Introduction
One of the most critical drawbacks in currently used steel reinforcement in reinforced concrete (RC) structures is its tendency to accumulate plastic deformation under excessive loads. In the case of earthquake loadings, structural ductility and energy dissipation capability are two important features that are sought in modern earthquake-resistant structures. Engineers rely on the nonlinear hysteretic behavior of steel reinforcement to provide such features, which results in permanent damage to the steel reinforcement and to the structure after the earthquake subside. Excessive steel deformation (beyond yielding) in RC moment resisting frames (MRFs) for example often results in permanent residual drifts, which not only cause overall capacity degradation but also pose safety issues for the occupants even under gravity loads. Thus post-damage functionality of ductile RC structures reinforced with steel reinforcement has started to be questioned because of their non re-centering performance.
To deal with the issue of damage accumulation associated with ductility; this paper presents experimental investigation of a new type of shape memory alloy (SMA)-based composite reinforcement with ability to withstand large elongation without accumulation of permanent residual strains. Often referred to as smart materials, SMAs have shown highly nonlinear behavior in terms of shape memory effect, super-elasticity and martensite damping, which encouraged researchers to study their use in civil engineering applications (e.g., Ref [
1-
4]). This study investigated the use of shape memory alloy (SMA) fibers as reinforcing material in the composite which can further be employed as primary reinforcement for RC members and structures. Named as ‘shape memory alloy-FRP (SMA-FRP) composite’; the proposed reinforcement has the capability of re-centering and energy dissipation through hysteretic action. SMA-FRP reinforcing bar comprising of polymeric resin reinforced with small diameter superelastic SMA fibers (Fig. 1(a)) were manufactured and then tested under quasi-static loading to get constitutive material behavior.The nonlinear, yet pseudo-elastic flag shape hysteresis typically associated with superelastic SMAs (Fig. 1(b)) is a direct result of a reversible stress-induced phase transformation between austenite and martensite phases.
In this study, Nickel-Titanium (NiTi) alloy has been used which is the most commercially available alloy. This nonlinear, yet pseudo-elastic behavior of SMA fibers will allow SMA-FRP composite reinforcement to exhibit ductile and energy dissipating behavior with minimal damage to reinforced concrete structure. This is not likely in steel reinforcement which exhibit ductility through permanent damage to the steel and concrete which often yields significant residual (permanent) deformations in the structure. The proposed SMA-FRP composite reinforcement is sought as a mean to introduce the features of ductility and re-centering to RC members and structures. Early research involving SMA materials in composites was performed by Rogers and others at Virginia Tech in early 1990s [
5,
6]. Their research focused mainly on vibration control of a composite with embedded SMA wires. Xu et al. [
7] explored the use of small diameter SMA wires to improve the interfacial bonding of SMA within composite. Sterzla et al. [
8] investigated two way actuation of SMA composite to produce actuators. Other researchers [
9-
11] investigated the interaction of SMA ribbons, foils and particles with host materials of the composites and explored diversity in their applications. The idea of using SMA-FRP reinforcement in concrete structures was recently studied by Wierschem and Andrawes [
12], whose study was mainly numerical with the aim of exploring the damping and ductility characteristics of superelastic SMA-FRP reinforcement in RC members. Zafar and Andrawes [
13] further investigated analytically the seismic behavior of this new composite reinforcement on the structural level. They also explored the fabrication and manufacturing of the new SMA based composite rebars [
14]. However, none of the previously mentioned studies investigated the experimental flexural behavior of concrete beams reinforced with the new SMA-FRP composite.
This study presents the experimental testing of SMA-FRP composite rebars embedded in small scale concrete beam under quasi-static flexural loading. The beam was tested under 3 point bending to failure with displacement controlled cyclic regime, which allowed studying the hysteretic and re-centering property of proposed SMA-FRP composite.
The paper explores the uniaxial tensile-compressive behavior of the SMA-FRP composite coupon specimen to observe the effects of compression cycles on tensile behavior of composite. The paper also explores the microstructure of the composite after testing to visualize defects like de-bonding and delamination within composite materials. For this purpose scanning electron microscope (SEM) images are used as means to verify manufacturing and fabrication process.
