Filtering characteristics of isolation layer in base-isolated structures and shaking table test verification.

Base-isolated structures are extensively used in critical infrastructure and lifeline projects. Current response spectrum methodology for determining seismic actions has limitations in designing long-period base-isolated structures. Based on a two-degree-of-freedom (2DOF) model that accounts for the non-proportional damping characteristics of isolation bearings, this study analyzes the filtering characteristics of the isolation layers through spectral analysis of both original and attenuated ground motions. Theoretical and experimental results demonstrate that designing the superstructure using the "filtered response spectrum(FRS)" is more rational compared to conventional design response spectra(DRS). Shaking table tests under white noise and seismic excitations verify the filtering effect of isolation bearings: filtered acceleration time-history are reduced to 1/3-1/2 of the original peak values, with the FRS exhibiting decreased peaks, rightward shifts, and significantly shortened platform segments. The filtering principle, rather than period elongation alone, better explains the isolation layer's effectiveness, particularly for structures with inherently long natural periods.Current code-based design spectra inadequately capture the filtered seismic input, necessitating revisions to incorporate isolation-specific filtered spectra for rational and resource-efficient designs.