect of a thermal cycles where specimens were exposed into a constant temperature of 40 oC for 5 hours followed by 7 hours of gradual cooling down to 30 OC. Specimens were tested after 70 cycles (35 days), 180 cycles (90 days) and 730 cycles (365 days). Results indicated a small increment in the bond strength which is attributed for two reasons 1) the maximum applied temperature was lower than Tg (47 oC); 2) the small range of temperature (30-40 oC) was not capable to produce thermal strain (thermal expansion and contraction). It was reported that a cohesive failure within the concrete substrate was the dominant failure mode for most specimens due to the increase of the bond strength. 
Ferrier et al. [3] utilized double lap shear test to studied the effect of high and low temperature on the bond strength between CFRP and concrete. All tested specimens were placed in thermal chamber for 3 hours before testing. Two strengthening systems were investigated which are unidirectional sheets applied in wet lay-up technique and pultruded FRP strips bonded by epoxy adhesive. Test results of the first system revealed that decreasing the temperature from 20 oC to -40 oC can reduce the ultimate load capacity by 5%. However, increasing the temperature from 20 oC to 60 oC and 80 oC can reduce the ultimate capacity by 38% and 70%, respectively. The failure mode was also shifted from a cohesive failure within concrete substrate into an adhesive failure within the bonded interface at temperature of 80 oC, which is higher than class transition temperature. For the second strengthening system, results revealed that decreasing the temperature from 20 oC to -20 oC can reduce the ultimate load capacity by 40%. However, increasing the temperature to the range 40-50 oC, which represents the service temperature range in many parts of the worlds, can reduce the ultimate load capacity by 16%. Increasing the temperature up to 100 oC can reduce the capacity by 30%. The author concluded that the second system is more sensitive to low temperature than the first one due to the differences in the characteristics of the materials and the higher axial stiffness of the second system.
Firmo et al. [4] investigated the influence of elevated temperature on the bond strength of the CFRP strips externally bonded to concrete substrate with epoxy adhesive. In this study, double-lap joint shear test was implanted to study two testing conditions. The first one was a steady condition where specimens were heated first to a specific predefined degree, then they were loaded until failure. In the second testing conditions (transit condition), specimens were placed under a certain level of sustain load, then they were heated by a constant rate until failure. Test results revealed a significant reduction in the bond strength when the temperature raised higher than class transition temperature. Moreover, it was found that changing testing condition had only limited effect. It is worth to note that at a temperature of 1.9 and 2.6 times the Tg, the residual tensile strengths were 32% and 23% of the residual strength at the ambient temperature, which represent a relatively high residual strength.   
Mikami et al. [17] investigated the effect of high temperature ( 100 and 180 oC) on the bond strength between FRP and concrete using a pull-off test [18]. In this study, all the specimens were strengthened with bi-directional CFRP sheets applied in wet lay-up technique. Then, they were placed into a chamber where specimens exposed into numerous thermal cycles at a temperature of 100 oC or 180 oC. The length of each cycle was two hours. Relative humidity was kept constant at 0% or 100% within the cycles. In this research, the accelerated aging behavior of water and temperature was also explored by immersing additional specimens in distilled water and salt water of the same temperature for the same number of cycles. Test results revealed that exposing the specimens into 250 cycles at 100 oC and 0 % relative humidity can reduce the bond strength by to 11 %. However, the reduction in the bond strength increased up to 92 % when the temperature increased to 250 oC under the same number of cycles and relative humidity. Most of the tested specimens developed an adhesive failure for the both cases.  

5       Effect of high temperature on axial strain distribution and effective bond length

Numerous studies [3][4]utilized double-lap and single-lap shear tests to explore the effect of elevated temperature on strain distribution along the FRP strips/sheet bonded to the concrete. Rising the temperature closer to the Tg will reduce the overall axial strain measured along the FRP in comparison with the one measured within normal ambient temperature (20-25 oC). It was noticed that the higher the temperature is, the higher the reduction is. This reduction is attributed to the decreasing in the strength and the stiffness of the materials especially the epoxy in the higher temperature. Nevertheless, at the same load level, axial strain recorded for higher temperature is more than the one of the lower temperature as shown in Figure 1. In other words, although increasing the exposure temperature can reduce the overall strain, it increases the axial strain measure of the same load level.    
 Two more important observations were made by studies [19] [12] [4]. First, increasing the exposure temperature shifts the stain distribution into a linear form, particularly for low and moderate load level, due to the dropping in the epoxy stiffness that smooths the stress/strain distribution [4]. Second, it is  detected that rising the temperature can also increase the effective bond length (least length required to transfer tensile stress from concrete substrate to FRP strips/sheet by direct shear)  more than the minimum length recommended by some design guidelines [19] [12] [4][16]. 
 