SMA-FRP composite coupon specimens
NiTi (NiTi-51%-49%) SMA wires with a diameter of 500µm were selected for manufacturing of SMAFRP composites. SMA wires were first trained through continued cycling to stabilize properties such as forward and reverse transformation stress, residual strain, Young’s modulus and energy dissipation capacity. More information about the SMA training process can be found in Zafar and Andrawes [
14]. Next step was exploration of host resin material for the composite. Most of the commercially available resins used in manufacturing conventional FRPs exhibit high strength and stiffness properties because of their application in rehabilitation and reinforcement of civil structures. To fully explore the elongation potential of NiTi SMA (6-8% strain), the selected resin matrix in the composite needs to have high elongation properties, while exhibiting stable behavior over course of cyclic tensile stretching. After conducting detailed testing [
14], the host material selected for use in the manufacturing of thermoset resin was Bisphenol/epichlorohydrin derived liquid epoxy containing a commercial grade of n-butyl glycidyl ether (Epoxy-862). Curing agent FB-84 Polyamine (Epikure-3274) was used as hardeners. Both epoxy and curing agents were mixed with a ratio of 100:40 pbw and cured for 48 h at 35°C in order to get high elongation resin matrix with stable behavior over course of cyclic tensile stretching. After undergoing resin manufacturing, trained SMA wires and supplementary glass fibers were embedded in this resin matrix to fabricate SMA-FRP composite under specified temperature and pressure, using hot-press. From here onwards, host matrix for all the composites discussed in this study is combination of Epoxy-862 and Epikure-3274 as curing agent in 100:40 pbw ratio, respectively.
Uniaxial tension composite specimens
Two composite specimens, one reinforced with only SMA wires (Fully reinforced composite, FRC), and second reinforced with SMA wires and additional glass fibers (Partially reinforced composite, PRC) were fabricated. Both composite specimens were designed to achieve target initial composite modulus (
Ec) of 13.74 Gpa, which required FRC and PRC specimens to be reinforced with 7 and 3 SMA wires, respectively. The reinforcement volumetric fraction including SMA wires and additional glass fibers for FRC and PRC specimens was 20.3% and 17.7%, respectively. Table 1 shows specifications of both SMA-FRP composite specimens. Figure 2 shows the uni-axial tensile testing results of the two SMA-FRP composite. Details regarding manufacturing process of the SMA-FRP composite specimens have been presented in Zafar and Andrawes [
14].
Results show that the PRC specimen while exhibiting reasonable elongation and hysteretic properties also show higher stiffness and strength behavior in comparison with FRC specimens, while the SMA wires undergo austenite to marten site phase transformation (post-yield behavior). Both stiffness and strength of hybrid SMA-FRP composite specimen (PRC) were improved compared to that of the 100% SMA-FRP (FRC), while the FRC composite specimen exhibited better ductility and energy dissipation capability. This shows that both types of composites have their own advantages; hence could be used for different applications.
Uniaxial tension-compression composite specimens
There has been limited research on compression behavior of NiTi SMA material including the studies by Liu et al. (1998) and Lim and McDowell (1999). However no study exists which addresses the behavior of SMA-FRP composite under compression or under complete cyclic loading. Therefore, in order to make the study conclusive and complete, pure SMA composite with 13 SMA wires (FRC-13) was manufactured and tested under complete cyclic behavior (tension and compression). Results from the tests helped in assessing the effects of compressive behavior of composite on its tensile behavior. Same testing loading frame was used which has been used for tensile loading tests. Since the composite specimens will be undergoing buckling while under compressive loading, a loading protocol was developed to limit the compressive strains to around 1/10th of tensile straining in each cycle. Although, no de-bonding was observed in previously tensile tested composite specimens, addition of compressive loading will definitely make the slippage of SMA wires with its surrounding resin matrix more prominent. To study this effect, the specimen was subjected to six tensile and compressive loading cycles. In the first cycle, the FRC-13 specimen was loaded in tension up to strain of 1% and there after it was incremented by 0.5% in every subsequent cycle. Cyclic stress strain curve of FRC-13 specimens has been shown in Fig. 3. As expected, the composite specimen was able to dissipate energy through hysteretic action of SMA and was able to recover residual strains due to superelasticity of SMA.
