Abstract
A new type of beam-to-column joint used in prefabricated concrete frames was proposed in this study. In this joint, the longitudinal bars at the top of the beam are anchored to the column using straight thread sleeves, and the bars at the bottom are welded to the steel fastener that is bolted to the column. Cyclic loading tests of three specimens, namely, two beam–column joints of this type and a cast-in-place beam–column joint, were conducted to study the seismic behavior and feasibility of this type of joint. The difference between the two prefabricated joints is the shape of the holes on the end plate. Failure modes of the specimens were observed and analyzed. The hysteretic curves, bearing capacities, stiffness degeneration, ductility, and energy-dissipating capacities of the specimens were compared and studied. Test results indicated that all beam–column joints exhibited beam hinge failure. No slippage was observed between the concrete and horizontal plates of the steel fasteners used in the new type of joint. The bearing capacity and initial stiffness of both prefabricated specimens compared with the cast-in-place ones were increased. The steel fastener could increase the distance between the plastic hinge and the side surface of the column while enlarging the length of the plastic hinge. The trend of energy dissipation and stiffness degeneration of the specimens were similar, and the ductility coefficient ranged from 2.7 to 4.91. The displacement angles of the joints exceeded 1/50 before the failure of the specimens. The mechanical behavior of both prefabricated joints was similar, but the joint with U-shaped holes on the end plate was convenient to create.
Introduction
Prefabricated concrete (PC) structures exhibit several advantages, such as less in-place concrete pouring, higher construction efficiency, lower labor cost, and more reliable member quality, compared with traditional cast-in-place concrete structures. The Modernization Development Guidelines on the Building Industry issued by the Chinese government has clearly pointed out that prefabricated buildings will account for over 20% of new buildings in China by 2020, and the proportion will reach more than 50% by 2025 (Liu et al., 2017). However, the destruction of PC structures after earthquakes, such as the 1994 Northridge (Mitchell et al., 1995), 1998 Adana-Ceyhan, Turkey (Gulkan, 1998), 2012 Emilia (Zoubek et al., 2013), and 1999 Turkey Marmara (Arslan et al., 2006) earthquakes, shows that joints between PC structural members are susceptible to failure, which may cause local or overall structural collapse. The qualities of industrially produced members can be easily ensured. However, if the reliability of prefabricated beam-to-column joints cannot meet the design requirement, then the integrity and seismic performance of PC structures are unguaranteed. The quality of the joints used in a PC frame structure is the key factor that influences their mechanical behavior (Park, 1995). Further investigation of the properties of prefabricated beam-to-column joints, especially their seismic properties, is essential for the promotion of PC structure application.
Several beam-to-column joints used in PC frame structures have been proposed. The prefabricated joints can be divided into wet and dry types, according to whether concrete pouring is necessary or unnecessary on site. Zhao et al. (2017) studied the mechanical behavior of beam-to-column joints that used pressed sleeves to connect longitudinal bars. Results showed that the sleeves could effectively transfer the force, and the prefabricated joints demonstrated higher bearing and deformation capacities compared with the cast-in-place ones. Guan et al. (2016) and Chen et al. (2012) investigated the seismic behavior of prefabricated beam-to-column joints connected by grouting sleeves. Their studies showed that prefabricated joints exhibited similar seismic performance compared with cast-in-place specimens. Cheok (1991, 1993) conducted experimental tests on an eight-scale model of precast concrete beam-to-column connections. The research program considered variables, such as location of the post-tensioning steel, use of post-tensioning bars versus prestressing strands, and fully bonded versus partially bonded strands. Compared with those of the monolithic test specimens, results indicated that the post-tensioned precast concrete specimens showed similar connection strengths, higher ultimate displacement ductility, and total energy dissipation to failure but lower energy dissipation per cycle. Ha et al. (2014) proposed a new type of PC beam-to-column joint that used U-shaped strands to transfer force. Results of the cyclic loading tests indicated that this new type of joint demonstrated a promising performance under seismic load. Ertas (2006) conducted experimental tests on five concrete beam-to-column joints, including a cast-in-place specimen, two postcast specimens, a welded specimen with corbel, and a bolted specimen. The prefabricated joint specimens exhibited similar mechanical behavior. All specimens, except for the welded joint, could sustain up to 3.5% story drift. A risk of the steel box or pipe of the bolted joint was detected with respect to the concrete beam and proposed design suggestions. Hanaor et al. (1998) applied high-strength bolted joints to the steel end plate of a PC beam-to-column joint to solve the strain transfer problem between the bolts and longitudinal bars of the prefabricated beam. Results indicated that this type of joint could enhance the energy-dissipating capacity of the specimen. Parastesh (2014) proposed a new connection that set the U-shaped groove at the prefabricated beam end. The seismic test showed that this joint exhibited considerably higher ductility and energy dissipation compared with similar monolithic specimens. Kulkarni et al. (2008) used numerical simulation to investigate the influences of axial compression ratio, thickness of junction plate, strength grade of rebar, and other factors on the seismic performance of PC beam-to-column joints.
