System Identification and Information Processing in Seismic Vulnerability Analysis of Structures

Seismic vulnerability analysis of structures generally involves processing many judgments which are based on experience and much data which is ill-defined or conjectural. A cascade model is presented for combining site hazard, design, architectural and building characteristics data that incor-porate both specific and vague information. The cascade model is Bayesian in approach in that new information is used to update prior estimates of vulnerability. The development of the seismic vulnerability evaluation model depends on rules encoded as a set of relations between linguistic variables (for example IF (antecedent 1, antecedent 2, . . .) THEN (conse-quence 1, consequence 2, . . .)). These statements are combined through a fuzzy set system identification model.