Uncertainty Quantification in the Assessment of Progressive Damage in a 7-Story Full-Scale Building Slice

Abstract

In this paper, Bayesian linear finite-element (FE) model updating is applied for uncertainty quantification (UQ) in the vibration-based damage assessment of a 7-story RC building slice. This structure was built and tested at full scale on the University of California at San Diego-Network for Earthquake Engineering Simulation shake table: progressive damage was induced by subjecting it to a set of historical earthquake ground motion records of increasing intensity. At each damage stage, modal characteristics, such as natural frequencies and mode shapes, were identified through low-amplitude vibration testing; these data are used in the Bayesian FE model updating scheme. To analyze the results of the Bayesian scheme and gain insight into the information contained in the data, a comprehensive uncertainty and resolution analysis is proposed and applied to the 7-story building test case. The Bayesian UQ approach and subsequent resolution analysis are shown to be effective in assessing uncertainty in FE model updating. Furthermore, it is demonstrated that the Bayesian FE model updating approach provides insight into the regularization of its often ill-posed deterministic counterpart.