Different types of prefabricated beam-to-column joints still demonstrate certain difficulties in application, especially in terms of mechanical behavior and construction. When a wet-type joint connection is adopted, concrete needs to be poured to the core of the joint after erection of the members to connect the prefabricated members to an integral structure. In this manner, the mechanical behavior of prefabricated joints can be ensured. The construction process will be difficult, because the reinforcement ratio is high in the joint area. For dry-type joints, the duct should be reserved in the core of the joint for prestressed and bolted connection to weaken the integrity of concrete in the core of the joint to a certain extent; thus, ensuring the seismic properties of joints is difficult. If the duct is particularly long, then the joint stiffness will hardly satisfy the requirements for rigid joints and cannot realize the same ductility and deformability as cast-in-place joints. Therefore, prefabricated beam-to-column joints must be further explored to improve fabrication efficiency and connection quality.
A new type of beam-to-column joint used in PC frames was proposed in this study. The top longitudinal bars in the beam were anchored to the column using straight thread sleeves, whereas the bars at the bottom with steel fastener were bolted to the column. As shown in Figure 1(a), the beam and column were prefabricated, whereas the column was connected to the core of the joint. The steel fastener was pre-embedded at the bottom of the beam to meet the bolted connection with the column by applying prestressing force to the high-strength screws. Figure 1(a) shows that sleeve connection was met between the longitudinal bar at the top of the beam and the straight thread sleeves that were pre-embedded in the column. Then, the monolithic joint was formed by pouring secondary concrete on the upside of the prefabricated beam and the preformed groove at the end of the beam. As illustrated in Figure 1(b) and (c), the pre-embedded steel fasteners at the bottom of the prefabricated beam presented two forms—U-shaped and round holes on the end plate. The high-strength screws could also provide a temporary support for prefabricated beam with steel fasteners; therefore, construction would be convenient. Pouring concrete to the core of the joint was unnecessary, and pouring secondary concrete was simple and convenient. Pseudostatic tests were conducted for two prefabricated joint specimens and a cast-in-place joint specimen to study the seismic performance of this type of joint. The seismic behavior of the joint specimens, including crack development, failure mode, ductility, and energy-dissipating capacity, were comparatively analyzed, thereby providing a reference for prefabricated joint engineering practice.
Prefabricated concrete frame beam-to-column upper anchor lower bolt joint: (a) assembly diagram of prefabricated concrete frame beam-to-column joint; (b) diagram of U-hole steel fastener; and (c) diagram of round-hole steel fastener.
Test program
Details of test specimens
Three full-size concrete frame beam-to-column joint specimens were designed in accordance with the design principles of “strong joint and weak member” and “strong column and weak beam.” The beam and column sizes of XJ, ZP1, and ZP2 were the same. The length, width, and height of the beam were 1700, 300, and 600 mm, respectively. However, the loading point was 1500 mm away from the column edge. The length, width, and height of the column were 550, 550, and 2800 mm, respectively. The height of the hinged support at the column base was 200 mm, and the designed axial compression ratio was 0.08. The beam-to-column reinforcement detailing of three specimens was consistent, and the details of these specimens are shown in Figure 2. The joint structures of the specimen were different. As shown in Figure 2(a), the entire longitudinal bar of the beam to XJ ran through the core of the joint. Figure 2(b) shows that the high-strength screws of ZP1 were pre-embedded in the concrete column, and the hole at the end plate was U shaped. Moreover, the top longitudinal bar of the beam was connected through straight threaded sleeves, whereas that at the bottom was connected through high-strength screws. A prestressing force of 355 kN was applied. As shown in Figure 2(c), the duct was reserved in the column of ZP2, the hole at the end plate was circular, and the top longitudinal bar of the beam was connected through a straight threaded sleeve. The high-strength screws were inserted in the column during assembly and filled with high-strength grouting materials. Then, a prestressing force of 355 kN was applied.
