Input and state estimation for earthquake-excited building structures using acceleration measurements

Abstract

Estimating both state and ground input for earthquake-excited building structures using a limited number of absolute acceleration measurements is critical to post-disaster damage assessment and structural evaluation. Input estimation in this case is particularly challenging due to the lack of direct feedthrough term, which renders the system weakly observable for its input. Hence, input estimation in this scenario is sensitive to modeling error and measurement noise. In this thesis, a two-step strategy is proposed to estimate both state (displacement and velocity) and ground input using a limited number of absolute acceleration measurements for building structures. First, the ground input is estimated by solving a least squares problem with Tikhonov regularization and Bayesian inference. In the second step, floor states are estimated using Kalman filter with input obtained from the first step, the least squares with Tikhonov regularization and Bayesian inference. The proposed strategy was numerically and experimentally evaluated based on shear-type building structures.