Abstract
Coupled shear walls are widely adopted as lateral force-resisting systems in high-rise structures, requiring coupling beams to possess high flexural stiffness and shear strength for effective wall interaction, alongside superior ductility for seismic energy dissipation. Conventional steel bar-reinforced concrete (RC) or steel-concrete composite coupling beams generally suffer from common failure modes such as concrete spalling and steel buckling, which limit their ductility. This paper proposes a novel form of ductile coupling beam, fiber-reinforced polymer (FRP)-Steel-Concrete Coupling Beam (FSC-CB), which consists of an FRP confining tube and a concrete-encased steel I-section. The FSC-CBs feature an innovative truncated elliptical cross-section, which leads to not only effective confinement to the concrete, but also sufficiently large flexural stiffness and shear strength. An experimental program, including axial compression tests, four-point bending tests, and a reversed cyclic loading test, was executed to demonstrate the rationale of this proposed concept. The results confirmed the effectiveness of the proposed sectional form in both column and beam configurations, exhibiting significantly enhanced ductility and load-carrying capacity. Specifically, the FSC-CB specimens demonstrated excellent hysteretic behavior, high flexural and shear capacities, and stability, with common failure modes of existing coupling beams, including concrete spalling, steel buckling, and shear failure, largely suppressed due to the dual confinement mechanism provided by the FRP tube and the steel section. The present study verifies the structural advantages of the novel configuration and provides essential data for the design of comprehensive future experimental and analytical investigations into the seismic behavior of these hybrid coupling beams.
Introduction
Coupled shear walls are widely recognized as one of the most efficient lateral force-resisting systems for medium-to-high-rise buildings due to their high stiffness, strength, and energy dissipation capacity (Harries et al., 2000; Paulay and Santhakumar, 1976; Su and Lam, 2009). Within these systems, coupling beams serve a critical role in linking adjacent wall piers, functioning to transfer shear forces, coordinate lateral deformation, and dictate the overall seismic performance. To ensure satisfactory structural behavior during major seismic events, coupling beams must exhibit sufficiently large flexural stiffness and shear strength, along with high ductility for effective seismic energy dissipation (Harries et al., 2000; Paulay and Santhakumar, 1976; Su and Lam, 2009).
Existing coupling beam configurations primarily fall into three main categories: reinforced concrete (RC) beams, steel beams, and various steel-concrete composite beams (Gong and Shahrooz, 2001a; Han et al., 2015; Harries et al., 1993; Lam et al., 2003; Nie et al., 2014; Paulay, 1971; Paulay and Binney, 1974; Tegos and Penelis, 1988; Teng et al., 1999). Conventional RC coupling beams, particularly those with small span-to-depth ratios common in shear walls, are inherently susceptible to brittle shear or sliding shear failure mechanisms (Paulay and Binney, 1974). While alternative configurations, such as diagonally reinforced beams (Fortney et al., 2008; Paulay and Binney, 1974), rhombic truss reinforcement (Tegos and Penelis, 1988) and bundled diagonal reinforcement (Han et al., 2015), enhance shear capacity and ductility, their intricate reinforcement details significantly increase labor costs and complicate concrete placement. Moreover, these members are still vulnerable to the buckling of compression rebars following the spalling of concrete cover, which critically limits their post-peak deformation capacity and energy dissipation under extreme demands (Canbolat et al., 2005).
Extensive studies have investigated the behaviour of steel coupling beams in the form of a single steel I-section (e.g., Harries et al., 1993; Park and Yun, 2005; Wang et al., 2017) or a steel truss (e.g., Li et al., 2019; Lin et al., 2011)). Compared to RC coupling beams, steel coupling beams generally have a larger energy dissipation capacity (Fortney et al., 2007). The use of prefabricated steel coupling beams also leads to ease and good quality control of construction (Harries et al., 1993). However, the ductility of steel coupling beams is often limited by the potential compressive/shear buckling (e.g., at the web of an I-section) (Harries et al., 1993). To prevent/delay such buckling, a significant number of stiffeners are often needed (Harries et al., 1993; Wang et al., 2017), which may lead to significant increases in material/fabrication costs. To address steel buckling, composite forms such as concrete-encased steel I-sections (Gong and Shahrooz, 2001a; Lam et al., 2003) or concrete-filled steel tubular (CFST) beams (Nie et al., 2014) have been explored. The reinforced concrete encasement has been shown to effectively prevent/delay the web buckling of steel I-section, leading to enhanced structural performance and reduced need for web stiffeners (Gong and Shahrooz, 2001a, 2001b). Nevertheless, these composite beams are susceptible to spalling of the external concrete encasement, resulting in abrupt losses of stiffness and strength (Motter et al., 2017; Su and Zhu, 2005). In CFST coupling beams, the steel tube provides confinement to, and prevent the spalling of, the infill concrete, while its inward buckling is constrained by the concrete; however, inelastic outward buckling of the steel tube may still occur under large deformation, leading to degradation in the strength and ductility of the beam (Nie et al., 2014).
