Wang, Kong, Zhang, Chu and Chen 2017
The authors evaluated a self-centering modular panel with a slit steel plate shear wall (SCMP-SW) in order to test its seismic load resistance while also measuring its recentering capabilities and energy dissipation. This was accomplished with cyclic quasi-static loading on nine specimen of varying arrangement and dimension.
System Concept
Designed as a posttensioned steel moment resisting frame, an SCMP-SW combines slit walls, that provide primary energy dissipation and partial strength, with PT connections that give the frame its recentering capabilities. The SCMP was installed in the east bay of the main frame and bolted at both ends of the panel beams, with the slit wall either installed within the SCMP or the west bay. The bottom through beam was connected to the reaction blocks at five locations using high-strength bolts. All posttensioned strands were placed symmetrically about the centerline of the panel beam and anchored to the outer flanges of the panel columns using multi-strand barrel anchors. To protect against local failure from large compressive loads on the flanges after decompression, horizontal stiffeners in the panel columns were added. To prevent local buckling or damage due to the forces exerted by the panel columns, transverse stiffeners were installed. A boundary plate was welded to the panel beam, with the slit walls connected to the boundary plate with connection plates and high-strength bolts. To avoid vertical sliding of the panel column during rocking, filler plate was welded to the panel beam and extended to the outside of the end of the panel beam with a sloped cutout in the overhang segment. This filler plate also prevented the pounding between the outer flange of the panel column and the lower flange of the through beam during large displacements of the frame. Lateral supports were added to prevent the moment frame from moving out of plane while cover plates were welded to the panel-beam flanges near the panel beam ends to prevent displacement of the panel beam columns.
Analytical Study
The theoretical behavior of the SCMP-SW was predicted prior to testing in order to identify the ideal response of the SCMP-SW. This theoretical analysis revealed that the SCMP exhibits a bilinear elastic behavior with an initial stiffness equivalent to that of a frame having fully welded moment resisting connections. An ideal hysteretic model of the SWs was assumed to be an elastic-linear hardening backbone, due to the post yield stiffness ratio, with a small compressive strength upon load reversal. An idealized lateral force versus displacement response was a superposition of the individual responses of each component. The SCMP-SW had a stiffness equal to that of a conventional steel moment resisting frame combined with a slit wall, during the initial lateral sway. Following this the slit walls continued to resist additional load until yield, when the lateral stiffness decreased substantially. After unloading PT connections recompress with a recentering stiffness. The idealized PT connection response loads and unloads along a bilinear elastic moment-gap rotation curve, in which the initial stiffness is infinitely large.
Experimental Study, Results and Discussion
Nine, 1-story, 2-bay, full scale specimen of varying dimension and arrangement were tested in order to determine the best configuration for seismic resistance, energy dissipation and recentering. The four variants of the slit walls had different width-to-thickness ratios (b/t), as that is the controlling factor for the behavior of the slit walls. Loading was applied via a servohydraulic actuator at a height of 3,490 mm from the lower flange of the bottom through beam. The PT strands were tensioned to 38% of their ultimate strength prior to testing to provide sufficient self-centering capabilities while avoiding a PT strand yielding at 4% story drift. The target cyclic drift history for the test was a modified version of the loading protocol outlined in ANSI/AISC 341-10 with fewer cycles up to 1% story drift and the same number of cycles after.
The specimen with SCMP-SW exhibited the expected flag shaped hysteresis, with no observed strength degradation at 4% drift occurring in any of the specimen when the test was terminated. As indicated by the reduction in specimen stiffness, significant slit wall yielding was typically observed near 0.6% drift. Slit walls having a relatively large b/t ratio suffered pinches hysteresis when inner flexural links buckled out of plane and each link twisted in a similar manner to lateral-torsional buckling of beams. However, slit walls with a relatively small b/t ratio, experienced a transverse deformation mode of the whole panel. Believed to be caused by the severe plastic deformation and exacerbated by the twisting of flexural links, the slit walls saw fracture that was initiated at the end of the slits during large drifts.
The target value for the pretensioned force in SCMP1 and SCMP2 were 600kN and 800kN respectively; the actual results were lower due to the limitations of the equipment. While the measured PT force curves agree with the expected behavior, there were some signs of loss of prestress resulting from seating the PT strands anchorage during each cycle of loading. The short PT strand length in both SCMP1 and SCMP2 cause such a significant prestress loss that reduction coefficient should be considered. Theoretical decompression moments of the PT connection were overestimated, likely due to the uneven bearing surfaces at the panel-beam end and the panel-column flanges from construction tolerances.
Actual slit wall behavior observed in the experiment was similar to that of the idealized model. Confirmed by comparing hysteresis curves, the slit wall was able to resist load after yielding with a small postyielding stiffness. The compressive strength of the slit walls with varying b/t ratios differed, which may have affected the recentering capabilities after unloading.
All specimen were able to recenter after the 2% drift cycle and the 5% drift cycle with the exception of the positive loading of Specimen S1W29 and the negative loading of Specimen S2W21. Specimen with smaller b/t ratios of slit walls had larger residual drifts due to the additional lateral load resistance from the compressive strength of the slit walls during unloading. The theoretical recentering stiffness aligned well with the experimental results.
Energy dissipation was practically identical across SCMP1 and SCMP2 however, SCMP2 was able to provide a larger stiffness and strength while also giving a larger recentering ability than SCMP1. Hysteresis curves suggest that SCMP-SWs with smaller b/t ratios had larger stiffness, strength and energy dissipation while giving the system larger resistance to recentering during unloading by large residual drift. It can be concluded that a properly designed SCMP-SW can provide satisfactory seismic resistance performance, recentering capabilities and energy dissipation capacity. This also indicates that the SCMP would potentially reduce the postearthquake structural repair workload demand as damage was concentrated in replaceable fuse elements.
Reference
Wang, W., Kong, J., Zhang, Y., Chu, G., and Chen, Y. (2017). “Seismic behavior of self-centering modular panel with slit steel plate shear walls: Experimental testing.” Journal of Structural Engineering, 144(1), 04017179–1-04017179–13.