 
 
 
 
Figure 1:  axial strain distribution at temperatures (20, 55, 90, 120 oC) at axial load 5.5kN [4].

6       Effect of outdoor exposure on the bond strength

When it comes to the effect of natural ageing on the bond behavior between FRP and concrete, there are very limited numbers of studies. Kabir et al. [16] investigated the effect of outdoor environment in Sydney, Australia on the bond strength of CFRP attached via wet lay-up technique. Specimens were exposed to the following ranges of temperature (5-35 oC), humidity (4-100%) and solar exposure (1-33 MJm-2) for 2 months, 6 months, 12 months and 18 months. Results of single shear test indicated a 15.2% reduction in the bond strength after 6 months. However, this reduction decreased into 8.6 % after 18 months. Regarding failure mode, most specimens developed an adhesive failure within the CFRP to concrete interface. 

7       Effect of hygrothermal environment on the FRP interface

Benzarti et al. [12]  utilized the pull-off test [20] to study the effect of the accelerated ageing conditions (40 C temperature and 95% relative humidity) on the bond behavior of two strengthening systems which are the CFRP sheets bonded in wet lay-up technique and pultruded CFRP strips bonded with adhesive. Test results of the second system revealed that the average bond strength decreased from 2.7 MPa to 0.8 MPa after 13 months of exposure. Moreover, failure mode was shifted from a cohesive failure within a concrete substrate into a cohesive failure within the adhesive. However, bond strength and failure mode of the first system were not significantly affected by the ageing conditions. This difference in the behavior of the two systems attributed to the higher deterioration and sensitivity of the adhesive used in the second system towards the ageing conditions than the resin used in the first system.  
Benzarti et al. [12] also investigated the bond behavior between CFRP and concrete using the single-lap- joint shear test under 40 C temperature and 95% relative humidity. Test results of the specimens strengthened with CFRP strips bonded with epoxy adhesive indicated a slight reduction in the bond strength between CFRP strips and concrete after 13 months of exposure which conflicts the result obtained from the pull-off test. The authors attributed this contradictory finding into the increment of the effective bond length due to the elasto-plastic behavior of the adhesive. In other words, the shear capacity of the composite joint was slightly affected because of the modification in the load transfer mechanism due to the plastic deformation of the joint. Regarding the failure mode, the authors reported a shifting from a cohesive failure within concrete substrate into an adhesive failure. For the CFRP sheets bonded in wet layup technique, test results did not indicate any significant change in the bond strength. Nevertheless, the failure was shifted from a cohesive failure within concrete substrate into a resin failure within the interface.
Ceroni et al. [8] investigated two hygrothermal conditions. The first condition included water immersion combined with elevated temperature (30 and 40oC). The second condition was exposure to humid ambient (RH = 100%) in conjunction with elevated temperature (30oC). Results indicted a significant reduction in the bond strength up to 72% for the first condition after 4 days of water immersion. Two important findings were revealed from the test results (1) water immersion can cause the same effect of the moisture ambient but with shorter exposure period; (2) raising the temperature of the water can further reduce the exposure period for the same strength degradation. This finding can be utilized to decrease of the durability test. Failure was shifted from a cohesive failure within concrete substrate into cohesive within the epoxy resin. Finally, the authors reported a slight strength recovery when the specimens exposed to drying cycles for 30 days.
Maljaee et al. [13] investigated the effect of the hygrothermal cycles (temperature 10-50oC and relative humidity of 90%) on the bond behavior between GFRP and masonry. Test results revealed a slight increment in the bond strength due to the post-curing effect that increases the cross-linking. Post-curing effect can occur by exposing specimens into a thermal cycles with a temperature higher than the curing temperature and lower than the Tg. It is worth to mention that the authors also reported a 7% increment in the Tg after exposure to the hygrothermal which is attributed to the post-curing effect.   