Fiber-reinforced polymer (FRP) composites, known for their high strength-to-weight ratio and corrosion resistance, have been widely used in strengthening RC structures due to their tailorability of mechanical properties (Teng et al., 2002; Yu et al., 2016). Several studies have investigated the use of FRP to strengthen RC coupling beams (e.g., Honarparast et al., 2019; Meftah et al., 2013; Riyazi et al., 2007). These studies have found that discrete closed FRP wraps or “U-shape” FRP wraps, with fibres in the transverse direction of the beams, could significantly enhance the shear strength and ductility of RC coupling beams (Li et al., 2016; Riyazi et al., 2007). Recently, the use of rectangular double-skin tubular members (DSTMs) as coupling beams has also been explored (Chen et al., 2021; Sui et al., 2021; Zeng, 2016). The hybrid DSTMs, consisting of an outer FRP tube, an inner steel tube and concrete filled between the two tubes, were first investigated at The Hong Kong Polytechnic University and have been extensively investigated since then (Jiang et al., 2023; Lin et al., 2024; Xie et al., 2023; Yu, 2007). While hybrid DSTMs generally possess excellent ductility due to the confinement of the FRP tube, their hollow core leads to potential inward buckling of the steel tube under large deformation which may cause degradation in the strength and ductility of the beams. Furthermore, a rectangular FRP tube was used in the above studies (Chen et al., 2021; Sui et al., 2021; Zeng, 2016) which is not effective in providing confinement to concrete due to its flat sides (Lam and Teng, 2003; Zeng et al., 2021; Zhu et al., 2020).
Proposed novel hybrid FRP-steel-concrete coupling beams
To meet the high energy dissipation demand in high seismicity regions, this study proposes a novel form of ductile coupling beams, namely, hybrid FRP-steel-concrete coupling beams (FSC-CBs). As depicted in Figure 1, the proposed FSC-CBs consist of an outer FRP confining tube and a concrete-encased steel I-section. Crucially, the cross-section adopts an innovative truncated elliptical geometry which is strategically designed to achieve high flexural stiffness and provide superior confinement to the concrete infill. This is realized through a dual confinement mechanism: the curved sides (part of an ellipse) of the FRP tube are known to provide much stronger lateral pressure to the concrete core compared to flat-sided sections (Teng et al., 2016), while the lateral expansion of the core concrete towards the flat sides of the FRP tube is strongly constrained by the flanges of the embedded steel I-section (Huang et al., 2021). As the concrete dilates, it comes into contact with the steel flanges, generating outward pressure on the flanges, which is transmitted to the steel web as transverse tensile stresses. Consequently, the steel section acts as an internal skeletal restraint that, in synergy with the external FRP tube, provides a dual confinement effect for the inner concrete. This multi-level restraint enhances the triaxial stress state of the concrete core, significantly improving its compressive strength and strain capacity. The flat sides, smoothly connected to the curved sides, also facilitate the accurate positioning of the steel I-section and the seamless seating of an adjacent floor slab. The curvature can be optimized by adjusting the elliptical aspect ratio to ensure strong confinement while minimizing the increase of mid-height sectional width. Typical cross-sections of FSCC and FCCC specimens.