Due to addition of compressive loading and straining, de-bonding of SMA wire with resin matrix becomes an issue, as stated earlier. During all the compressive loading cycles, FRC-13 composite specimens experienced buckling, as expected which affected its tensile behavior. During the tensile unloading in 5th cycle, after experiencing reverse phase transformation by SMA, the wires started to de-bond with its surrounding resin matrix because of buckling effects. This was evident from the jagged lines in 5th and 6th cycle of the test, as seen in the Fig. 3. Each drop in force / stress indicates de-bonding which continued once the specimen underwent compressive strains. By the end of the loading of specimen in compression in 6th cycle, the composite specimen had experienced complete de-bonding, thus resulting in crushing of resin matrix in compression as seen in Fig. 3 with a jump to large strain value. Results show that the compressive loading does have impact on the tensile properties of the composite beyond strains of 5%. However in real application, the SMA-FRP composite rebar will be surrounded with concrete which will provide bracing along the length of the rebar. This bracing / confinement of rebar by concrete will prevent it from exhibiting buckling tendency, thus limiting de-bonding effects.
SEM imaging of SMA-FRP composite
Scanning electron microscope (SEM) is a powerful magnification tool that utilizes focused beams of accelerated electrons to generate a variety of signals at the surface of solid specimens. These signals include secondary electrons (SE) and backscattered electrons (BSE) which can be utilized to generate high-resolution images that provide topographical and compositional information. Secondary electrons (SE) are most valuable for showing topography on the sample surface and backscattered electrons (BSE) are most valuable for illustrating contrasts in composition of the composite specimen. For this study, SEM images were utilized to investigate the damages and anomalies like de-bonding / de-lamination between resin and SMA, fracture of FRP fibers, efficacy of resin in filling all air voids and overall layout of composite specimen. These images acted as tools to confer the manufacturing technique which was established after many trials.For this purpose, all composite specimens used for SEM imaging were acquired from tested / damaged specimens and underwent surface polishing.
Figure 3 shows SEM images of PRC composite specimen with both SE and BSE image at a 200X magnification. The image helps in understanding the layout of the specimen with different constituents. As mentioned earlier, the images were observed after the specimen underwent tensile testing and show broken individual glass fibers spread around the SMA wires in resin matrix which is seen in the Fig. 4 with darker tone. Image also shows complete enveloping of resin around tiny glass fibers and SMA wire, proving good penetration capability and cavity filling of the selected resin matrix system. Figure 4(a) shows the contour of the PRC specimen around SMA wires and glass fibers. Figure 4(b) shows contrast image of the PRC sample, in which SMA wires are shown in white while resin which is opaque, is revealed in dark tone. Glass fibers can also be distinguished clearly in BSE image.
Figure 5(a) shows a back-scatter electron (BSE) image of PRC composite specimen with 200X magnification and a blow up with 1200X and 5000X magnification in Figs. 5(b) and (b), respectively. BSE images allow achieving contrasting images of the cross section which helps in identifying material with different densities. Figure 5(d) shows secondary electron (SE) image at 5000X magnification to see the depth and contours of the cross section.
Figure 6 highlights the penetration and surrounding of SMA wire and individual glass fibers by resin in PRC specimen. Image is able to show all three constituents of the hybrid composite specimen i.e., SMA, glass fibers and resin. Figure 6 which is BSE image with 25000X magnification, also show absence of voids and de-lamination between SMA, glass fiber and resin matrix, depicting good bond.
Concrete beam details
The research work presented in the previous sections of this paper has helped in proving the concept of using SMA-FRP composite as reinforcement with or without supplementary conventional fibers at a material level. However to fully explore the efficacy of the proposed composite reinforcement, fabrication process needs to be extended beyond the realm of dog bone coupon specimen to actual rebar. These rebars are the used as primary reinforcement in small scale concrete beam to explore their flexural behavior under 3-point bending test. Concrete beam was first designed using preliminary numerical models developed using OpenSees software [
15] which has been specifically designed for nonlinear fiber based analysis and earthquake simulations. Because of limitation of hot-press which restricted the length of SMA-FRP composite rebars, the concrete beam was also required to be scaled down accordingly. So length of rebar and thus the beam was considered as one of the design constraints during the numerical modeling. Finally, a T-beam with 250 mm in length, 51 mm in depth with 64 and 25 mm in width for flange and web, respectively, was established as dimensions. Because of small scale of reinforcement, the concrete beam was designed without lateral reinforcement (stirrups). The longitudinal reinforcement ratio was kept as 1.25% as a design parameter, which allowed calculation of beam cross sectional area required to achieve the target reinforcement ratio. The main aim of test was to investigate the ductility and re-centering capability of SMA composite reinforced concrete beam, which requires SMA composite rebar to sustain high straining. Since no stirrups were provided, the concrete was left un-confined. To delay the crushing of concrete at top of the beam, T-beam cross section was selected for design purpose. This will allow increase in the curvature capacity of the beam’s section, hence force SMA composite rebars to experience high strains during testing. The T-beam cross sectional dimensions were optimized to allow embedment of the two manufactured SMA-FRP composite rebars as tension reinforcement with 6 mm cover from bottom. Schematics of the designed T-beam with dimensions of the cross section are shown in Fig. 7.
