According to the “Specification for Seismic Test of Buildings”⁴⁴, definitions of relevant mechanical characteristics can be illustrated as follows. The cracking load (F_cr) and cracking displacement (Δ_cr) refer to the load and deformation at the onset of cracking, respectively. Referring to the corresponding parameters of steel bars shown in Table 3, the specimen was deemed to have yielded when the maximum strain of the longitudinal steel bars reached 2210 µε, or the specimen showed a sudden increase in deformation rate, the corresponding load and deformation were defined as yield load (F_y) and yield displacement (Δ_y) severally. The peak load (F_p) and peak displacement (Δ_p) are the maximum load and corresponding displacement during testing. Following the peak load, the specimen’s horizontal resistance decreased with increasing displacement and loading. Generally, the specimen was considered to have reached its peak state when the horizontal resistance was reduced to 85% of the peak load or when visible damage occurred. The load and displacement at this moment were defined as ultimate load (F_u) and ultimate displacement (Δ_u) severally.

Damage process and failure mechanisms

Observations for the XJQ1 specimen

During the force-controlled phase, each load level was applied with a single cycle of reciprocation. Initial cracking at the bottom of the shear wall was observed when the horizontal load reached 450 kN. As the load continued to increase, multiple horizontal cracks began to form along the height of the shear wall, with the original cracks also extending. Upon reaching a horizontal load of 550 kN, the outermost reinforcement in the confined boundary zone yielded, accompanied by a significant increase in horizontal displacement at the top of the wall. This was identified as the yield point for specimen XJQ1, at which point the loading mode was switched to displacement control.

When the horizontal displacement reached 2Δ_a, cracks started to develop diagonally, and the width of the original cracks increased significantly. During the loading process of 3Δ_a cycles, there were no new cracks occurred, whereas the original cracks developed and extended dramatically, with the primary diagonal crack nearly penetrating the entire wall. Additionally, slight spalling of the concrete at the bottom of the shear wall was observed. By the time the displacement reached 4Δ_a, the load carrying capacity of the specimen had reached its peak, and significant crushing and spalling of the concrete occurred. Due to the extensive damage sustained by the specimen, the test was terminated at this stage.

Observations for ZPQ1

The horizontal cracks occurred initially at the PC-CIP interface (the junction between the precast concrete shear wall and the post-cast concrete) when the horizontal force reached 250 kN. As the loading process continued, these cracks propagated significantly, extending both within the precast wall panel and the post-cast concrete zones. During the 2Δ_a loading cycles, when the loading displacement was applied to the maximum values in the positive and negative directions respectively, two diagonal cracks developed sequentially, penetrating the wall panel and intersecting with each other. In the duration of the 3Δ_a loading cycles, the load carrying capacity of the specimen reached its peak value by the time the loading displacement attached 3Δ_a in the first cycle. At this point, the interface crack width extended dramatically, and noticeable crushing of the bottom concrete occurred. When the loading displacement reached 3Δ_a in the second cycle, the concrete at both ends of the post-cast zone experienced severe crushing and spalling, leading to the termination of the test.

Observations for ZPQ2

When the horizontal load reached 275 kN, initial cracks formed at the PC-CIP interface. As the load increased, these interface cracks continued to develop and eventually penetrated the interface at a horizontal load of 550 kN, corresponding to the yield load. With the gradual increase in loading displacement, several horizontal cracks appeared along the height of the shear wall, while the original cracks extended further. During the 2Δa loading cycles, two diagonal cracks formed sequentially, penetrating the wall panel and intersecting with each other, indicating clear signs of shear damage in the wall. When the loading displacement reached 3Δa, the specimen reached its peak load carrying capacity, accompanied by a dramatic widening of the interface cracks. Moreover, numerous diagonal cracks emerged in the post-cast zone, the cracks at both ends of the post-cast zone were densely distributed company with significant crushing and spalling of concrete, the test was terminated at this stage.

Observations for ZPQ3

Horizontal cracks first appeared at the PC-CIP interface when the horizontal force reached 225 kN. The specimen yielded as the load approached 550 kN, at which point a horizontal crack formed in the PC wall panel approximately 200 mm above the interface between the PC and CIP concrete. After yielding, cracks developed gradually, with two diagonal cracks penetrating the wall panel and intersecting with each other during the 2Δa loading cycles. During the 3Δa loading cycles, the specimen reached its peak load carrying capacity. Although several short diagonal cracks emerged in the post-cast zone, the crack development did not cause significant damage, with only minor concrete spalling observed.