Compared to conventional concrete-encased steel coupling beams (Gong and Shahrooz, 2001a), the novel feature of FSC-CBs is the use of an additional FRP confining tube which possesses only a small axial stiffness so that its confinement effectiveness is not compromised by buckling. This simple addition leads to a number of advantages: (1) enhanced ductility as the concrete is well confined by the FRP tube and the buckling of steel is well restrained; (2) enhanced shear capacity due to the contributions from the FRP tube; (3) ease for construction as the FRP tube serves as a stay-in-place formwork for concrete while steel bar reinforcement in the concrete encasement is no longer needed here; (4) excellent durability in harsh environments as the corrosion-resistant FRP tube acts as a permanent protective layer for steel and concrete. The FSC-CBs also offer significant structural and constructional benefits. By eliminating the need for concrete cover, the steel I-section can be positioned closer to the outermost edges, maximizing the effective depth and ensuring efficient utilization of the steel material for high flexural stiffness and strength. Moreover, as the main structural function of the FRP tube is to confine concrete (i.e., not as a longitudinal reinforcement), it can be relatively thin with a minimal material cost and is not required to be anchored into the wall piers, and thus existing, reliable beam-wall connection forms developed for steel or steel-concrete composite beams can be readily adopted (e.g., Motter et al., 2017; Park and Yun, 2006).
To demonstrate the rationale of the proposed concept and provide fundamental first insight, a pilot experimental program, including axial compression tests, four-point bending tests, and reversed cyclic loading test, was conducted in the present study. The experimental program and results are discussed in the following sections.
Experimental program
Overview and objectives
Details of test specimens.
Note: n: number of specimens; n
f
: number of FRP layers; b: cross-sectional width; h: cross-sectional height; L: specimen length; h
s
: steel sectional depth; b
s
: flange width; t
sf
: flange thickness; t
sw
: web thickness;
Test specimens and design
FSCC and FCCC specimens
A total of eight stub column specimens in two batches, including four FSCCs and four corresponding FCCCs, were fabricated and tested under uniaxial compression (Table 1). The cross-sectional configurations of the two types of specimens are illustrated in Figure 1(a) and (b), respectively. Both FSCC and FCCC specimens shared an identical truncated elliptical envelope shape, which had a cross-sectional width (b) and height (h) of 100.0 mm and 152.4 mm (excluding the thickness of the FRP tube, t f ), respectively, leading to a cross-sectional aspect ratio of around 1.5; the radius of the four rounded corners was set at 25 mm. All specimens had a clear overall length (L) of 305 mm. The only difference between the FSCC and FCCC specimens was the presence of an embedded steel I-section in the former. The steel I-sections in the FSCCs were characterized by a clear sectional depth (h s ) of 152.4 mm, a flange width (b s ) and thickness (t sf ) of 37 mm and 7.7 mm, respectively, and a web thickness (t sw ) of 4.5 mm. Two concrete compressive strengths (target values in standard concrete cylinders: 30 MPa and 50 MPa) were examined, while nominally identical steel sections and FRP tubes were used across all specimens. The fibers in the FRP tubes were mainly oriented close to the hoop (transverse) direction, and thus the FRP tubes possessed negligible axial stiffness and served mainly to provide confinement to the inner concrete.
Each specimen was assigned a name, starting with the specimen type (FSCC or FCCC), followed by F4 denoting a four-layer FRP tube and C30 or C50 denoting the target concrete compressive strength. The names conclude with a Roman numeral to distinguish between two nominally identical repeating specimens.
FSCB specimens
Two FSCB specimens were fabricated and tested under four-point bending to evaluate their flexural capacity and stiffness. The key design parameters are summarized in Table 1. Both specimens featured a span length (i.e., distance between the centre lines of the two supports) of 1350 mm. The external cross-sectional envelope was similar to that of the columns as described above; however, the steel I-sections were slightly different with h s = 150.0 mm, b s = 37 mm, t sf = 6.4 mm, and t sw = 4.6 mm. The primary variable investigated in this series was the FRP tube thickness. The two FSCB specimens were named FSCB-F4 and FSCB-F6, corresponding to the use of 4-layer and 6-layer FRP tubes, respectively.
FSC-CB specimen
One full-scale FSC-CB specimen was fabricated and tested under reversed cyclic lateral loading. The cross-sectional configuration shown in Figure 1(a) was again adopted but the dimensions were increased relative to the columns and beams, with the details of summarized in Table 1.
To accurately simulate the boundary conditions of a coupled shear wall system, the FSC-CB was fabricated with two RC end blocks, as shown in Figure 2. The specimen was rotated 90° for testing, allowing the horizontal actuator to impose lateral shear forces consistent with those induced by earthquake loading. To ensure robust force transfer, the embedded steel I-section in the beam was securely anchored in each end block (see Figure 2). The two ends of the FRP tube were also extended 415 mm into the end blocks to suppress potential sliding shear failure and local damage (Figure 2). Geometric dimensions and reinforcement details of the FSC-CB specimen.
