Performance of insulated FRP-strengthened concrete flexural members under fire conditions

Abstract

This paper presents the results of fire resistance tests on carbon fiber-reinforced polymer (CFRP) strengthened concrete flexural members, i.e., T-beams and slabs. The strengthened members were protected with fire insulation and tested under the combined effects of thermal and structural loading. The variables considered in the tests include the applied load level, extent of strengthening, and thickness of the fire insulation applied to the beams and slabs. Furthermore, a previously developed numerical model was validated against the data generated from the fire tests; subsequently, it was utilized to undertake a case study. Results from fire tests and numerical studies indicate that owing to the protection provided by the fire insulation, the insulated CFRP-strengthened beams and slabs can withstand four and three hours of standard fire exposure, respectively, under service load conditions. The insulation layer impedes the temperature rise in the member; therefore, the CFRP–concrete composite action remains active for a longer duration and the steel reinforcement temperature remains below 400°C, which in turn enhances the capacity of the beams and slabs.

Keywords

concrete beams / concrete slabs / carbon fiber-reinforced polymers / fire resistance / FRP strengthening / repair / retrofitting

Introduction

Over the last few decades, fiber-reinforced polymers (FRPs) have emerged as effective and efficient materials for the retrofitting and strengthening of structural members fabricated from reinforced concrete (RC). The high strength and low weight, accompanied by excellent durability and corrosion resistance properties, make FRPs a preferred option over steel for strengthening applications.

Despite the aforementioned advantages, the use of FRPs in indoor applications is rather limited because FRPs have inferior fire resistance properties owing to the glass transitioning of the matrix polymer and bonding adhesives at temperatures as low as 60°C to 82°C [¹]. This transitioning of the polymer causes rapid deterioration in the modulus and strength properties of FRPs, while the softening of bonding adhesives causes the degradation of the FRP–concrete interfacial bond, thereby decreasing the overall stiffness of the strengthened concrete member. As a result, a strengthened concrete member has lower fire resistance compared with an un-strengthened one.

The provision of satisfactory structural fire resistance is a primary requirement in building design [²,³]. Typically, normal-strength concrete (NSC) structural members provide better fire performance than high-strength concrete (HSC) or high-performance concrete (HPC) ones. This is because of the slower deterioration in the strength and modulus of NSC [⁴,⁵]. In contrast, HSC and HPC undergo rapid deterioration in strength at elevated temperatures [⁶] and often experience a decrease in the cross-section owing to fire-induced explosive spalling, which primarily stems from low permeability of HSC/HPC [⁷]. To reduce such spalling in HSC structural members, several researchers have recommended the addition of fibers (steel or polypropylene) to the concrete mix [⁷,⁸]. Several studies [⁹–¹⁵] concluded that the addition of fibers in HSC reduces spalling and enhances the fire resistance of HSC structural members. Thus, concrete structures fabricated from NSC have good fire resistance prior to strengthening and do not require any additional fire protection to satisfy the fire safety requirements of design standards [²]. However, when an existing concrete member is strengthened by adding an external layer of FRP, the strengthened member exhibits lower performance under fire than the original un-strengthened member, as explained earlier. Therefore, the fire response of strengthened RC flexural members must be verified.

The fire response of strengthened concrete beams has been evaluated in several tests [¹⁶–²⁵]. However, only limited fire tests [²⁶–²⁸] have been conducted on FRP-strengthened RC slabs. The strengthening level, type of insulation material as well as layout and thickness, fire scenario, applied loading, and restraint conditions were some of the major factors examined in these tests. These studies recognized the degradation of the FRP–concrete interfacial bond as a primary cause of failure in a strengthened concrete member exposed to fire. These studies demonstrated that the fire resistance of strengthened concrete members can be increased by applying fire insulation over the strengthening system. Additionally, these studies demonstrated that limiting the increase in temperature at the end zones when FRPs are attached to beams (i.e., anchorage zones) by applying thicker insulation significantly improves the fire performance of strengthened structural members.

All the aforementioned studies evaluated the overall fire resistance of strengthened concrete members; however, only a few studies [²¹,²³] provided any insights into the behavior of the strengthening system during fire exposure. Furthermore, these tests did not provide any recommendations for the optimal layout and thickness of the fire insulation to achieve satisfactory performance under fire. As a thin insulation layer may not yield a reasonable enhancement of fire resistance, and an excessively thick layer could result in the delamination of the insulation from the concrete surface owing to the insulation’s additional self-weight, thereby leaving the strengthened structure unprotected. Furthermore, the type of insulation and its associated thermal properties dictate the fire performance of a strengthened concrete member. Therefore, whether the FRP-strengthened concrete member would achieve necessary fire rating requirements even after insulation is applied is significant.

To address these concerns related to the type of insulation, its thickness, and its adhesion properties, Structural Technologies Strongpoint LLC developed a strengthening scheme and a spray-applied fire protection system. An experimental program was conducted at Michigan State University to examine the efficacy of the fire insulation material in maintaining the CFRP–concrete composite action under fire conditions. In this program, four CFRP-strengthened beams slabs protected with an insulation material were subjected to thermal loading in the form of the ASTM E119 “standard fire” test under sustained structural loading. Furthermore, a case study on the rational fire design of strengthened beams was conducted using a previously developed numerical model, validated using the test data.

Fire tests on strengthened flexural members

The experimental program comprised fire tests on RC flexural members strengthened with V-Wrap C200HM FRP and insulated with V-Wrap FPS fire protection system insulation material. The details of the fire tests, such as the fabrication and instrumentation of the specimens, test setup and procedure followed, as well as the response parameters measured, are discussed in the following sections.