Observations for ZPQ4

Due to construction defects encountered during the casting of the post-cast concrete, horizontal cracks first appeared at the PC-CIP interface when the horizontal force reached 200 kN, which is lower compared to other specimens. The interface cracks continued to develop and eventually penetrated through the interface when the horizontal load reached 450 kN. The specimen yielded as the load approached 550 kN. During the 2Δa loading cycles, when the displacement was applied to its maximum in the positive direction, significant crushing of the concrete on the right side of the CIP zone was observed. Subsequently, asymmetric damage occurred, with the specimen exhibiting severe damage under positive loads and only minor damage under negative loads. This indicated notable deterioration of the concrete under reversed cyclic loading. When the 3Δ_a loading displacement was conducted, the load carrying capacity attained its peak value, and the longitudinal bars were bulged and exposed with the serious crushing of the concrete at the bottom of the shear wall.

Propagation and distribution of cracks

The schematic diagram of specimen cracks is presented in Fig. 9. The crack distribution diagram reveals that the crack development in the PC specimens was comparable to that in the CIP specimen, which indicated that the damage morphology of the PC specimens closely mirrored that of the CIP specimen. Both exhibited characteristics of compression-bending failure, which was marked by the buckling of tensile steel bars and the crushing of concrete at the bottom of the beam.

Crack propagation details and typical damages of specimens.

A comparison of crack development and failure modes between the PC specimens and the CIP specimen shows that, except the premature occurrence of cracks at the PC-CIP interface in the PC specimens, the concrete also experienced early crushing and spalling. This premature damage may be caused by the low strength of the concrete in the CIP zone, which failed to meet design requirements and thus could not provide adequate tensile strength and bond strength with the steel bars, impairing effective stress transfer. When comparing ZPQ1/ZPQ2 with ZPQ3/ZPQ4, it was observed that ZPQ1/ZPQ2 exhibited slightly more severe concrete crushing and spalling. In contrast, ZPQ3 and ZPQ4 showed only minor concrete spalling, which may be due to the presence of sleeves weakening the thickness of the protective layer. This indicated that the force transmission efficiency of steel bars connected by extruded sleeves is superior to that of the overlapping U-bar loop connections in the confined boundary zone.

Hysteretic loops and skeleton curves

As illustrated in Fig. 10(a)−(e), the hysteretic curves under positive loading are noticeably fuller compared to those under negative loading during the displacement loading phase. This discrepancy may be attributed to two factors: Firstly, an installation gap between the lateral loading device and the specimen likely led to asymmetric loading displacements applied to the specimen. Secondly, the loosening of the anchored threaded bars that secured the specimens to the laboratory floor contributed to the asymmetric horizontal movement of the specimen under lateral loading. The hysteretic properties of all specimens were relatively similar. Before yielding, the specimens exhibited nearly constant stiffness, with horizontal displacement increasing approximately linearly with lateral force. After yielding, the stiffness of specimens decreased rapidly with the increase of the horizontal displacement. The CIP specimen displayed a slight pinching effect, while the hysteresis curves of the PC specimens were fuller and did not exhibit a pronounced pinching effect.

Hysteretic loops and skeleton curves of specimens.

The skeleton curves of the five specimens, as compared in Fig. 10(f), show that the curves for ZPQ3 and ZPQ4 closely matched that of XJQ1 prior to the peak cycle. In contrast, the skeleton curves for ZPQ1 and ZPQ2 exhibited significantly lower stiffness compared to XJQ1. Moreover, the load carrying capacity of the PC specimens was lower than that of the CIP specimen, likely due to the insufficient strength of the post-cast concrete, which did not meet the design requirements.