Material properties
FRP tubes
All the FRP tubes were fabricated via filament winding using foam molds with the desired shapes as the mandrels (Figure 3). 2400-tex (g/km) direct roving glass fibers with a tensile strength of 2741 MPa and an elastic modulus of 81.2 GPa, as provided by the manufacturer, and epoxy resin were used for the filament winding. The nominal fiber thickness of each ply of all the FRP tubes was 0.17 mm. The FRP tubes of the columns had 4 layers of fibers with a winding angle of around 89.5°, while those of the beams had 4 or 6 layers of fibers, both wound at an angle of 80°. The FSC-CB specimen utilized a 6-ply FRP tube with a winding angle of 80°. Fabricating of FRP tubes via filament winding process.
Concrete
The compressive strength of each batch of concrete was determined by testing three standard 300 mm × 150 mm cylinders in accordance with ASTM C39/C39M-20 (2020). The average compressive strengths of Batch 1 and Batch 2 of the columns were 31.1 MPa, and 51.2 MPa, respectively. The concrete compressive strength of the beam specimens was 57.3 MPa. In the reversed cyclic loading test, the coupling beam and the RC end blocks were cast in two separate batches, with the compressive strengths measured to be 24.7 MPa and 61.6 MPa, respectively.
Steel I-sections and reinforcement
Tensile tests on steel coupons cut from the internal steel I-sections were conducted in accordance with ASTM E8/E8M-16a (2016). The I-sections for the compression and four-point bending tests were fabricated by trimming flanges of standard hot-rolled I-sections, while that for the reversed cyclic test was fabricated by welding steel plates. The average yield stresses of steel I-sections in the compression tests and the bending tests were 370.9 MPa and 290.6 MPa, respectively, while that in the reversed cyclic loading test was 310.8 MPa for both web and flanges.
Preparation of specimens
FSCC, FCCC, and FSCB specimens
The FRP tubes of the FSCC, FCCS and FSCB specimens were used as molds for casting concrete. For FSCC and FSCB specimens, the steel I-sections were first instrumented with strain gauges and then inserted into the FRP tubes. The resulting assemblies were secured onto timber plates to maintain precise vertical alignment (Figure 4(a) and (b)), and then concrete was cast into these assembled formworks. To avoid the risk of undesirable local premature failure, additional CFRP strips were wrapped around the top and bottom ends of all column specimens prior to testing. Preparation process of test specimens.
FSC-CB specimen
The specimen was fabricated in two distinct steps: (1) fabricating the coupling beam and (2) casting the RC end blocks. The procedure for fabricating the coupling beam was similar to that for the FSCC and FSCB specimens, while the second step involved accurately positioning the coupling beam and the end steel I-sections within the pre-assembled reinforcement cages of the end blocks and then hoisting the entire assembly into the wooden formwork (see Figure 4(c)) for casting the second batch of concrete to form the RC end blocks.
Test set-up, instrumentations, and loading scheme
Compression tests on FSCC and FCCC specimens
The compression tests were conducted on an MTS testing machine with a capacity of 4600 kN (Figure 5(a)). Two linear variable displacement transducers (LVDTs) were installed to measure the axial shortenings of the 100-mm mid-height region. Additionally, four LVDTs were placed between the loading platens to capture the overall axial shortenings. Extensive strain gauges were installed to capture the axial and lateral strains of the FRP tube and the steel I-section, as shown in Figure 5(b). A monotonic compressive load was applied under displacement control at a constant loading rate of 0.3 mm/min until the specimen failure. Test set-up and instrumentations of column specimens.
Bending tests of FSCB specimens
A loading frame equipped with a 500 kN static actuator was employed to apply vertical monotonic loads, and the vertical load was transferred through a spreader steel beam, resulting in two symmetric point loads with a distance of 450 mm on the specimens (Figure 6(a)). To prevent out-of-plane lateral instability, lateral supports were provided at both reaction points and near the two loading points as shown in Figure 6(b). Six LVDTs were deployed to capture the vertical deflections of the beam specimens: two at the mid-span, two at the supports, and another two beneath the loading points. Comprehensive strain measurement was performed on both the FRP tubes and the steel I-sections, with the detailed strain gauge layout shown in Figure 6(c). The monotonic load was applied with a displacement control mode with a rate of 0.2-2 mm/min. Test set-up and instrumentations for four-point bending tests.