Figure 7 shows projection of composite rebars beyond the face of the beam. This projection on both ends is intended for anchoring the rebars at both ends using U-clamps to restrict slippage of the rebars along the beam length during testing. Since the manufactured rebars have smooth surface finish, in the likelihood of them de-bonding along the interface with concrete, the proposed U-clamp attachments will be beneficial in restraining this slippage. To achieve the desired reinforced concrete T-beam, wooden mold were prepared for casting of the desired cross section. A wooded mold with inner length, width and depth as shown in Fig. 7 was constructed with opening at each ends for the two projecting composite rebars. Manufactured wooden mold for T-beam with embedded SMA composite rebars is shown in Fig. 9. Ordinary Portland cement (OPC) was used in casting the beam. Because of miniature nature of the SMA composite reinforcement and beam cross section, coarse aggregate were not used. Only fine aggregate (sand) was used to produce grout with adopted mix proportion shown in Table 2. The concrete mix was designed to achieve strength of 42 MPa or more. Achievement of higher grout strength would help in further delaying crushing of concrete in compression at top of the beam. Plasticizer admixture was used to increase the workability and flow characteristics of the grout in and around the SMA composite reinforcement.
American standard for testing materials (ASTM) C305 (ASTM C305, 2012) [
16] was followed for mixing the contents of the grout mix. Figure 8 shows the casting of T-beams with grout. Pouring of grout in the wooden mold was done on a vibrating table to allow better flow of grout between and around the rebars. The vibration was restricted to 5 s to avoid segregation and bleeding in the grout.
Same grout mix was used to prepare six 50 mm cubes which were tested after 14 days of curing to obtain compressive strength of the beams. These cubes were manufactured and tested in accordance with ASTM C109/ C109M (ASTM C109/C109M, 2013) [
17]. Average compressive strength achieved from the six cubes was 55 MPa, which exceeded the target strength. This high strength grout will allow SMA composite rebar to sustain higher strains before failure of concrete in compression.
Manufacturing of square SMA-FRP composite rebar
Square SMA-FRP composite rebar were manufactured for use as longitudinal reinforcement in small scale concrete T-beam. The length of SMA-FRP composite rebar had to be restricted to 280 mm due to the limitation of the dimensions of hot-press. This length of rebar in turn affects the length of concrete beam which will be reinforced with the manufactured SMA-FRP composite rebar. Keeping this in mind, a cutout of 280 mm in length, 3.2 mm in width was made in the silicone mold which was 3.2 mm in thickness. This cutout in silicone was used as dam for assembling the SMA and resin matrix layer by layer to manufacture square SMA-FRP rebar. All the SMA wires used in the manufacturing process were 280 mm in length and underwent training to stabilize their mechanical properties. 22 trained SMA wires of 500 μm in diameter were embedded in resin matrix to fabricate one composite rebar. In total, two SMA composite rebars were manufactured for reinforcing concrete beam Fig. 9 shows the layout of the silicone mold dam with cutouts, ready for manufacturing process.
The 22 trained SMA wires were distributed in four layers along the height of the cutout in silicone dam. All the layers were separated by segments of porous fabric cloth along the length of the composite rebar (L). The function of these porous fabric segments was to bifurcate each layer of SMA wires while at the same time, not hinder the flow of resin into the gaps between SMA wires. Figure 10 shows schematics of the layout of the SMA-FRP composite rebar.
Resin matrix was degassed for 30 min before pouring into the molds, layer by layer. Once the laying of all 22 wires in each cutout was complete, the whole mold setup was covered by permeable nylon and bleeder cloth to capture any over flowing resin. The mold setup was then subjected to 172.3 kPa pressure and 35°C curing temperature in the hot-press for a period of 24 h. After the curing period, the SMA composite rebars were extracted from the mold and were ready for placing in concrete beam. The first SMA composite rebar (SC-1) with 22 SMA wires and composite area of 10.24 mm2 allowed achievement of 39.3% FVF. SC-2 achieved 40.6% FVF. The final area and FVF which was achieved in the SMA composite for each rebar is shown in Table 3.