Details of the test specimens (in mm): (a) XJ, (b) ZP1, (c) ZP2, and (d) steel fastener.
The width, height, and thickness, except for the shape of the hole, of end plates ZP1 and ZP2 were 155, 300, and 30 mm, respectively. The width of the U-shaped hole of ZP1 was 33 mm, whereas the diameter of the round hole of ZP2 was 33 mm. The heights of the horizontal plate, baffle, and connecting plate were all 300 mm; their widths were 235, 100, and 125 mm; and their thicknesses were 20, 10, and 25 mm. The thickness of ribbed plates 1 and 2 was 10 mm, and the width and height were 125 and 200 mm, respectively. The two right-angle sides of ribbed plate 2 were 100 and 75 mm. Six studs were welded onto the horizontal plate to transfer the shear force between the steel fastener and concrete beam. Each stud exhibited a diameter of 8 mm and a length of 50 mm. The external diameter and length of the straight thread sleeve were 31 and 45 mm, respectively. This sleeve was connected to the rebar with a diameter of 20 mm. The measured properties of steel and rebar are listed in Table 1.
Rolled steel material properties.
t: thickness of steel plate; d: rebar diameter; Fy: yield strength of rebar and steel plate; Fu: tensile strength; δ: elongation.
Figure 1(c) showed that each joint specimen was separately poured twice in the experiment, with the prefabricated column and bottom of the prefabricated beam poured first. Subsequently, the reinforcement bar on top of the beam and preformed groove was poured. The same batch of concrete was used in each pouring process of all specimens. Cubic compression tests were conducted to acquire the material properties of concrete according to the Chinese Standard GB/T50081-2002 (2002). Calculation results show that the compressive strength values of the cubes at 28 days were 36.8 and 50.0 MPa.
Test setup and loading history
A schematic view of the test setup is illustrated in Figure 3(b). An end plate was bolted to the base of the column, and the column was connected to the hinged support at the bottom. The applied lateral support restricted the horizontal displacement at the top of the column. The hinged support was set to restrict the horizontal displacement at the bottom of column. A vertical concentrated force load was applied at the top of the column using a hydraulic jack to simulate the axial force of the column. The jack was driven by an oil pump, and the vertical concentrated force load was kept constant by ensuring that the oil pressure of the oil pump was unchanged. Cyclic load was applied to the free end of the beam using a hydraulic jack that was driven by another oil pump. The pressure borne by the top of the prefabricated beam was positive, and then the phenomenon description and data processing were used as the standard. The force sensor was placed at the free end of the beam to measure vertical force.
(a) Loading history and (b) test setup (in mm).
The loading history of the experiment is shown in Figure 3(a). The load–displacement hybrid loading mode was used at the free end of the beam. Before the specimen yield, load-controlled loads were used. Each load stage was repeated once with a 20 kN increment. After the specimen yield, displacement-controlled loads were applied at displacement levels of 1Δy, 2Δy, 3Δy, and so on. Each load step was repeated twice until the load dropped to 85% of the peak load or until the failure of the steel fastener; thus, continuous loading was inappropriate.
Instrumentation
The arrangement of displacement transducers is shown in Figure 4(a). T1 measured the vertical displacement of the free end of the beam. T2 and T3 were used at the ends of the column to measure possible horizontal slippage. T4 measured the relative rotation angle of the beam and the column. T5 and T6 were used to measure relative displacements at diagonal measuring points to determine the shear deformation of the core of the joint. The strain gauge was used to measure the strain of the rebar and steel plate, and the arrangement of the strain gauge is shown in Figure 4(b) to (d). Strain gauges R1–R20 measured the longitudinal bar strain of the columns, and R21–R28 calculated the longitudinal bar strain of the beams. G1–G3 measured the stirrup strain in the core of the joint, whereas G4 and G5 were used to gauge the stirrup strain of the beam. GB1–GB6 were used to measure the plate strain of the steel fasteners.
Arrangement of displacement meter and strain gauges: (a) arrangement of displacement meter, (b) strain gauge arrangement of the XJ specimen, (c) strain gauge arrangement of the ZP1 specimen, and (d) strain gauge arrangement of the ZP2 specimen.