Test specimens

Four flexural members, i.e., two RC T-beams designated as TB1 and TB2, and two RC slabs designated as S1 and S2, were designed and fabricated to conduct the fire tests. The T-beam and slab specimens were designed and strengthened as per provisions of ACI 318-14 [²⁹] and ACI 440.2R-17 [¹], respectively. The dimensions of these flexural members were selected to represent typical residential or office building flexural elements.

Figure 1 shows the geometry of the test specimens. Both the beams and slabs were 3960 mm long. The T-beams had cross-section dimensions (width × depth) of 432 mm × 127 mm and 254 mm × 279 mm for the flange and web, respectively, whereas the slabs had cross-section dimensions (width × depth) of 406 mm × 152 mm. TB1 and TB2 were reinforced at the bottom with 3-19 mm φ (#6) and 3-16 mm φ (#5) bars, respectively, and at the top with 4-12 mm (#4) bars. Additionally, both beams consisted of 10 mm φ (#3) stirrups and transverse rebars (in the flange) spaced at 152 mm c/c and 305 mm c/c, respectively. Both slabs were reinforced with 3-12 mm (#4) bars along the length and 8-12 mm (#4) rebars along the width. The longitudinal reinforcement had a clear cover of 38 mm in the beams and 19 mm in the slabs.

A carbonaceous-aggregate-based concrete batch mix was used to fabricate the specimens. The concrete batch mix had a compressive strength of 43 MPa at the end of 28 d, which increased to 46 MPa on the test day. The flexural (bottom) steel reinforcement in the beams and slabs had a yield strength of 500 MPa. The specimens were cast within the forms and sealed for 31 d. Thereafter, the forms were removed, and the specimens were stored at 10°C–40°C for six months in the Civil Infrastructure Laboratory at Michigan State University.

FRP strengthening and insulation installation

In retrofitting applications, concrete structural members are typically strengthened to enhance the flexural capacity by 20% to 45%. To achieve this increased capacity in the beams and slabs, one layer of V-Wrap C200HM FRP was installed in the tension zone (at the soffit) of the beams and slabs over a length of 3660 mm. Before the CFRP sheets were installed, the soffit of the beams and slabs was sandblasted to roughen the surface and partially expose the aggregates. Specimen surface preparation and installation of the CFRP sheets were conducted according to the specifications provided by the manufacturer. The strengthened specimens were then cured for 24 h, as recommended by the manufacturer.

The dimensions of the CFRP sheets applied to the beams and slabs are summarized in Table 1. A 100 mm × 1 mm (width × thickness) CFRP sheet was applied to the beams (TB1 and TB2), while a 75 mm × 1 mm CFRP sheet was applied to the slabs (S1 and S2). The cured CFRP laminate had a design tensile strength of 1172 MPa and failure strain of 0.011, with a design tensile modulus of 96.5 GPa. The epoxy polymer (V-wrap 700S) used for bonding the CFRP sheet had a tensile strength of 45 MPa and tensile modulus of 2.07 GPa with a rupture strain of 0.055 measured according to the ASTM D638 [³⁰] method. Furthermore, the bonding adhesive had a glass transition temperature (T_g) of 80°C, as reported in the technical data sheet (evaluated in accordance with ASTM D4065 [³¹] specifications). The applied CFRP strengthening increased the moment capacity of TB1 and TB2 by 23% and 33%, respectively, and that of S1 and S2 by 40%, computed according to ACI 440.2R [¹]. These levels were based on typical strengthening levels recommended by design standards for building applications.

After the installation and curing of the CFRP sheet, the beams and slabs were insulated with a V-Wrap FPS. The V-Wrap FPS is an incombustible insulation material, primarily comprising vermiculite, gypsum, Portland cement, and a few other additives. The thermal conductivity, density, specific heat, and bond strength of the V-Wrap FPS, as reported in the technical sheet supplied by the manufacturer, was 0.156 W/m·K, 425 kg/m³, 1888 J/(kg·K), and 105 kPa, respectively.

The insulation material was spray-applied in several passes at the soffit and sides of the beams and slabs using a hopper gun. The thickness of the insulation layer applied to the test specimens is listed in Table 1. The insulation was applied on either side of the beams’ webs up to a depth of 152 mm and along the entire depth on both sides of the slabs, maintaining the same thickness as at the bottom of the specimens. A uniform insulation thickness with a tolerance of ±3 mm was maintained throughout the length of the beams and slabs.

Instrumentation

The temperature progression was monitored at a quarter and half the span length of the test specimens using seven K-type thermocouples installed at each of the sections (Fig. 2). The thermocouples were designated as TC1, TC2, TC3, TC4, TC5, TC6, and TC7. These thermocouples measured the temperature rise at insulation-FRP interface, FRP-concrete interface, stirrup, mid-rebar, corner rebar, mid-depth of concrete, and top rebar, respectively, in beams, whereas in case of slabs, the thermocouples measure the temperature rise at insulation-FRP interface, FRP-concrete interface, mid-rebar, corner rebar, temperature rebar, mid-depth and top surface of concrete, respectively. The strains in the tensile and compressive rebars of the beams and in the tensile rebars of the slabs were recorded by installing normal temperature strain gauges, SG1 and SG2 on rebars (Fig. 2) at the mid-span section. In case of slabs the strains were recorded at the mid-rebar and top of concrete surface at the mid-span section through normal temperature strain gauges, SG1 and SG2, respectively. The vertical deflection was recorded at the mid-span of the test specimens using two displacement transducers placed along the center line at the top of the specimens. Additionally, for the beams, the vertical deflection was measured at one of the load extensions.

Test furnace and procedure

The strengthened flexural members were tested in two separate fire resistance tests conducted in a structural fire testing furnace (cf., Fig. 3) located at the Michigan State University. The testing facility comprises a heating chamber that is 1.68 m high and has a cross-section dimension of 3.05 m × 2.44 m and loading actuators mounted on a steel framework, which in turn is mounted on four steel columns. During a fire test, the heat (thermal energy) is provided within the chamber through six propane burners, and the temperature is monitored using six K-type thermocouples. The test furnace can simultaneously apply both structural and thermal (heating) loading conditions.