Load carrying capacity, deformation and ductility

The resistance and deformation of all specimens at different character loading steps are shown in Table 5. The yield load of PC specimens was very close to that of the CIP specimen. However, the obvious difference lay in the fact that the cracking load and peak load of PC specimens were significantly lower than those of the CIP specimen. The difference in cracking load may be caused by the presence of the PC-CIP interface weakening the integrity of the PC shear wall, while the difference in peak load may be due to the actual strength of the post-cast concrete being much lower than the design value, which cannot provide sufficient compressive bearing capacity, resulting in premature collapse of the concrete at the bottom of the shear wall. In terms of deformation, except for the yield displacements of ZPQ1 and ZPQ2 were significantly larger than those of the CIP specimen, while the yield displacements of ZPQ3 and ZPQ4 were much closer to those of the CIP specimens, the characteristic displacements of PC specimens in the cracking cycle and peak cycle were significantly lower than those of the CIP specimen. Moreover, considering that specimens had already been severely damaged when the peak load were reached and could not continue to withstand the load to obtain ultimate displacement, the ratio of the peak displacement Δ_p to the yield displacement Δ_y was defined as the ductility coefficient Δ_p/Δ_y to characterize the deformation capacity of specimens. It can be seen from Table 5 clearly, that the ductility of PC specimens was significantly smaller than that of the CIP specimen, but the ductility coefficients of ZPQ3 and ZPQ4 were closer to those of the CIP specimen.

Equivalent stiffness

Formula 7 defines the equivalent stiffness of each step as K_i under low reversed cyclic loading. In this formula, F_i represents the peak load of the step, Δ_i denotes the peak displacement of the step, and “+” and “-” indicate positive and negative directions, respectively. As illustrated in Table 6, the equivalent stiffness values of ZPQ3 and ZPQ4 were larger than those of the CIP specimen, while the stiffness of ZPQ1 and ZPQ2 was significantly lower than that of the CIP specimen. Moreover, by fitting these stiffness data of the five specimens, Fig. 11 demonstrates the variation pattern of equivalent stiffness with increasing displacement for the five specimens. It was evident that all specimens exhibited a similar degradation trend as displacement increased.

Stiffness degradation curves of specimens.

Energy dissipation performance

The energy dissipation properties of all specimens under low reversed cyclic loading are illustrated in Figs. 12 and 13. The energy dissipated per cycle is represented by the area enclosed within each corresponding load-displacement hysteretic loop. As shown in Fig. 12, ZPQ3 and ZPQ4 exhibited slightly lower energy dissipation than the CIP specimen, while ZPQ1 and ZPQ2 demonstrated the lowest energy dissipation overall. However, the trend in energy dissipation across the PC specimens was consistent with that of the CIP specimen, with ZPQ3 and ZPQ4 displaying a dissipation rate nearly equivalent to the CIP specimen.

Energy dissipation per cycle for specimens.

Evolution of normalized energy dissipation.

In order to analyze the evolution of energy dissipation, the normalized energy dissipation w/Wwas introduced for further examination⁴⁶. Figure 13 presents the normalized energy dissipation per cycle and its evolution trend derived by fitting the data across all samples. It is evident that the energy dissipation trend of the CIP specimen aligned closely with that of the PC specimens. Before yielding, the energy dissipation was minimal but increased approximately linearly with displacement. After yielding, the energy dissipation rate surged, contributing significantly to the total energy dissipation. As the peak phase approached, the energy dissipation rate gradually stabilized.

Rebar strain analysis

Strain gauges were mainly employed to measure the strain in vertical steel bars at the bottom of the shear wall, providing a reference for determining whether the specimen had reached the yield state. Additionally, the effectiveness of different connection methods in transmitting force could be assessed by examining the strain development in the corresponding steel bars of the PC specimens. The layout of the strain gauges is illustrated in Fig. 7. At the connection position of vertical steel bars in the confined boundary zone, if the strain development forms of the corresponding steel bars connected by extruded sleeve or overlapping U-bar were similar to each other, and the maximum strain is approximate, it would indicate that the connection method is reliable and capable of effectively transmitting stress. As depicted in Fig. 14, the force-strain curves of all specimens exhibit an approximate linear distribution before yielding, with the curve progression of the PC specimens closely resembling that of the CIP specimen. Moreover, the shape of the force-strain curves of the corresponding steel bars at the connection position in the PC specimens was consistent across the different specimens. It should be noted that although the force-strain curves of ZPQ1 displayed notable differences from the other specimens that may caused by the damage of the strain gauges or installation defects, the force-strain shape of the corresponding steel bars in ZPQ1 remained similar. These findings indicated that both connection methods are reasonable and reliable.

Force-strain curves of vertical steel bars in the confined boundary zone.