Test results
Test phenomena and failure modes
The main test results are shown in Table 2. The failure modes of the specimens are shown in Figure 5(a) to (e). All specimens exhibited beam hinge failure. Furthermore, no signs of severe buckling were observed for the longitudinal bars of the specimens in which rebars were partially exposed even at the end of the tests. Spalling of the concrete cover, longitudinal bar buckling, and rebar failure were detected through visual inspection.
Test results of specimens in the main phases.
Failure patterns and cracks of specimens: (a) XJ, (b) ZP1, (c) ZP2, (d) weld joint fracture (the connection plate and rebar weld joint of the ZP1 specimen are intact), and (e) weld joint fracture of the ZP2 specimen.
When XJ was loaded to 80 kN, several slightly vertical flexural cracks appeared at the beam end, which is near the core of the joint (abbreviated as “fixed beam end”). The longest crack extended to 1/5 height of the beam section. At this time, the cracking load of XJ was 80 kN. After the load was controlled by displacement loading, cracks propagating at the top and bottom of the beam were gradually connected diagonally with the increase in displacement. These cracks continued to extend along their respective directions and presented crossed ones. The maximum crack width reached 0.15 mm, and the tensile crack developed within an approximately 1050 mm range of the fixed beam end. When the displacement reached 12 mm, the joint specimen started entering the yield phase, and a plastic hinge was formed within an approximately 100 mm range at the fixed beam end. Moreover, when the displacement reached 32 mm, the specimen reached the maximum load (Pmax = 195 kN). Afterward, the bearing capacity gradually declined and concrete crack development accelerated. When the displacement reached 36 mm, the load declined to 85% of the maximum load. Simultaneously, spalling of the concrete cover in the plastic hinge area occurred under tension and compression. Specimen failure is shown in Figure 5(a).
Cracks initiated at 180 mm from the top of the fixed beam end when ZP1 was loaded to –60 kN. Diagonal cracks appeared at 400 mm from the bottom of the fixed beam end when loading was increased to 80 kN. This finding indicates that the cracking load of ZP1 was 60 kN. Considerable curved diagonal cracks appeared 700–1100 mm away from the bottom of the fixed beam end and extended to the top of the beam. When the displacement reached 12 mm, ZP1 entered the yield phase. The positive load was 213 kN, and the negative one was –172 kN. A plastic hinge was formed within the 50–250 mm range from the fixed beam end. The main diagonal cracks were gradually formed along the steel fastener edge at the bottom of the fixed beam end, which intersected with the upper cracks. The maximum crack width was 7 mm. Simultaneously, spalling of the concrete cover occurred. When the displacement reached 32 mm, ZP1 reached the ultimate load. The maximum positive and negative loads were 234 and –219 kN, respectively. The fracture of weld between ribbed plate 1 and the baffle of ZP1 under 36 mm displacement is shown in Figure 5(d). Therefore, the failure of ZP1 could be analyzed. The load declined by 91% of the peak load, and specimen failure is shown in Figure 5(b).
When ZP2 was loaded to –60 kN, cracks initiated at 200 mm from the upper part of the fixed beam end. Diagonal cracks initiated at 450 mm from the bottom of the fixed beam end under an 80 kN condition. This finding indicates that the cracking load of ZP2 was 60 kN. After displacement loading was adopted for control, horizontal cracks appeared at the connection plate of the beam bottom with the increase in displacement. A diagonal crack propagated toward an inclined upward direction. Substantial curved diagonal cracks appeared from 650 mm to 1150 mm at the bottom of the fixed beam end, which extended to the top of the beam. When the displacement reached 8 mm, ZP2 entered the yield phase. The positive and negative loads were 199 and –160 kN, respectively. A plastic hinge was formed within the 50–300 mm range from the upper part of the fixed beam end. Two main diagonal cracks were gradually formed along the steel fastener edge at the bottom of the fixed beam end, which intersected with the upper crack. The maximum crack width was 5 mm, and spalling of the concrete cover occurred. When the displacement reached 32 mm, the ultimate load was reached. The maximum positive and negative loads were 232 and –210 kN, respectively. The fracture of weld between ribbed plate 1 and the baffle of ZP2 when the displacement became 36 mm is shown in Figure 5(e). Therefore, the failure of ZP2 could be analyzed. This specimen failure is shown in Figure 5(c).