TB1, TB2, S1, and S2 were tested simultaneously in the first and second fire tests, respectively. During each fire test, the specimens were simply supported at the ends with a clear span of 3.66 m and were subjected to two concentrated loads at a distance of 430 mm on either side of mid-span. At the beginning of the test, a pre-determined structural load was gradually applied to the specimens through a pneumatically driven hydraulic actuator until a constant deflection was observed. Subsequently, the gas burners in the furnace were turned on and the fire temperatures were simulated according to the time–temperature requirements of the ASTM E119 [³²] standard fire test while the structural loading on the specimens was kept constant.

The total structural load applied by the actuator on each specimen is summarized in Table 1. Total loads of 128, 97, 21, and 26 kN were applied to TB1, TB2, S1, and S2, respectively. These applied loads represented the service load as required by ASTM E119 [³²] for fire testing conditions and corresponded to approximately 50%–55% of the ultimate strengthened capacity of the beams and slabs.

During the tests, TB1 and TB2 were exposed to fire on three surfaces, i.e., two sides and the soffit, while only the soffits of S1 and S2 were subjected to fire exposure. Additionally, only 2.44 m of the unsupported length of the specimens was exposed to fire temperatures, while the remainder were thermally protected by the furnace walls and thermal blankets. This configuration was adopted to evaluate the beneficial effects of maintaining a low temperature in the CFRP sheets in the anchorage zones. The fire exposure on the beams and slabs was continued for four and three hours, respectively, during which the mid-span deflection, strains, and temperatures rise in the cross-section were monitored.

Test results and discussion

The effects of different insulation thicknesses, strengthening levels, and load levels on the behavior of CFRP-strengthened RC flexural members under fire were evaluated using the measured test data. The behavior was analyzed separately for TB1 and TB2 and S1 and S2 by comparing the thermal response, i.e., cross-sectional temperatures and structural response, i.e., mid-span deflections. Although the temperatures were measured at the quarter and mid-span sections, the assessment of the test data indicated that the temperature progression measured at both sections followed a similar trend. Therefore, only the temperatures measured at the mid-span section of the beams and slabs are presented here.

Results from the test on the beams

Thermal response

To evaluate the thermal response of beams TB1 and TB2, the time-temperature progression at the mid-span section in both beams is plotted in Fig. 4. The measured furnace temperature was compared with the standard tire temperature as defined in ASTM E119 [³²] (Fig. 4(a)). The furnace temperature was slightly higher than the standard fire temperature. However, the deviation of the area under the furnace fire and ASTM E119 standard fire curves, for a duration of four hours, was within the permissible limits (<5%). Hence, the temperature in the furnace fire was considered equivalent to the standard fire temperature.

Figure 4(a) shows the increase in temperature measured by thermocouples TC1 and TC2 (at the insulation–FRP and FRP–concrete interfaces) of TB1 and TB2. The temperatures at both the interfaces in TB1 increased at a faster rate than the temperatures at the interfaces in TB2 throughout the fire exposure duration owing to the smaller thickness of the insulation layer on TB1. For instance, the FRP–concrete interface temperature exceeded the adhesive T_g after 18 and 44 min of fire exposure in TB1 and TB2, respectively. The temperature change at TC1 and TC2 in TB2 plateaued after approximately 15 min of the start of the fire because of the energy consumed in the evaporation of chemically bound moisture in the insulation.

The temperature at TC1 and TC2 of TB1 abruptly increased to 800°C after 170 min of fire and then continued to increase rapidly until the end of the fire exposure. This abrupt increase in temperature was due to the localized delamination of the fire insulation at the mid-span of TB1, which was confirmed through visual observations during the fire tests. For TB2, the temperature at both interfaces increased rapidly after 180 min of fire exposure. This rapid increase was attributed to the formation of cracks in the insulation, resulting in an increase in the heat flow, which in turn resulted in higher measured temperatures. After approximately 220 min of fire exposure, the temperature at both the interfaces in TB2 increased abruptly to 1000°C, indicating a delamination of the insulation.

Figure 4(b) compares the temperature progression measured in TC4, TC5, and TC6 (i.e., bottom reinforcing bars and at the middle of the concrete, cf., Fig. 2(a)) of TB1 and TB2. The figure shows that at any time instant, the temperature rise in the concrete and rebars of TB2 was slightly lower than that in TB1. This was because the insulation layer applied on TB2 was thicker than that on TB1. In both beams, the temperature at the steel rebars and mid-depth of concrete increased gradually, followed by a concrete water evaporation plateau at 100°C. Furthermore, the temperature in the steel rebars remained below 400°C in TB1 and below 350°C in TB2 until the end of fire exposure. Additionally, the highest temperature recorded in concrete at the mid-depth of the cross-section was below 270°C (TB1) and 250°C (TB2). This indicated that the strength of the steel rebars and concrete degraded slightly.

Structural response

Figure 5 shows the progression of deflection with time in TB1 and TB2. At the beginning of the fire exposure, the deflection in both beams increased gradually at the same rate; however, the deflection began to increase rapidly after approximately 100 and 150 min of fire exposure in TB1 and TB2, respectively. This was because of the degradation of the CFRP–concrete interfacial bond as the interface temperature exceeded 200°C, which was significantly higher than the adhesive T_g (80°C). The degradation of the bond resulted in a complete loss of the CFRP–concrete composite action, which in turn reduced the stiffness of the beams. Note that the deflection in the beams increased rapidly at a much later stage of the fire exposure, although the temperature at the FRP-concrete interface of TB1 and TB2 exceeded the adhesive T_g at 18 and 44 min of fire exposure, respectively. This infers that the attainment of T_g at the interface did not result in an immediate loss of the FRP–concrete bond; hence, the FRP continued to contribute to the strength and stiffness of the beams at temperatures much higher than the adhesive T_g.