Reversed cyclic loading test of FSC-CB specimen
The set-up for the reversed cyclic loading test is illustrated in Figure 7(a) and (b). The system included a reaction wall, an L-shaped beam, lateral bracing, a 1000 kN horizontal actuator, and two 500 kN vertical actuators. The horizontal load was applied via the L-shaped beam, and the horizontal actuator was in line with the coupling beam’s centerline. The L-shaped beam was pin-connected to the two vertical actuators, which prevented rotation at the specimen’s top surface. Test set-up and instrumentations for reversed cyclic loading test.
Instrumentation for the test involved both a non-contact optical technique and traditional contact sensors (i.e., strain gauges and LVDTs) (Figure 7(b)). Several LVDTs and one laser displacement sensor were used to measure deformations of the coupling beam specimen, as shown in Figure 7(b). Among them, the measurements Sensor H1 and LVDT H2 were used to calculate the lateral drift Δ (i.e., by taking the difference between the two readings) and then the drift ratio Δ/L (where L is the clear span of the coupling beam) in the analysis of results. Furthermore, strain gauges were used to record the local strain development of the steel I-section and the FRP tube. The cyclic loading was applied in displacement control based on previous work of the last author (e.g., Hou et al. 2018).
Test results and discussions of compression tests
General observation and failure mode
All the FCCC and FSCC specimens exhibited excellent ductility, characterized by large axial deformations, until they failed suddenly due to hoop rupture of the FRP tubes which generally occurred on the flat sides near the corners and the specimen mid-height (see Figure 8). Following the failure of the FSCC specimens, the internal steel I-sections were extracted for post-test inspection, and no obvious global or local buckling was observed (Figure 9). In the compression tests, the axial compressive forces/stresses/strains and circumferential tensile stresses/strains are taken as positive unless otherwise specified. Failure mode of FCCCs and FSCCs. Typical steel I-sections after tests.
Axial load-axial strain curves
The axial load-axial strain curves of the test specimens are shown in Figure 10, in which the axial strains were obtained from the average readings of the LVDTs measuring the full-length shortenings of the specimens, noting that local strains measured from the strain gauges on the FRP tubes or LVDTs covering only the mid-height region may not accurately capture the overall deformation of the concrete due to localized behaviors (Huang et al., 2020). The curves are all terminated at the point of FRP rupture. Axial load-axial strain curves of the column specimens.
The axial load-strain behaviors differed significantly based on the concrete batch utilized (Figure 10(a) and (b)). The FCCC and FSCC specimens fabricated with the lower strength concrete (Batch 1) exhibited the ideal ascending bilinear axial load-strain curves (Figure 10(a)), indicating that the specimens were provided with a strong FRP confinement. By contrast, specimens cast with the higher strength concrete (Batch 2) also displayed two-stage axial load-strain curves, but their second branch is generally characterized by a softening segment (gradual load reduction) which may then be followed by a hardening segment (gradual load recovery). The observed differences in behavior between the two batches of specimens are attributed to the variation in concrete compressive strength: a higher concrete strength results in a lower FRP confinement ratio, leading to less effective enhancement (Lam and Teng, 2003; Lin et al., 2016; Lin and Teng, 2020; Teng et al., 2015). Due to the above reason, the specimens in Batch 1 achieved much larger ultimate axial strains (approximately 2.75%). Nevertheless, the specimens in Batch 2 still reached ultimate axial strains of around 1.3%, which is 5-6 times the crushing strain of plain concrete.
To clarify the confinement mechanism, the axial load-strain curves of the FSCC specimens are compared against the corresponding superposition curves in Figure 11. The superposition curve represents the arithmetic sum of the load resistances from the two individual components: the confined concrete (derived from the corresponding FCCC specimens) and the steel I-section (derived from the steel coupon tests). It is a notable observation that the measured axial load-strain curves of the FSCC specimens were generally higher than the corresponding superposition curves throughout the loading history. This outcome indicates that the structural performance of the hybrid section exceeded the simple combined resistance of its constituent parts (i.e., confined concrete and steel). This superior performance strongly suggests that the embedment of the steel I-section resulted in a synergistic effect within the system. This improvement is primarily attributed to the additional confinement to the concrete provided by the steel I-section, especially for the concrete near the flanges (Huang et al., 2023). Comparison of test and superposition load-strain curves for FSCC specimens.