Comparative analysis of the failure phenomena of the three specimens showed that no cracking of concrete was detected in the core of all joint specimens throughout the loading process. All specimens exhibited beam hinge failure. Cracks were uniformly distributed in XJ and fully propagated in the plastic hinge area. Most cracks were vertical ones passing through the upper and lower parts. The length of the plastic hinge of ZP1 and ZP2 was enlarged with minor concrete spalling under the influence of steel fasteners. Most cracks at the bottom of the beam were diagonal. The steel fastener could increase the distance between the plastic hinge and the side surface of the column by 50 mm compared with XJ. This finding indicates that these new type of joints could shift the plastic hinge outward. At the end of the trial period, the concrete above the steel fasteners of ZP1 and ZP2 was chipped away. The stud was effectively bonded to the concrete, and no slippage was observed between the horizontal plate and concrete, thereby indicating that they were reliably connected.
Analysis of test results
Hysteretic and skeleton curves of the specimens
Figure 6(a) to (c) depict the hysteretic load–displacement relationships of all specimens. The skeleton curves of the specimens were compared in Figure 6(d). The method of determining characteristic points is shown in Figure 7. “Ultimate load” is the load that equals to 0.85 Pmax at the descending part of the skeleton curve. The displacement corresponding to this load is called the ultimate displacement of the specimen. The yield load for the skeleton curve without evident yield point can be determined using the Park method (Park and Paulay, 1975; Figure 7).
Load–displacement hysteresis and skeleton curves of the specimens: (a) XJ, (b) ZP1, (c) ZP2, and (d) specimen skeleton curves.
Park R method.
As shown in Figure 6(a) to (c), the load–displacement curve was almost linear during the load control stage, thereby indicating that the beam performed in an elastic stage. The load–displacement curves presented a nonlinear behavior after yielding because of the plastic deformation of the beam. The slope of the load–displacement curve declined with the increase in displacement. After the beams reached the peak load, the applied loads gradually decreased with the spalling of the concrete.
The positive and negative hysteretic curves of ZP1 and ZP2 were basically symmetrical before reaching the peak load. This finding indicates that the straight threaded sleeve connection at the top of the beam and the high-strength screw connection at the bottom of the beam could transfer the load. The positive and negative peak loads are symmetrical, indicating that the combination of new and old concrete of the prefabricated beam had minimal influence on their positive and negative bending capacities. However, the load–displacement curves of ZP1 and ZP2 after the peak load were partially symmetrical during the loading process, because the residual deformation and damage under negative loading were larger than those under positive loading. After the specimens yielded, the hysteretic loops of the two cycles of the displacement control loading stage were different from those under the smaller slope of the second cycle. This finding indicates that the specimens experienced accumulative damage. The yield and ultimate loads of ZP1 and ZP2 were higher than those of XJ, indicating that the existence of steel fasteners can improve the bearing capacities of the specimens. At the displacement control stage, ZP1 and ZP2 weld joint fractures occurred in ribbed plate 1 and the baffle with the increase in displacement. The load suddenly dropped, and the test ended. Therefore, the welding quality of steel fasteners is important and should be extensively investigated.
Bearing capacity, deformability, and ductility
The characteristic loads and corresponding displacement are listed in Table 2. For each specimen, Fcr and Δcr are the cracking load and displacement; Fy and Δy are the yield load and displacement; Fmax and Δmax are the ultimate load and displacement; and Fu and Δu are the failure load and displacement, respectively. The positive cracking loads of ZP1 and ZP2 and the corresponding positive cracking displacements were larger than those of the negative cracking loads and displacements. The flexural moment of the upper concrete crack point was higher than that of the bottom concrete under the same load due to the existence of steel fasteners. Positive yield loads of ZP1 and ZP2 and their corresponding positive yield displacements were larger than their negative yield loads and displacements. This finding indicates that the section of flexural capacity at the steel fastener part at the bottom of the beam end was higher than that at the top. Compared with that of XJ, the positive and negative yield loads of ZP1 and ZP2 increased by 19.69% and 21.09%, whereas their negative yield loads increased by 8.03% and 2.66%, respectively. The yield displacements of ZP1 and ZP2 were larger than that of XJ, except for the positive yield displacement of ZP2. The positive ultimate bearing capacities of ZP1 and ZP2 increased by 23% compared with that of XJ, whereas the negative ones increased by 13% and 8%, respectively. This finding indicates that this new type of joint connection mode can effectively improve the bearing capacity of the specimens.