In both beams, the deflection remained relatively small until the end of the fire exposure, i.e., four hours. The maximum deflection observed at the end of four hours in TB1 and TB2 was 40 and 27 mm, respectively. These lower deflections were attributed to the lower temperature in the reinforcing bars and concrete, as discussed in the previous section, which reduced the degradation of the strength of the steel and concrete; this in turn reduced the deflection of the beams. Additionally, the CFRP sheets in the cooler anchorage zones of the beams remained intact. Therefore, carbon fibers detached from the beam at the mid-span owing to bond degradation continued to bear the tensile forces through the cable mechanism, thereby reducing deflection in the beams. Similar behavior was reported by Ahmed and Kodur [²⁰], Firmo et al. [³³], and Firmo and Correia [²³].

Fire resistance

The temperature, strength, and deflection criterion specified in ASTM E119 [³²] were used to determine the failure and fire resistance times of the beams. TB1 and TB2 were subjected to 4 h of the ASTM E119 [³²] standard fire test under sustained structural loading. During this entire fire exposure, the maximum temperatures in the steel reinforcements of TB1 and TB2 were recorded to be 400 and 293°C, respectively. These temperatures are below the critical temperature of 593°C for tensile reinforcement specified in ASTM E119 [³²]. Moreover, the maximum mid-span deflections in TB1 and TB2 were 30 and 16 mm, respectively. These deflections are below the critical deflection limit (L_c²/400d = 82 mm), as specified in ASTM E119 [³²]. Moreover, the strengthened and insulated RC T-beams withstood the service loads for the entire four hours of fire exposure. Therefore, the beams achieved four hours of fire-resistance rating according to the ASTM E119 [³²] limiting criterion.

Results from the tests on CFRP-Strengthened RC slabs

Thermal response

Figure 6(a) shows the comparison of the measured furnace temperatures with the ASTM E119 standard fire temperatures. The temperature in the furnace varied rapidly in the beginning and slightly exceeded the standard fire temperature after approximately 15 min of the ignition of the burners. The furnace temperatures remained slightly higher than the standard fire temperature for a significant duration of the test; however, the deviation in the area under the furnace fire and ASTM E119 standard fire curves was less than 5% during three hours of fire exposure. Hence, according to the ASTM E119 [³²] specifications, the furnace fire curve could be considered equivalent to the standard fire curve.

Figure 6(a) also shows the temperatures measured at thermocouples TC1 and TC2 (at the insulation–CFRP and CFRP–concrete interfaces) of slabs S1 and S2. The temperature at both the interfaces in S1 and S2 increased gradually after the moisture evaporation-induced plateau. The temperature rise measured at TC1 in S2 was higher than that at TC1 in S1. This can be attributed to the increased heat flow from the cracks developed in the insulation because of the higher load level applied on S2 during the fire test. However, a closer examination of the measured temperature change at the quarter span (not presented in this paper) indicated the expected trend, i.e., the temperature rise in S1 was higher than that in S2 throughout the fire test. Owing to the protection provided by the V-Wrap FPS, the temperature at the FRP–concrete interface (TC2) in both slabs remained below 300°C during the entire test, which was significantly below the decomposition temperature of the polymer matrix.

Figure 6(b) shows the time–temperature progression measured in TC3, TC4, and TC5 (rebars and middle of the concrete cf. Fig. 2(b)) of S1 and S2. The figure shows that the temperature in the rebars and concrete increased at a slower rate until the end of the fire test. Furthermore, the temperature change at the same locations in both slabs had negligible differences. This was attributed to the relatively smaller difference in the thickness of the fire insulation. The temperatures in the steel rebars and concrete remained below 300°C in both S1 and S2 until the end of the fire test. Therefore, it can be inferred that the steel rebars and concrete experienced minimal degradation in their strength and modulus properties, and as a result, the slabs retained most of their strength and stiffness until the end of the test.

Structural response

Figure 7 compares the measured progression of deflection with time in S1 and S2. The deflection in S1 increased gradually and remained low throughout the fire test, whereas for S2, the deflection increased very rapidly from the beginning of the fire exposure despite the low temperature in the concrete and steel rebars. This was attributed to the relatively higher loading applied on S2 compared with S1. Because of the relatively higher loading, S2 experienced a surge in the deflection (runoff deflection) at approximately 150 min. However, this increase in deflection did not cause the slab to fail. The maximum deflections observed at the end of fire exposure in S1 and S2 were 39 and 100 mm, respectively.

Fire resistance

During this entire fire exposure time, the maximum temperatures in the tensile steel rebars of S1 and S2 were recorded as 316°C and 232°C, respectively. These temperatures were below the critical temperature of 593°C for steel rebars according to ASTM E119 [³²] specifications. The maximum mid-span deflection in S1 and S2, during the fire exposure time, was below the critical limit (L_c²/400d = 219.4 mm) specified in ASTM E119 [³²]. Moreover, S1 and S2 sustained the applied service loads throughout the fire exposure. Thus, both slabs performed satisfactorily during fire exposure and achieved a 3 h fire resistance according to the ASTM E119 criterion.

Numerical evaluation of fire performance of strengthened flexural members

A macroscopic finite-element-based numerical model was applied to evaluate the fire behavior of the tested beams and slabs. The model was originally developed by Kodur and Ahmed [³⁴] and was later extended by Kodur and Bhatt [³⁵]. The analysis procedure followed in the model is briefly described in the following section; however, for a detailed description of the model, the reader is encouraged to refer to the above-cited articles.