Instrumentation and test setup
The small scale beam was investigated in flexural under three point bending test in a simply supported configuration. A notch was provided at the mid span on the bottom of the beam starting just below the reinforcement. Provision of notch will facilitate initiation of flexural crack at the specified location. Since maximum moment in simply supported beam occurs at mid span, so the notch was created so as to facilitate crack at pre-determined location. The beam was rested on a roller and a pin support at each end through support brackets, while the load will be applied in the middle of the beam through load cell. Linear variable displacement transducer (LVDT) was used to measure the vertical deflection of the beam under the load. Location of various gauges used for measurements during the testing has been shown in Fig. 11.
Three strain gauges were used along the height of beam at mid span to measure strains in concrete. SG-1was provided right above the notch. Reading from this strain gauge will give indication of crack appearance and progression. SG-2 was provided at the top of the web, right under the flange. Reading from this strain gauge will give indication of progression of crack and shifting of neutral axis. SG-3 was provided at the T-beam flange. Reading from this strain gauge will give indication of crushing of concrete in compression. Crack mouth opening device (CMOD) was used to measure the opening of the provided notch after initiation of the crack. CMOD reading will also help in determining of crack closing thus indicating re-centering capability of SMA composite rebar. All the measuring instruments were connected to data acquisition system (DAQ) for recording of readings during the test.
The 3-point bending test on RC T-beam reinforced with SMA composite reinforcement was controlled using displacement reading from actuator. Constant actuator displacement rate of 0.25 mm/min was used throughout the test. For loading segment of the cycle, limit detection was placed in loading protocol using reading from LVDT. This limit detection was user selected and allowed the loading segment of the cycle to stop. After this, the unloading cycle begins which is again controlled by displacement reading from actuator. However again limit detection from load cell (zero force) was placed to stop the unloading segment of the cycle. This procedure was repeated for all subsequent loading and unloading cycles. For each cycle the target displacement from LVDT was set at an absolute value of 1.15 mm. This absolute value was chosen in order to have an initial cycle before the phase transformation in SMA. The final test setup along with all the instrumentation used during the testing has been shown in Fig. 12.
Testing and results
The test results from load cell, LVDT, CMOD and strain gauges were collected and were plotted to assess the behavior of the beam reinforced with composite. The results were able to offer insight about flexural behavior of proposed composite in addition to bondage with concrete. Plots were developed between LVDT vertical deflection at midspan, CMOD horizontal deflection and readings from strain gauges in relation to force to facilitate the understanding of the behavior of beam reinforced with proposed composite reinforcement.
Load-deflection and crack opening response
To facilitate examining and understanding the load-deflection curves of the beam, the response curve was segregated in terms of cycles and points. In total, 4 downward deflection cycles were conducted which included loading of the beam to target deflection and then unloading till the force is zero. All four cycles and the points (A to P) during the test are shown in force-deflection plot in Fig. 13. As the externally applied load increased beyond the initial cracking load of the beam, the crack depth and crack mouth opening displacement (CMOD) also increases. Figure 14 shows the CMOD readings of the tested beam in relation to loading. Since CMOD significantly influence the durability of structures and may accelerate corrosion of internal longitudinal steel reinforcement. Hence ability of reinforcement to close the cracks initiated due to nonlinear loading is also an important aspect for the sustenance of the structures.
At point-A, the beam initially was in an unloaded position as shown in Fig. 12, from where the load started to increase till cracking load. At point-B, the flexural crack initiated from provided notch at cracking load of 1.31 kN with deflection of 0.29 mm. From point-A till point-B, the CMOD showed almost negligible increase in horizontal displacement because of notch being intact. After point-B (initiation of flexural crack), the CMOD recorded jump in horizontal displacement because of initiation of crack. At point-C, web shear crack initiated on the right side of the beam at 2.63 kN load with deflection of 0.93 mm at mid-span. Figure 13a shows the initiation of flexural and shear crack at point-B and point-C, respectively.