Test results and discussions of four-point bending tests
General observation and failure mode
The FSCB specimens developed stable load resistances as the vertical loads were progressively applied. A clear transition from elastic to inelastic behavior was observed for both FSCB-F4 and FSCB-F6 when the mid-span deflection reached approximately 10 mm, accompanied by a significant reduction in the flexural stiffness of the beams. During the inelastic stage, transverse cracks appeared on the FRP tubes within the pure bending region (Figure 12). However, these cracks did not impair the load resistances of the FSCBs, as the FRP tubes had fibers primarily oriented in the hoop direction and contributed negligibly to the longitudinal flexural resistance. The tests were terminated when sliding occurred at the supports of both specimens without rupture of FRP tube in the hoop direction, thereby demonstrating the excellent ductile behavior of these specimens. In the four-point bending tests, downward thrusts of the actuator, downward deflections, and tensile strains are taken as positive. Failure modes of test FSCB specimens.
Load-deflection curves
Figure 13 shows the load-mid-span deflection curves of specimens FSCB-F4 and FSCB-F6. The beam specimens exhibited typical bi-linear load-deflection curves with a clear ascending trend in the post-yield stage, demonstrating their excellent flexural ductility. Despite the difference in FRP tube thickness, the load-deflection curves of the two specimens were similar, implying that the 4-ply FRP tube was already sufficient to provide the necessary confinement. It should be noted again that the loss of load resistances and the subsequent termination of the tests were not caused by the rupture of FRP tubes or internal material failure, but rather by support slips. Therefore, the FSCB specimens still possessed significant untapped load-carrying potential at the time of test cessation, indicating that the true flexural capacity was likely higher than the recorded maximum load. Load-deflection curves of FSCB specimens.
Longitudinal strain development
Figure 14 illustrates the development of longitudinal strains (i.e., axial strains) measured on the top (compression) and bottom (tension) fibers of the FRP tubes at the mid-span. In the early loading stage, the relationship between longitudinal strain and mid-span deflection was approximately linear for both specimens. However, when the deflection reached approximately 6-8 mm, the longitudinal strain gauges installed at the tension fiber ceased to show an increase in readings. This behavior is attributed to the onset of interlaminar and/or resin cracking in the FRP tube at the extreme tension zone, leading to local unloading of FRP tube where the strain gauges were attached. Conversely, the longitudinal strains at the extreme compressive fiber of the FRP tubes continued to increase almost linearly with the deflection until the measured strain reached approximately 20,000 με (micro strain) in both specimens, when the strain gauges damaged without structural member failure. It should be noted that the fibers of the FRP tubes were oriented nearly perpendicular to the longitudinal direction. As a result, the measured longitudinal strains reflected primarily the deformation of the resin matrix in the longitudinal direction. Longitudinal strain development in FRP tube of FSCB specimens at the mid-span.
Test results and discussions of reversed cyclic loading test
General observation and failure mode
Figure 15 shows the FSC-CB specimen after the reversed cyclic loading test, which exhibited superior shear capacity and excellent ductility. The hysteretic loops observed were stable, with no apparent pinching or asymmetry, indicating robust energy dissipation. Minor cracks were initiated at the beam-wall interface in the RC end blocks at a chord rotation of approximately 0.03 rad, which propagated outward from the interface towards the edges of the end blocks (Figure 15(c)). At this stage, the coupling beam itself largely remained within the elastic range, and the applied load continued to increase. The specimen reached its transition point with an obvious decrease in stiffness at a rotation level of about 0.03 rad, beyond which the stiffness decreased, yet the applied load continued to increase. The test was terminated when the actuators reached their load limits without an obvious failure in the specimen. By the time of cessation, the specimen had undergone significant deformation with slight fiber ruptures and resin cracks in the FRP tube near the beam-wall interface (Figure 15(b)), which occurred without any noticeable reduction in the resisted load. This sustained performance was further corroborated by the relatively small hoop strains measured on the FRP tube, indicating substantial residual confinement capacity remained in the system. FSC-CB specimen after test.