Analysis procedure

The numerical model traces the thermo-mechanical response of a structurally loaded and fire-exposed strengthened concrete structural member through a sequentially coupled thermal and structural analysis at incrementing time steps. At the beginning of the analysis, the length and cross-section of the structural members (beam or slab) were discretized into segments and rectangular elements (Fig. 8). Subsequently, fire loading was applied to the structural member through predetermined time–temperature relationships. Subsequently, a heat transfer analysis was conducted to compute the temperature rise in each element of the cross-section. The computed temperatures were applied as thermal loading on the member; subsequently, the temperature-dependent moment–curvature relationships at various segments were computed, which in turn were used to compute the moment capacity and deflection of the structural member. The capacity and deflection were then compared with the applicable limit states, which are stipulated in ASTM E119, to evaluate the failure time or fire resistance.

During the heat transfer and moment–curvature analysis of the structural member, the temperature-dependent variation of the thermal, mechanical, and deformation properties of concrete, reinforcing steel, CFRP, and insulation were incorporated into the model. These high-temperature properties of concrete and steel were defined using the property relationships specified in Eurocode-2 [³⁶]. The thermal properties of FRP was neglected owing to the small cross-sectional area [³⁷]. Therefore, only the temperature-dependent strength properties of CFRP, according to Bisby et al. [³⁸], were considered in the model. Additionally, the FRP–concrete interfacial bond degradation was incorporated using the elevated temperature bond-slip relationships proposed by Dai et al. [³⁹]. When fire insulation was considered, only the temperature variation of thermal conductivity, specific heat, and density, as determined from thermal property tests, were considered in the model.

Analysis of tested beams and slabs

The aforementioned model was applied to conduct a fire resistance analysis of tested beams and slabs with the same geometrical configuration, material properties, loading, and boundary conditions as in the tests. The cross-section of various segments along the length of the beams was discretized into 20 mm × 20 mm quadrilateral elements in concrete and 10 mm × 10 mm elements in the FRP and insulation. The entire cross-section of the slabs was discretized into 10 mm × 5 mm elements. The response parameters predicted from the analysis, namely temperature, deflection, and fire resistance, were compared with those measured in fire tests.

The predicted and measured temperatures are compared in Figs. 9(a) and 9(b) for TB1 and TB2, respectively, and in Figs. 10(a) and 10(b) for S1 and S2, respectively. The temperatures at the FRP–concrete interface and in the bottom steel rebars of the beams and slabs were compared. These figures show that the predicted and measured temperatures compared well with minor discrepancies. For instance, the concrete moisture evaporation plateau observed in the temperature rise measured at steel rebars was not predicted in the model because the concrete moisture evaporation effect was not considered in the model. Further, the model slightly overpredicted the temperature in the steel rebar and at the FRP–concrete interface in the beams and slabs. However, the predicted temperature rise change in the steel rebars of the strengthened beams and slabs did not exceed 400°C until the end of fire exposure. Hence, these slightly higher temperatures did not affect the strength and stiffness degradation of the rebars, and therefore, did not affect the overall stiffness of the beam.

Figure 11 compares the predicted and measured deflections of the beams and slabs. The predicted deflection compared well with the measured values for the entire duration of the fire exposure. The predicted values were slightly higher than the measured values in the later stages of fire exposure. This was because of the higher temperature rise predicted by the model, which increased the degradation in the strength and stiffness of the rebars at a relatively higher rate than that in the tests. Overall, the trends of the predicted deflection values for both the beams and slabs were in close agreement with the trends measured in the test. Thus, the model can predict the fire response of FRP-strengthened flexural members with sufficient accuracy.

To further evaluate the structural response of the beams and slabs, the strength contributions from the CFRP and steel rebar at the mid-span section of the beams and slabs are plotted against time of fire exposure as shown in Figs. 12(a) and 12(b), respectively. The strength contributions from the CFRP and steel rebar were normalized with respect to the room-temperature ultimate tensile strength of the CFRP and yield strength of steel rebars, respectively. Additionally, the moment capacity degradation with time, as predicted by the model, is plotted in Figs. 13(a) and 13(b) for the beams and slabs, respectively.

Figure 12(a) shows that in the first few minutes of fire exposure, the CFRP and steel rebar contributed their respective maximum strength toward the capacity of the beams. The contribution of strength from the CFRP toward the capacity of beams began decreasing after approximately 25 min and 65 min of fire exposure in TB1 and TB2, respectively. This was attributed to temperature-induced bond degradation (as the interface temperature exceeded the T_g of epoxy), which reduced the composite interaction between the CFRP and concrete; consequently, the strength contribution from the CFRP toward the beam capacity decreased. After approximately 105 and 150 min of fire exposure, the strength contribution from CFRP was reduced to 0 in TB1 and TB2, respectively. This was attributed to the interface temperature in both beams being greater than 200°C (cf., Fig. 9) and the CFRP–concrete interaction being completely lost. Therefore, the strength contribution from CFRP was completely lost. This indicates that the contribution of the CFRP cannot be completely neglected when the T_g of epoxy was attained. Furthermore, Fig. 12(a) shows that the strength contribution from the steel rebars remained intact throughout the fire exposure owing to the lower temperature change in the rebars.

For S1 and S2, the trends of the strength contribution from CFRP were similar to those of beams, i.e., complete contribution in the initial stages, followed by a rapid decrease owing to bond degradation and then reduced to zero owing to complete loss of the CFRP–concrete bond. However, for the slabs, the strength contribution from the steel rebar did not remain intact until the end of fire exposure, as was observed for the beams. The strength contribution from the steel rebars began decreasing after approximately 85 and 60 min of fire exposure in S1 and S2, respectively.

