A posteriori optimal intensity measures for probabilistic seismic demand modeling

Abstract

A probabilistic seismic demand modeling approach with the optimal intensity measure (IM) parameters determined a posteriori (POS-PSDM) is employed to examine and compare the full potential of the explanatory power of different IM formulations. A total of six IM formulations adopting the optimal IM parameters determined a posteriori (i.e. POS-IMs) are studied, including two spectral IMs with the optimal period T*, two fractional order IMs with the optimal fractional order α*, as well as two spectral IMs with the optimal period T* and the optimal damping ratio ζ*. A comprehensive IM comparative study is conducted based on hysteretic single-degree-of-freedom systems, considering a wide range of structural parameters. The POS-IMs manifest substantially improved performance (i.e. efficiency and sufficiency) compared with their conventional counterparts, revealing the value of adopting this POS-PSDM approach to ensure the PSDM predictive performance. In particular, the spectral acceleration at the optimal period and damping ratio, Sa(T*,ζ*), which is introduced as an IM candidate for the first time, not only consistently demonstrates superior explanatory power but also exhibits fairly good hazard computability. The POS-PSDM approach in conjunction with Sa(T*,ζ*) exhibits good potential in further improving the accuracy and reliability of probabilistic seismic risk assessment with negligible increase in computation cost.