Bayesian-based seismic resilience assessment for high-rise buildings with the uncertainty in various variables

Abstract

This paper presents a probabilistic methodology that can consider the uncertainty coming from various random variables to assess seismic resilience for high-rise buildings. A 42-story composite steel frame-RC core tube building is used to show the proposed method. Uncertainties associated with the input loads, annual discount rate, cost ratio, recovery time and unknown parameters (or coefficients) in the demand model for fragility analysis are taken into consideration. Firstly, the Bayesian updating rule is applied to determine the posterior distribution of unknown parameters in the demand model. Then, the epistemic uncertainty of unknown model parameters is considered according to the full probability theory, which is the main novelty of this study. Additionally, the record-to-record method is adopted to consider the uncertainty associated with input seismic ground motions. Furthermore, the Monte-Carlo simulation is used to generate samples of the annual discount rate, cost ratio and recovery time accounting for the impacts of their uncertainties on the structural functionality loss and seismic resilience. It can be concluded that the epistemic uncertainty associated with unknown parameters in the demand model has significant impacts on the fragility estimates and seismic resilience. This study provides an effective way to incorporate the uncertainty of various random variables into the seismic resilience assessment, which is helpful for the decision-makers to implement appropriate mitigation strategies.