Figure 13(a) shows that the flexural capacity of the beams decreased gradually in the initial stages of fire exposure. This was because of the presence of fire insulation, which reduced the temperature rise therefore, the deterioration in the strength of CFRP and steel rebars was almost negligible. Thus, the CFRP and steel rebar contributed their respective maximum strength toward the capacity of the beams (cf., Fig. 12(a)). The capacity of TB1 and TB2 began to decrease at a rapid pace after approximately 25 and 65 min of fire exposure, respectively. This rapid decrease in the capacity was due to bond degradation, which decreased the strength contribution from the CFRP (Fig. 12(a)). The capacity of TB1 and TB2 began to degrade at a significantly low rate after approximately 105 and 150 min of fire exposure, respectively. This is attributed to the fact that the CFRP–concrete bond interaction in both beams is completely lost and the beams behave as an insulated RC beam, wherein the decrease in strength in the steel rebar was negligible (Fig. 12(a)).

The degradation in the moment capacity of S1 and S2, as shown in Fig. 13(b), followed trends similar to those of the beams. The capacity of the slabs decreased rapidly during the fire exposure between 45 and 90 min for S1 and between 35 and 60 min for S2. Thereafter, the capacity of slabs decreased at a lower rate, as the CFRP–concrete composite interaction was completely lost (cf., Fig. 12(b)), and the slabs behaved as insulated RC slabs. The capacity of the slabs continued to decrease gradually but did not fall below the applied loading until the end of fire exposure. This was attributed to the minimal loss of strength in the steel rebars (cf., Fig. 12(b)), which was due to the slower temperature rise in the cross-section because of the thermal protection provided by the insulation.

Although the insulation layer applied on S2 was thicker than that applied on S1, the capacity of S2 degraded at a faster rate than that of S1. This was attributed to the higher load applied on S2, which caused higher internal stresses in the cross-section and higher shear stresses at the CFRP–concrete interface. The higher internal stresses increased the rate of degradation in the strength and stiffness of the constitutive material, whereas the higher shear stresses may have resulted in the earlier debonding of the CFRP from the concrete surface, thereby reducing the strength and stiffness of the slab. Moreover, a higher load level produces a large curvature in the slab, which increases the demand on the slab and decreases the reserve capacity. Thus, the load level on the strengthened members can significantly affect the rate of degradation in capacity and can reduce the fire resistance of a member. Overall, it can be concluded that the model can predict the thermal and structural behavior of strengthened concrete members exposed to fire with reasonable accuracy and can be utilized to predict the fire resistance of strengthened flexural concrete members.

Case study

The response of a strengthened concrete flexural member under fire conditions is governed by several factors such as the fire exposure scenario, applied load level, high temperature properties of insulation, CFRP, concrete, and steel rebars, as well as the configuration and thickness of the fire insulation [⁴⁰]. These parameters must be duly considered when evaluating the fire resistance of a CFRP-strengthened concrete member. A case study implementing the model for rational fire design of strengthened concrete flexural members is presented in this section, wherein the fire response of insulated and uninsulated CFRP-strengthened RC T-beams under two different fire scenarios is evaluated.

Four FRP-strengthened RC T-beams, namely, TB-i0-SF, TB-i12-SF, TB-i0-LF, and TB-i12-LF, were analyzed as part of this case study. Table 2 shows the details of the analysis parameters and loading for each of these beams. The cross-sectional geometry and material properties of these beams were identical to that of beam TB2; however, TB-i0-SF and TB-i0-LF were uninsulated, whereas TB-i12-SF and TB-i12-LF were protected with 12.5-mm thick fire insulation. Two different design fire scenarios, designated as short severe fire scenario (SF) and long severe fire scenario (LF), were considered in the analysis (Fig. 14). TB-i0-SF and TB-i12-SF were subjected to an SF fire scenario, whereas TB-i0-LF and TB-i12-LF were subjected to the LF fire scenario. The SF and LF scenarios comprised a heating phase that lasted for 90 and 120 min, respectively, followed by a cooling phase wherein the temperature decreased linearly by 10°C every minute until ambient temperature conditions were attained (Fig. 14). These design fires were representative of a severe fire in a parking garage or chemical plant. All four beams were analyzed under the same loading and boundary conditions as described in the tests; however, the structural load corresponded to 60% of the ambient temperature strengthened capacity.

The thermal and structural responses of the beams, as predicted by the model, are shown in Fig. 15, while the fire resistance time as well as the moment capacity and deflection of the beams at failure are described in Table 3. The progression of temperature with fire exposure time in the corner rebar of the beams is shown in Fig. 15(a) to evaluate the thermal response. The figure shows that the temperature in the corner rebar of the uninsulated beams (TB-i0-SF and TB-i0-LF) increased more rapidly than that of the insulated beams (TB-i12-SF and TB-i12-LF). However, the rebar temperatures increased gradually in all the beams compared with the fire temperatures, which increased very rapidly in the beginning of fire exposure. The rebar temperatures increased with a lag compared with the fire temperatures. For instance, the peak temperature was attained after 150 and 180 min in TB-i0-SF and TB-i0-LF, respectively, and after 180 and 210 min of fire exposure in TB-i12-SF and TB-i12-LF, respectively. The peak rebar temperature reached 631°C and 658°C in TB-i0-SF and TB-i0-LF, respectively, and 438°C and 466°C in TB-i12-SF and TB-i12-LF, respectively. Thus, the rebar temperatures in uninsulated beams increased beyond 593°C, which is a critical temperature for reinforcing steel bars as the rebar strength decreases to less than 50% of its ambient temperature strength.