Ultimate displacement angle (
where H is the distance from the loading point to the center of the column. The minimum displacement angle at the failure point of all specimens was 1/40, thus satisfying the requirement of the limiting value, namely, 1/50 for the elastic–plastic story drift of a reinforced concrete frame structure under a rare earthquake, as specified in the Seismic Design Specification of Building Structures (GB50011–2010). This finding indicates that the new joints exhibit good post-yield deformability.
The ductility coefficient (
The positive and negative ductility coefficients of XJ were 4.67 and 4.97, respectively. The difference between the positive and negative ductility coefficients was due to the residual deformation after positive and negative loading. The positive ductility coefficients of ZP1 and ZP2 were greater than the negative ones due to the existence of steel fasteners and prestressing force at the bottom of the beam. The ductility coefficients of ZP1 under positive loading and ZP2 under positive and negative loadings were greater than 3. Moreover, the ductility coefficient of ZP1 under negative loading was 2.7, which exceeded the requirement for concrete structures. The displacement ductility coefficients of concrete structures should be greater than 2. Although the weld joint fracture of ribbed plate 1 and the baffle was caused by construction deficiency, the ductility could still meet the requirements.
Stiffness degeneration
Stiffness of the specimens can be expressed using secant stiffness under the cyclic loading and can be calculated as follows
where Fi is the peak load of the ith cycle, and Δi is displacement corresponding to the peak load of the ith cycle. Relation curves between the secant stiffness and Δi/Δy of each specimen are shown in Figure 8.
Stiffness degradation.
The stiffness degeneration trends of XJ, ZP1, and ZP2 were similar. Moreover, stiffness rapidly decreased before the yielding of the specimen. Then, the speed of stiffness degeneration slowed down. Results showed that the stiffness values of all specimens were basically identical. The initial stiffness values of ZP1 and ZP2 were greater than that of XJ due to the existence of steel fasteners and the application of prestressing force.
Energy dissipation
Energy dissipation capacity, which is an important indicator for the assessment of seismic structural performance, refers to the energy absorbed by the specimen under cyclic loading in which the energy dissipation coefficient E or equivalent viscous damping coefficient ξeq is generally used for evaluation. The plumper the hysteretic loop, the greater the E or ξeq value, and the stronger the energy-dissipating capacity of the specimen.
The equivalent viscous damping coefficient ξeq was used in this work to evaluate the energy-dissipating capacity of the joint, which can be calculated by equation (4)
where SABC, SCDA, SOBE, and SODF are the areas of regions ABC, CDA, OBE, and ODF shown in Figure 9, respectively.
Diagram of the calculation of energy dissipation coefficient ξeq.
The first hysteretic loop at each loading step was selected for computation. The equivalent viscous damping coefficient ξeq and accumulative energy dissipation of specimens at the peak load, as listed in Table 3, were extracted. The values of the equivalent viscous damping coefficients ξeq of both prefabricated beam-to-column joint specimens were 0.19 and 0.20. The ξeq of the cast-in-place specimen was approximately 0.1. Thus, the energy-dissipating capacities of ZP1 and ZP2 met the requirement, and the shape of the hole on the end plate under the peak load had little effect on the energy consumption.
Energy dissipation coefficients of specimens.
The equivalent viscous damping coefficient (ξeq) against the displacement curve of all specimens is shown in Figure 10(a). The energy dissipation of all specimens under the elastic state in the load control phase was poor. After entering the displacement control phase, the energy dissipation capacity became highly dependent on the hysteretic dissipation at the hinge (Mitoulis and Rodriguez, 2016), and the energy-dissipating capacities had been improved due to accumulated damage; however, the growth was slow. The equivalent viscous damping coefficient continuously increased, and all specimens retained their growth state. The energy dissipation coefficients of ZP1 and ZP2 were slightly lower than that of XJ because of the steel fasteners at the bottom part of the fixed beam ends of ZP1 and ZP2. After the loading process entered the plastic phase, the plastic damage at the beam-end plastic hinge area became smaller than that of XJ; thus, the energy-dissipating capacity of the joint was lower.