In the present study, the tested FSC-CB exhibited highly ductile behavior, and no clear load reduction was observed even at large chord rotation levels. Therefore, the tested FSC-CB specimens were not deemed to reach the failure state. In a more general case, the failure of an FSC-CB may be defined as the occurrence of extensive rupture of the FRP tube or a substantial load drop after the peak, whichever occurs first.
Hysteretic curves
Figure 16 shows the hysteretic load-chord rotation curves and the corresponding skeleton curve for specimen FSC-CB. It is evident that the specimen exhibited full and stable hysteretic loops with minimal pinching, demonstrating superior energy dissipation capacity. The skeleton curve displays a bilinear characteristic without any load degradation. The specimen reached a substantial maximum chord rotation of 0.1 rad at the end of test, confirming its excellent deformation capacity. Upon the termination of loading, the specimen showed no significant sign of degradation. The promising hysteretic behavior validates the beneficial composite interaction among the different components in the coupling beam system, as explained in Section 2. Hysteretic load-chord rotation curves.
Energy dissipation capacity
The dissipated energy (E) for each cycle was quantitatively evaluated by calculating the area enclosed by the corresponding hysteretic loop. Figure 17 illustrates the dissipated energy (E) of each cycle and the accumulated dissipated energy (E
a
) at various chord rotation levels. As shown in the figure, the dissipated energy per cycle increased significantly with the increment of the chord rotation (θ), especially after the plastic deformation occurred at a chord rotation of around 0.03 rad. At the maximum achieved chord rotation of 0.1 rad, the dissipated energy per cycle reached approximately 30 kN·m, resulting in a large accumulated dissipated energy of approximately 227 kN·m (Figure 17(b)). In comparison with other types of coupling beams, which frequently suffer from severe strength degradation and prominent pinching especially at later loading stages (Li et al., 2020; Lim et al., 2016), the proposed hybrid system of coupling beam not only delayed damage progression but also significantly enhanced the overall seismic performance through stable and high energy dissipation. Energy dissipation.
Conclusions
This paper has presented a novel type of hybrid FRP-concrete-steel coupling beam (FSC-CB). To validate the proposed concept and to evaluate the performance of structural members with the novel sections, an experimental program was carried out, including axial compression tests, four-point bending tests, and a reversed cyclic loading test. Based on the experimental results and discussions, the following conclusions can be drawn. (1) The structural performance of FSCC specimens generally exceeded the simple combined resistance of their constituent parts (i.e., confined concrete and steel), successfully verifying the existence and rationale of the dual confinement (provided by the FRP tube and steel I-section) mechanism in the proposed coupling beam section. The steel I-sections in the FSCC specimens possessed no obvious global or local buckling due to the confinement from the FRP tube and the surrounding concrete. (2) The FSCBs demonstrated excellent ductility and superior load-carrying capacity, with typical bi-linear load-deflection responses, characterized by an ascending second branch. Pronounced hoop strains developed in FRP tubes at top corners and flat sides within the compression zone, implying that FRP confinement was effectively activated. (3) The FSC-CB specimen exhibited full and stable hysteretic loops with minimal pinching effects and limited strength and stiffness degradations, demonstrating the great potential held by the proposed coupling beams in resisting seismic forces and dissipating energy during earthquakes. The promising hysteretic behavior also validated the beneficial composite interaction among the different components in the coupling beam system. (4) The geometric and material properties of FRP tubes and steel I-sections significantly influence the overall performance of FSC-CBs. The quantitative identifications of the effects of different parameters, however, require further investigation in the future. Besides, in practical applications, the anchorage design of coupling beams remains a critical issue and constitutes an important direction for future research on FSC-CBs.
Footnotes
Acknowledgements
The authors gratefully acknowledge the assistance of Mr Wenxuan Chen in the reversed cyclic loading test.
Author Contributions
D. Han: Investigation, Data curation, Writing – original draft. T. Yu: Conceptualization, Methodology, Supervision, Validation, Writing – review and editing, Funding acquisition. G. Lin: Validation, Writing – review and editing. L. Huang: Supervision, Methodology, Validation. W. Hou: Supervision, Methodology, Validation.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The authors are grateful for the financial support received from the Research Grants Council of the Hong Kong Special Administrative Region (Project No. 15222321) and the State Key Laboratory of Climate Resilience for Coastal Cities at the Hong Kong Polytechnic University.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.