Figure 15(b) shows the predicted decrease in the moment capacity of the analyzed beams throughout the fire exposure. The figure shows that the moment capacity of TB-i0-SF and TB-i0-LF decreased dramatically within a few minutes of the initiation of fire. This rapid degradation in the capacity was caused by the temperature-induced bond degradation resulting from the rapid increase in temperature at the CFRP–concrete interface. After approximately 25 min, the capacity of the uninsulated beams decreased gradually and remained almost constant until 70 min of fire exposure. The capacity of the beams began to decrease again at a relatively faster rate after 70 min owing to the increase in rebar temperature beyond 400°C. The capacities of TB-i0-SF and TB-i0-LF continued to decrease at a constant rate and decreased below the bending moment resulting from applied loads after 100 and 115 min, respectively, indicating strength failure in the beam. In contrast, the capacities of the insulated beams TB-i12-SF and TB-i12-LF decreased gradually until 90 min of fire exposure and then remained constant until 150 min. Thereafter, TB-i12-SF and TB-i12-LF began gaining capacity owing to the decrease in the rebar temperatures. Therefore, the capacity of the insulated beam remained greater than the moment due to applied loading indicating no strength failure in the beam.

Figure 15(c) shows the progression of deflection in the analyzed beams throughout the fire exposure. The uninsulated beams TB-i0-SF and TB-i0-LF began deflecting rapidly from the beginning of fire exposure. However, the deflection and rate of deflection limit were exceeded only in TB-i0-LF, indicating failure through the breaching deflection limit state. For TB-i12-SF and TB-i12-LF, the deflections increased at a similar and relatively lower rate until 150 and 180 min, respectively. The deflection in both the beams began to decrease after the aforementioned time owing to the decrease in the rebar temperatures. The maximum deflection in the insulated beams was less than 40 mm, which is significantly below the deflection limit state, indicating no failure.

The failure time of the beams was evaluated by applying the strength and deflection limiting criteria as described earlier, which was considered to be the fire resistance time. The above discussion indicates that TB-i0-SF attained the strength limit state after 100 min of fire exposure, and TB-i0-LF attained the strength and deflection limit state after 115 min of fire exposure. However, both TB-i12-SF and TB-i12-LF did not attain the strength or deflection limit state throughout the fire exposure. Therefore, TB-i0-SF and TB-i0-LF had 100 and 115 min of fire resistance, respectively, whereas the insulated beams had 4-h fire resistance under the selected design fire scenarios. Thus, uninsulated CFRP-strengthened RC beams cannot withstand the selected design fire scenarios under sustained structural loads (equal to 60% of ambient temperature strengthened capacity) for more than 2 h. However, to achieve more than 2 h of fire resistance, at least a 12.5 mm-thick fire insulation is required.

This case study demonstrated that the model can be efficiently utilized to conduct a rational design of strengthened concrete beams under fire conditions. Although the analysis was illustrated for a CFRP-strengthened beam, the model can be utilized to examine the effect of different parameters on the fire response of CFRP-strengthened concrete slabs. Additionally, the level of fire insulation thickness required on a CFRP-strengthened flexural member can be optimized using numerical model to achieve satisfactory fire resistance.

Conclusions

The following are the key findings from the research presented in this paper.

1) RC beams strengthened with CFRP and supplemented with 19 and 32 mm-thick spray applied fireproofing at the bottom and sides can withstand service load levels for four hours under an ASTM E119 standard fire exposure.

2) CFRP-strengthened RC slabs protected with 19 and 25 mm-thick fire insulation can withstand three hours of ASTM E119 standard fire exposure under a constant structural loading equivalent to service loads on the strengthened slabs.

3) The CFRP–concrete interfacial bond degrades gradually; therefore, the CFRP–concrete composite action is lost completely at a temperature much higher than the T_g of the polymer. Thus, neglecting the contribution of the CFRP to the strength and stiffness of a member under fire conditions yields a conservative estimation of the fire resistance of the strengthened member.

4) The numerical model discussed in this paper can evaluate the fire performance of CFRP-strengthened concrete flexural members with reasonable accuracy and can be utilized to optimize various design parameters affecting fire resistance, including the thickness of the fire insulation.

References

[1] ACI 440.2R–17. Guide for the Design and Construction of Externally Bonded FRP Systems for Strengthening Concrete Structures. Farmington Hills: American Concrete Institute, 2017

[2] ACI 216.1–14. Code Requirements for Determining Fire Resistance of Concrete and Masonry Construction Assemblies. Farmington Hills: American Concrete Institute, 2014

[3] IBC. International Building Code. Country Club Hills: International Code Council, 2015

[4] Naus D J. The effect of elevated temperature on concrete materials and structures—A literature review. Washington, D.C.: Report of Oak Ridge National Laboratory, U.S. Nuclear Regulatory Commission Office of Nuclear Regulatory Research, 2006

[5] Kodur V K R, Dwaikat M M S, Dwaikat M B. High-temperature properties of concrete for fire resistance modeling of structures. ACI Materials Journal, 2008, 105(5): 517–527

[6] Phan L, Carino N. Review of mechanical properties of HSC at elevated temperature. Journal of Materials in Civil Engineering, 1998, 10(1): 58–65

[7] Kodur V K R. Spalling in high strength concrete exposed to fire—Concerns, causes, critical parameters and cures. In: Proceedings of the ASCE Structures Congress, Advanced Technology in Structural Engineering. Philadelphia: ASCE, 2000, 1–9

[8] Ali F A, O’Connor D, Abu-Tair A. Explosive spalling of high strength concrete columns in fire. Magazine of Concrete Research, 2001, 53(3): 197–204

[9] Kodur V R, Sultan M A. Structural behaviour of high strength concrete columns exposed to fire. In: Proceedings of International Symposium on High Performance and Reactive Powder Concrete. Sherbrooke: NRCC1998, 217–232

[10] Kodur V, Cheng F, Wang T, Sultan M. Effect of strength and fiber reinforcement on fire resistance of high-strength concrete columns. Journal of Structural Engineering, 2003, 129(2): 253–259

[11] Ali F, Nadjai A, Silcock G, Abu-Tair A. Outcomes of a major research on fire resistance of concrete columns. Fire Safety Journal, 2004, 39(6): 433–445

[12] Bilodeau A, Kodur V R, Hoff G C. Optimization of the type and amount of polypropylene fibers for preventing the spalling of lightweight concrete subjected to hydrocarbon fire. Cement and Concrete Composites, 2004, 26(2): 163–174