8       Bond-Slip models

Due to the limited data available regarding the effect of the hygrothermal environment on the FRP-to-concrete interface, there is no model evaluate the degradation of the bond strength. However, there are numerous models proposed regarding the behavior FRP-to-concrete interface in the ambient and elevated temperature. In this section, some of these models are going to be reviewed to determine the suitable model that can be extended to include the effect of the hygrothermal conditions.    
Essentially, there are three methods used to develop bond-slip model of the FRP-to-concrete interfaces: 1) directly from the reading of the closed space strain gauges attached to the surface of the FRP sheets/plates [21] [22]; 2) using mesoscale finite-element modeling of the debonding failure combined with test data[23]; 3) from the global load-displacement responses of the FRP bonded to concrete [24] [10].
Dai et al. [10] extended his bond-slip model developed for ambient temperature [24] to make it applicable for elevated temperature. This model utilized two keys parameters: (1) the interfacial fracture energy (Gf­); (2) the interfacial brittleness index (B). Despite the simplicity of the proposed model (only two parameters need to be determined from the test data), it also has some weakness in attaining the highly accurate shape description of nonlinear bond-slip curves. In addition, the value of B was determined based on the load-strain distribution not based on the load-displacement curves due to the limited experiments reporting the load-displacement curves. Using the measured strain can show considerable variations from one place to another due to the discrete nature of concrete cracks and heterogeneity of concrete.  
Dai et al proposed the following expressions to determine interfacial fracture energy and interfacial brittleness index.
                                                                           (1)
                                                               (2)
Where Gf (T) and B(T) are the interfacial fracture energy and brittleness index at the elevated temperature, respectively; Gf0 (N/mm) and B0 (mm-1) are the interfacial fracture energy and brittleness index at the ambient temperature; and b2 = 3.206, b3 = 1.313, c1 = 0.485, c2 = 14.053, and c3 = 0.877 are constant derived from the available experimental data then. The value of Gf0  and Br should be determined from the shear test because they change from one system to another depending on the concrete strength and properties of adhesive[23] [22][25]. If such test cannot be performed, Dai et al. [10] reamended use the values from Lu et al. [23] model Gf0 = 0.545 N/mm (for normal compressive strength of 35 MPa and commonly used adhesive) and Br is within the range of 8 to 14.1 (for normal concrete with cylindrical compressive strength ranging from 15 MPa to 50 MPa)   
Dai et al. [10] also revised Bisby’s model [26] for the degradation of modules of elasticity at elevated temperature of prefabricated FRP to account for the degradation of the FRP sheets (wet layup strengthening system). It is important to mention that the revised model was only validated with data from Chowdhury et al. [27] [28]  due to the limited available test data (therefore, it is need to be validated with a bigger data to check its accuracy).
                                                        (3)
Where Ep0 and Ep(T) are the elastic modulus of FRP at the ambient temperature and elevated temperature T(oC), respectively; a1 = 0.729, a2 = 9.856, a3 = 0.609 are empirical factors.
The ultimate load that the FRP can resist before failure can be determined using the following formula:
                                                                (4)
                                                                                                                   (5)
Where PuT is the ultimate thermotical pull load due to both mechanical and thermal effects; Gf is the interfacial fracture energy at a specified temperature; bp and tp are the width and thickness of the FRP plate/sheet, respectively; bc and tc are the width and thickness of the concrete prism; αp and αc are the thermal expansion coefficient of FRP and concrete, respectively.
Dai et al. [10] proposed the following formulas to determine the bond-slip relationship
                                                                                (6)
                                                                                           (7)
                                                                      (8)
                                                                                                      (9)
                                                                              (10)
Where τ(x) is the shear stress at distance x from the loaded end; δ(x) is the slip at the same distance x ; ε(x) is the strain at the same distance x; P is the experimental load acting at the load end of the FRP; A is the area underneath the τ ⁓ δ curve which is equal to the interfacial fracture energy [ ; B is the brittleness index.
 
 

9       Observations

1-     There are limited number of experimental studies dedicated to investigating the bond behavior of the FRP-to-concrete interface under hygrothermal environment.
2-     Up to this moment, there is not any study investigated the effect of the hygrothermal environment on the bond behavior at a temperature higher than Tg  
3-     Up to now, most of the theoretical models proposed to determine the behavior of FRP-to-concrete interface are developed for ambient temperature and elevated temperature. Therefore, they may not applicable for the hygrothermal environment. 
4-     Epoxy resin cured in the ambient temperature often has a low Tg (45-82). Plasticization induced by moisture can reduce Tg more make it more vulnerable to temperature. Consequently, more studies about the effect of the hygrothermal environment on the FRP resin and FRP coupons cured in the ambient temperature are required.    
5-     None of the available studies have investigated the effect of the heating procedure on the bond strength. By heating procedure, it is meant the difference the between thermal cycle and monotonic increasing of the temperature up to predefined limit and maintain that limit until test.
6-     There are very limited number of studies dedicated to investigating the effect of the elevated temperature on the FRP interface under sustain load.     