(a) Equivalent viscous damping coefficient curves and (b) accumulated energy dissipation–displacement relation curves of the specimens.
The cumulative energy dissipation (Ea) against displacement curve of all specimens is shown in Figure 10(b), where Ea is the sum of the area of each load–displacement hysteretic loop before the ith cycle. As shown in the figure, the accumulated energy-dissipating capacity difference between joints ZP1 and ZP2 with U-shaped and circular holes opened on the pre-embedded steel fasteners at the bottom was 4.7%. Thus, the energy-dissipating capacities were basically identical, thereby indicating that the influence of different hole shapes was minimal. The cumulative energy dissipation increased with the increase in displacement. The development trend of specimens ZP1 and ZP2 was similar to that of specimen XJ; the cumulative energy dissipation increased slowly under load control and rapidly after displacement control. Specimens ZP1 and ZP2 showed poorer accumulative energy dissipation compared with XJ. This performance could be attributed to the advance fracture of weld between ribbed plate 1 and baffle of ZP1 and ZP2.
Joint deformation
The shear deformation of the joint can be measured/evaluated using shear angle, and the calculation formulas are shown as follows
where Δ1, Δ2, Δ3, and Δ4 are the expanded or shortened forms of A–A′, B–B′, C–C′, and D–D′, respectively; a and b are the heights of the column and beam sections, respectively. Figure 11(a) shows the schematic of shear angle calculation of joints.
(a) Diagram of the shear angle calculation of joints, (b) load–shear angle hysteretic curves of the XJ specimen, (c) load–shear angle hysteretic curves of the ZP1 specimen, and (d) load–shear angle hysteretic curves of the ZP2 specimen.
On this basis, shear deformations were calculated, and load–shear angle curves are shown in Figure 11(b) to (d). The figures show that the shear deformations of all the joints were small, and the maximum shear angle was 0.002 rad. This finding indicates that the specimen loading process realized the requirements of “strong joint and weak member” and “strong column and weak beam.”
Conclusion
The seismic behavior of a new type of beam-to-column joint used in PC frames was studied. Two prefabricated joint specimens and a cast-in-place joint specimen were tested under cyclic loading. Seismic behavior, including ultimate bearing capacity, failure mode, energy dissipation, ductility, and strength degeneration, was analyzed. The following conclusions could be obtained on the basis of the test results:
All specimens experienced beam hinge failure. Diagonal cracks mostly appeared at the bottom parts of the ZP1 and ZP2 beams compared with the cast-in-place specimen. The distance between the plastic hinge and the side surface of the column and the length of plastic hinge were enlarged because of the steel fasteners, thereby enhancing the fixed beam end.
The bearing capacity and initial stiffness of prefabricated joints were higher than those of the cast-in-place joint. The trends of stiffness degeneration were similar to those of the cast-in-place joint.
The energy dissipation coefficients of ZP1 and ZP2 were lower than that of XJ because of the existence of steel fasteners at the bottom of the ZP1 and ZP2 beams. After the yielding of the specimens, damage in the plastic hinge area was smaller than that of XJ. Thus, the energy-dissipating capacity of the prefabricated joints was low.
The ductility coefficient of ZP1 under negative loading was 2.7 because of the weld joint fracture of ribbed plate 1 and the baffle. Moreover, the ductility coefficients of ZP1 and ZP2 were within 3.34–4.91. Therefore, the ductility requirements of concrete frame joints can be satisfied. The ultimate displacement angles of the two new types of PC frame beam-to-column joints were approximately 1/36, which were larger than the limiting value of 1/50 of the plastic interlayer displacement angle of a frame structure under a rare earthquake, as specified in the Chinese standard.
After the test, concrete on the steel fasteners of ZP1 and ZP2 were chipped away. No slippage was observed between the concrete and the horizontal plates of the steel fasteners used in the new type of joint. Therefore, straight thread sleeve and bolt connections, which feature reasonable structure and reliable force transfer, were used for this new type of joint. The mechanical behaviors of both prefabricated joints were similar, but the joint with U-shaped holes on the end plate was more convenient to create.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclose receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by the National Natural Science Foundation of China (Grant no. 51578369), the Tianjin Science and Technology Major Projects, China (Grant no. 17ZXCXSF00080), and the Development and Demonstration of High-performance Green Assembled Storage Buildings (2017-379). The authors wish to express their sincere gratitude to the sponsors.