[13] Ali F, Nadjai A. Fire resistance of concrete columns containing polypropylene & steel fibers. ACI Special Publications, 2008, 255: 199–216

[14] Khaliq W, Kodur V. High temperature properties of fiber reinforced high strength concrete. ACI Special Publications, 2011, 279: 1–42

[15] Bhatt P P, Kodur V K R, Matsagar V. Numerical approach to evaluate elevated temperature behavior of steel fiber reinforced concrete columns. Indian Concrete Journal, 2019, 93(8): 8–15

[16] Deuring M. Fire Tests on Strengthened Reinforced Concrete Beams. Research Report No. 148’795. Dubendorf: Swiss Federal Laboratories for Materials Testing and Research, 1994

[17] Blontrock H, Taerwe L, Vandevelde P. Fire tests on concrete beams strengthened with fiber composite laminates. In: Konrad B, eds. Proceedings of the 3rd Ph.D. Symposium in Civil Engineering. Vienna: Institute of Structural Engineering, University of Agricultural Sciences, 2000, 151–161

[18] Williams B, Kodur V K R, Green M F, Bisby L A. Fire endurance of fiber-reinforced polymer strengthened concrete T-beams. ACI Structural Journal, 2008, 105(1): 60–67

[19] Gao W Y, Hu K X, Lu Z D. Fire resistance experiments of insulated CFRP strengthened reinforced concrete beams. Chinese Civil Engineering Journal, 2010, 43(3): 15–23 (in Chinese)

[20] Ahmed A, Kodur V. The experimental behavior of FRP-strengthened RC beams subjected to design fire exposure. Engineering Structures, 2011, 33(7): 2201–2211

[21] Firmo J P, Correia J R, França P. Fire behaviour of reinforced concrete beams strengthened with CFRP laminates: Protection systems with insulation of the anchorage zones. Composites. Part B, Engineering, 2012, 43(3): 1545–1556

[22] Yu B, Kodur V. Fire behavior of concrete T-beams strengthened with near-surface mounted FRP reinforcement. Engineering Structures, 2014, 80: 350–361

[23] Firmo J P, Correia J R. Fire behaviour of thermally insulated RC beams strengthened with EBR-CFRP strips: Experimental study. Composite Structures, 2015, 122: 144–154

[24] Dong K, Hu K, Gao W. Fire behavior of full-scale CFRP-strengthened RC beams protected with different insulation systems. Journal of Asian Architecture and Building Engineering, 2016, 15(3): 581–588

[25] Carlos T B, Rodrigues J P C, de Lima R C A, Dhima D. Experimental analysis on flexural behaviour of RC beams strengthened with CFRP laminates and under fire conditions. Composite Structures, 2018, 189: 516–528

[26] Blontrock H, Taerwe L, Vandevelde P. Fire testing of concrete slabs strengthened with fiber composite laminates. In: Burgoyne C J, eds. In: Proceedings of the 5th International Conference on Fiber-Reinforced Plastics for Reinforced Concrete Structures FRPRCS-5. Cambridge: Thomas Telford Publishing, 2001, 547–556

[27] Williams B, Bisby L, Kodur V, Green M, Chowdhury E. Fire insulation schemes for FRP-strengthened concrete slabs. Composites. Part A, Applied Science and Manufacturing, 2006, 37(8): 1151–1160

[28] López C, Firmo J P, Correia J R, Tiago C. Fire protection systems for reinforced concrete slabs strengthened with CFRP laminates. Construction & Building Materials, 2013, 47: 324–333

[29] ACI 318–14. Building Code Requirements for Reinforced Concrete and Commentary. Farmington Hills: American Concrete Institute, 2014

[30] ASTM D638. Standard Test Method for Tensile Properties of Plastics. West Conshohocken: American Society for Testing and Materials, 2014

[31] ASTM D4065. Standard Practice for Plastics: Dynamic Mechanical Properties: Determination and Report of Procedures. West Conshohocken: American Society for Testing and Materials, 2012

[32] ASTM E119. Standard Test Methods for Fire Tests of Building Construction and Materials. West Conshohocken: American Society for Testing and Materials, 2016

[33] Firmo J P, Arruda M R T, Correia J R. Contribution to the understanding of the mechanical behaviour of CFRP-strengthened RC beams subjected to fire: Experimental and numerical assessment. Composites. Part B, Engineering, 2014, 66: 15–24

[34] Kodur V K R, Ahmed A. Numerical model for tracing the response of FRP-strengthened RC beams exposed to fire. Journal of Composites for Construction, 2010, 14(6): 730–742

[35] Kodur V K R, Bhatt P P. A numerical approach for modeling response of fiber reinforced polymer strengthened concrete slabs exposed to fire. Composite Structures, 2018, 187: 226–240

[36] Eurocode-2 EN 1992–1-2. Design of Concrete Structures, Part 1–2: General Rules-Structural Fire Design. Brussels: European Committee for Standardization, 2004

[37] Kodur V K R, Bhatt P P, Naser M Z. High temperature properties of fiber reinforced polymers and fire insulation for fire resistance modeling of strengthened concrete structures. Composites. Part B, Engineering, 2019, 175: 107104

[38] Bisby L A, Green M F, Kodur V K R. Response to fire of concrete structures that incorporate FRP. Progress in Structural Engineering and Materials, 2005, 7(3): 136–149

[39] Dai J G, Gao W Y, Teng J G. Bond-slip model for FRP laminates externally bonded to concrete at elevated temperature. Journal of Composites for Construction, 2013, 17(2): 217–228

[40] Kodur V K R, Naser M Z. Structural Fire Engineering. New York: Mc Graw-Hill Professional, 2020

RIGHTS & PERMISSIONS

Higher Education Press