10   State of the Art

Vast number of studies dedicated to study the structural failure, design parameters, and failure modes of externally bonded FRP. The performance of this system is dominated by the bond behavior between FRP and concrete. This bond behavior can be greatly influenced by the environmental factors that can lead into premature failure. Understanding the long-term behavior and degradation of this system under various environmental factors is as important as the short-term performance. Despite of the extensive literature dedicated for studying the durability of the FRP system, most of these studies focused on the effect of an isolated degrading agent such solar radiation (UV), freeze and thaw, moisture and temperature. Thus, less attention was given to the combined effect of two or more agents, which represents the real case situation. The hygrothermal environment including the combined effect of moisture and temperature is considered the most aggressive environment for FRP [7] [8] [13]. On one hand, moisture can plasticize and degrade the epoxy bonding between FRP and concrete through the hydrolytic breakdown between matrix and FRP. Moreover, it can reduce the Tg.  On the other hand, temperature close or higher than the Tg can significantly reduce the bond strength [14][3][4]. Temperature coupled with moisture can accelerate the degradation by increasing the rate of moisture absorption.
In this research, the behavior of FRP sheets bonded with epoxy resin in wet layup technique is going to be investigated. There are two reasons for investigating this system (1) this system is widely used in the rehibition and strengthening of the masonry wall and infrastructure that is commonly exposed to hygrothermal environment; (2) since It has a lower Tg, it is expected to be more effected by hygrothermal environment especially in a temperature close to the service temperature.
The experimental program will explore the degradation of the bond after exposure into two hygrothermal conditions which are (1) moisture combined with different ranges of temperature; (2) water immersion combined with various range of temperature. Single-lap shear and pull-off test will be used to investigate bond degradation as shown in the Figure. The effect of the hygrothermals conditions on the modules of elasticity of the FRP and Tg will be also evaluated and measured. Finally, an analytical model is going to be proposed to determine the degradation of the bond strength.
a
b
 
         
           
 
 