It is known fact that concrete shear strength is the combination of resistance from compressive zone, aggregate interlocking, dowel action and shear reinforcement. This premature appearance of shear crack could primarily be related to absence of shear reinforcement and aggregate interlocking (because of use of grout instead of regular concrete). Point-D is the completion of loading phase of the 1st cycle, as shown in force-deflection plot in Fig. 14(a). The peak load and deflection at this point were 3.05 kN and 1.28 mm, respectively. Hereafter the load was reduced at control rate till the force in the load cell was zero (Point-E). At point-E, the beam accumulated 0.57 mm residual deflection (should be close to zero because SMA should still be in the Austenite phase, elastic range) which suggest slight slippage of smooth SMA composite rebars only in the mid-span region. This slippage would also trigger nullifying of dowel action which would add to the explanation provided earlier for development of premature shear cracks.
Point-F is the load where forward transformation in SMA (yielding), i.e., start of phase transformation from Austenite to Martensite is observed. This happened at 3.69 kN load with corresponding deflection of 1.96 mm at mid-span. Hereafter the test was continued till the achievement of target deflection of 2.33 mm (point-G). At this point, maximum flexural and shear crack opening was observed for the 2nd cycle. Point-H marks the completion of 2nd cycle. At this point the observed residual crack opening at mid-span was 1.01 mm. Even after the SMA wires experienced phase transformation within composite rebar, because of super-elastic property, SMA composite rebar was able to recover much of the residual deflection at mid-span. Although results suggest that SMA composite rebar experienced some slippage along interface with concrete, recovery of deflection by 75% shows re-centering capability. Point-I again marks the forward transformation in SMA in the 3rd cycle. This happened at 3.55 kN load with corresponding deflection of 1.99 mm at mid-span. Hereafter the loading continued till reaching of point-J which is the peak load point for the test with 3.73 kN and corresponding deflection of 2.2 mm. Point-K is the maximum deflection of the beam in the 3rd cycle. The peak deflection measured at this point was 3.37 mm with corresponding load of 3.62 kN. Figure 15(b) shows beam configuration at this point with maximum flexural crack opening at mid-span.
Point-L and M mark the Martensite to Austenite start and Austenite to Martensite finish phase transformation in SMA wires. Because of super-elastic property associated with SMA wires, SMA composite rebar is able to recover residual strains and displacements. Point-N, which is the completion of 3rd cycle, shows excellent re-centering of T-beam due to presence of SMA composite rebars. This crack closing and recovery would have not been possible in steel reinforced beam because of permanent damage and plasticity in steel material. Better engagement of SMA composite rebar in 3rd indicates drop in slippage at the interface between SMA composite rebar and concrete. This could be either because of engagement of SMA composite rebar with concrete due to presence of unintentional deformations along the length of rebar (From manufacturing process) or restraining of composite rebar by U-clamps at both ends. At this point the observed residual crack opening at mid-span was 1.21 mm as shown in Fig. 14(c). This is recovery of 92% of deflection, shows unique ability of re-centering of SMA composite rebar. At the end of 3rd cycle the beam experience addition of only 0.2 mm residual deflection. Picture of beam at point-N is shown in Fig. 15(c) with excellent crack closing and deflection recovery ability. At point-O shown in Fig. 13(d), there was more progression of shear crack relative to flexure crack. The load and deflection at this point were 3.4 kN and 2.53 mm, respectively. Because of this progression of web shear crack, which extended along the horizontal plane of the reinforcement, complete shear failure was observed at point-P at deflection value of 3.94 mm. At point-P the shear cracks expose the SMA composite rebar and even extended into T-beam flange. Point-P marked the culmination of the test. Summary of test data including mid span deflection, force, CMOD readings at different points during the tests is presented in Table 4.
Test results show excellent crack mouth closing of beam by SMA composite reinforcement even after yielding, especially in the 3rd cycle, once the composite rebar got fully engaged. In the 3rd cycle the SMA composite rebars were able to close the mouth opening by 96.4%, which provides insight to the re-centering and damage control ability of the SMA composite. Results show much superior performance of SMA composite rebars in terms of ductility, energy dissipation through hysteretic action, re-centering to recover residual deflections, without permanent damage to the reinforcement. This is not possible for beams reinforced with steel or conventional FRP rebars.
