Abstract
This paper examines the applicability or validity of the recently developed analytical approach to seismic reliability analysis, in which nonwhite random process was used for earthquake excitation and bi-linear and peak-oriented hysteretic models were used for the building structures. For this purpose, 1995 Kobe earthquake accelerograms and the associated damages in Kinki-district were used. From comparative discussions for the analytical reliability and the damage magnitude in selected eleven points in the district, it was found that the approach is effective for predicting earthquake damage provided pertinent random earthquake process and structural models are used for the analysis.