Figure 1: a) single-lap shear test; b) pull-off test
Reference
[1]       J. G. Teng, J. F. Chen, S. T. Smith, and L. Lam, FRP-strengthened RC structures. Chichester, U.K.: Wiley, 2002.
[2]       L. C. Hollaway and J.-G. Teng, Strengthening and rehabilitation of civil infrastructures using fibre-reinforced polymer (FRP) composites. Elsevier, 2008.
[3]       E. Ferrier, O. Rabinovitch, and L. Michel, “Mechanical behavior of concrete-resin/adhesive-FRP structural assemblies under low and high temperatures,” Constr. Build. Mater., vol. 127, pp. 1017–1028, 2015.
[4]       J. P. Firmo, J. R. Correia, D. Pitta, C. Tiago, and M. R. T. Arruda, “Experimental characterization of the bond between externally bonded reinforcement (EBR) CFRP strips and concrete at elevated temperatures,” Cem. Concr. Compos., vol. 60, pp. 44–54, 2015.
[5]       ACI committee 440, Guide for the design and construction of externally bonded FRP systems for strengthening existing structures. 2008.
[6]       B. Ghiassi, G. Marcari, D. V. Oliveira, and P. B. Lourenço, “Water degrading effects on the bond behavior in FRP-strengthened masonry,” Compos. Part B Eng., vol. 54, pp. 11–19, Nov. 2013.
[7]       R. Sen, “Developments in the durability of FRP-concrete bond,” Constr. Build. Mater., vol. 78, pp. 112–125, 2015.
[8]       F. Ceroni, A. Bonati, V. Galimberti, and A. Occhiuzzi, “Effects of environmental conditioning on the bond behavior of FRP and FRCM systems applied to concrete elements,” J. Eng. Mech., vol. 144, no. 1, p. 4017144, 2018.
[9]       J. L. Clarke, “EUROCOMP design code and handbook: Structural design of polymer composites,” E FN Spon, London, 1996.
[10]     J.-G. Dai, W. Y. Gao, and J. G. Teng, “Bond-Slip Model for FRP Laminates Externally Bonded to Concrete at Elevated Temperature,” J. Compos. Constr., vol. 17, no. 2, pp. 217–228, 2013.
[11]     International Standard Organaziation ISO and 527-1, “Plastics Determination of tensile properties – Part 1: General principles,” Geneva, Switzerland., 2012.
[12]     K. Benzarti, S. Chataigner, M. Quiertant, C. Marty, and C. Aubagnac, “Accelerated ageing behaviour of the adhesive bond between concrete specimens and CFRP overlays,” Constr. Build. Mater., vol. 25, no. 2, pp. 523–538, 2011.
[13]     H. Maljaee, B. Ghiassi, P. B. Lourenço, and D. V. Oliveira, “FRP-brick masonry bond degradation under hygrothermal conditions,” Compos. Struct., vol. 147, pp. 143–154, 2016.
[14]     J. R. Cromwell, K. A. Harries, and B. M. Shahrooz, “Environmental durability of externally bonded FRP materials intended for repair of concrete structures,” Constr. Build. Mater., vol. 25, no. 5, pp. 2528–2539, 2011.
[15]     H. Blontrock, “Analysis and modeling of the fire resistance of concrete elements with externally bonded FRP reinforcement,” Ghent University, 2003.
[16]     M. I. Kabir, R. Shrestha, and B. Samali, “Effects of applied environmental conditions on the pull-out strengths of CFRP-concrete bond,” Constr. Build. Mater., vol. 114, pp. 817–830, 2016.
[17]     C. Mikami, H. C. Wu, and A. Elarbi, “Effect of hot temperature on pull-off strength of FRP bonded concrete,” Constr. Build. Mater., vol. 91, pp. 180–186, 2015.
[18]     ASTM D4541 – 09, “Standard Test Method for Pull-Off Strength of Coatings Using Portable Adhesion,” 2009.
[19]     M. Leone, S. Matthys, and M. A. Aiello, “Effect of elevated service temperature on bond between FRP EBR systems and concrete,” Compos. Part B Eng., vol. 40, no. 1, pp. 85–93, 2009.
[20]     EN 1542, “Products and systems for the protection and repair of concrete structures. Test methods. Measurement of bond strength by pull-off.,” 1999.
[21]     K. Nakaba, T. Kanakubo, T. Furuta, and H. Yoshizawa, “Bond behavior between fiber-reinforced polymer laminates and concrete,” Struct. J., vol. 98, no. 3, pp. 359–367, 2001.
[22]     B. A., F. C., M. E., and N. E., “Indirect Identification Method of Bilinear Interface Laws for FRP Bonded on a Concrete Substrate,” J. Compos. Constr., vol. 16, no. 2, pp. 171–184, Apr. 2012.
[23]     X. Z. Lu, J. G. Teng, L. P. Ye, and J. J. Jiang, “Bond-slip models for FRP sheets/plates bonded to concrete,” Eng. Struct., vol. 27, no. 6, pp. 920–937, 2005.
[24]     J. Dai, T. Ueda, and Y. Sato, “Development of the Nonlinear Bond Stress–Slip Model of Fiber Reinforced Plastics Sheet–Concrete Interfaces with a Simple Method,” J. Compos. Constr., vol. 9, no. 1, pp. 52–62, 2005.
[25]     T. Houssam, H. Meng, and G. Elhem, “Interfacial Bond Strength Characteristics of FRP and RC Substrate,” J. Compos. Constr., vol. 16, no. 1, pp. 35–46, Feb. 2012.
[26]     L. A. Bisby, Fire behaviour of fibre-reinforced polymer (FRP) reinforced or confined concrete. Queen’s University, 2003.
[27]     E. U. Chowdhury et al., “Mechanical characterization of fibre reinforced polymers for numerical fire endurance modelling,” in Proceedings of the 5th International Conference: Structures in Fire, 2008, pp. 28–30.
[28]     E. U. Chowdhury, R. Eedson, L. A. Bisby, M. F. Green, and N. Benichou, “Mechanical Characterization of Fibre Reinforced Polymers Materials at High Temperature,” Fire Technol., vol. 47, no. 4, pp. 1063–1080, 2011.