Ductility and hysteretic behavior
From Fig. 13, it is evident that the T-beam reinforced with SMA-FRP composite was able to dissipate energy through hysteretic action without accumulation of much residual deflection. This recovery of vertical deflection is because of re-centering capability while at the same time dissipating energy using wide hysteresis. For 1st cycle, the energy dissipated (area under force-deflection curve) was 1.66 J. For 2nd cycle, in which the SMA rebar showed forward and reverse stress transformation, was able to dissipate 2.9 J energy. This is 74% increase in energy dissipation which was effected by yielding of SMA composite rebar. In the 3rd cycle the beam showed wide hysteresis and was able to dissipate 3.72 J energy which is 124% increase from 1st cycle. The wider hysteretic behavior and re-centering capability exhibited by SMA composite reinforced concrete element is the hallmark of proposed composite rebar for use in seismic zones. T-beam exhibited yielding at a deflection value of 1.96 mm, thus it was able to achieve 1.2 and 1.8 ductility (μ) in 2nd and 3rd cycle. Beam reinforced with GFRP rebar would not be able to show any hysteresis and ductility if subjected to same loading protocol.
Concrete strains
As shown in Fig. 11, three strain gauges were used along the height of beam at mid span to measure strains in concrete. SG-1 was provided right above the notch, while SG-2 was installed on the web, right under the flange. SG-3 was installed on the T-beam flange side face. Reading from this strain gauge will give indication of crushing of concrete in compression. Readings from all three strain gauges are presented in Table 5. SG-1 was able to record initiation of crack but soon reached its capacity (7.46%) as the crack opened in the first cycle. SG-2 experienced compressive stresses initially before initiation of flexural crack. However right after the crack initiation and progression (point-B), SG-2 started to experience tensile strains. Like SG-1, SG-2 also reached its capacity by point-C. Table 5 shows that in 1st and 2nd cycles, SG-3 experienced compressive strains (negative sign) and exhibited residual compressive strains of 0.019% and 0.21%, respectively. In the 3rd cycle, SG-3 started to really stretch and reached its compressive capacity of 4%. During the compressive straining of SG-3 in 3rd cycle, the gauge did not get damaged and was able to record data during the unloading phase starting at point-M. The residual compressive strain in SG-3 at end of the 3rd cycle was 0.57%.
Conclusions
This study focused on the use of SMA-FRP composite as a new reinforcement for RC members to improve their re-centering capability. SMA-FRP composite specimens were tested under quasi-static loading to achieve constitutive behavior of these composites under tensions and compression. Microscopy using SEM was performed on SMA-FRP composite specimens after being tested to establish manufacturing technique of composites. SMA-FRP composite was further explored experimentally by manufacturing square rebars which were embedded in small scale concrete T-beam and tested under 3-point bending. Following conclusions can be drawn from the study presented in this paper:
1) Hybrid composite PRC specimens (manufactured with S-glass) showed better stiffness and strength as compared to the FRC specimen, while the FRC composite exhibited better ductility and energy dissipation capability. This makes FRC composite ideal for high seismic demand applications. It can be debated that both types of composites have their own advantages; hence could be used for different applications.
2) FRC-13 specimen, which experienced compressive cycles in addition to tensile cycles, was able to show ductility of 3.9 before exhibiting de-bonding between SMA wires and surrounding resin. Due to buckling of specimen during compression cycles, SMA wires de-bonded with resin after 6% strain cycle with the resin crushing in compression by the end of 7% strain cycle. Results show that compressive behavior does affect the tensile behavior of composite after 5% strains.
3) SEM images did not show any anomalies like de-bonding/de-lamination between resin and SMA wires in the manufacturing process. Efficacy of resin in filling all air voids and cavities around glass fibers and SMA wires was proven.
4) Results from flexural testing showed that SMA-FRP composite rebars are able to close the tensile cracks (measured using CMOD) due to its re-centering capability. In the 3rd cycle, the SMA composite rebars were able to close the mouth opening by 96.4% with negligible residual deflection.
5) SMA-FRP reinforced beam showed excellent hysteresis in the 3rd cycle. The hysteretic energy dissipated in the 3rd cycle was 124% more as compared to 1st cycle. The beam initially experienced flexural crack from the provided notch, but later developed shear crack because of absence of lateral reinforcement and aggregate interlock. The ultimate failure of the beam in 4th cycle was due to shear.
6) This study showed that the use of SMA-FRP rebars in concrete members reduces significantly the accumulation of permanent damage and residual displacements.The re-centering capability achieved by SMA-FRP composite reinforced beam while dissipating energy is a hallmark characteristics for superior performance in high seismic zones, thus ideal replacement for existing reinforcements (steel and conventional FRP).
Higher Education Press and Springer-Verlag Berlin Heidelberg
PDF
(702KB)
