Application of Bayesian State Estimation in Real-time Loss Estimation of Instrumented Buildings

Abstract

The focus of this paper is real-time Bayesian state estimation using nonlinear models. A recently developed method, the particle filter, is studied that is based on Monte Carlo simulation. Unlike the well-known extended Kalman filter, it is applicable to highly nonlinear systems with non-Gaussian uncertainties. The particle filter is applied to a real-data case study: a 7-story hotel whose structural system consists of non-ductile reinforced-concrete moment frames, one of which was severely damaged during the 1994 Northridge earthquake. An identification model derived from a nonlinear finite-element model of the building previously developed at Caltech is proposed. The particle filter provides consistent state and parameter estimates, in contrast to the extended Kalman filter. Finally, recorded motions from the 1994 Northridge earthquake are used to illustrate how to do real-time performance evaluation by computing estimates of the repair costs and probability of component damage for the hotel.