Reliability Analysis of Operational Metro Tunnel Based on a Dynamic Bayesian Copula Model

Abstract

This paper develops a hybrid Bayesian copula model for the reliability analysis of an operational metro tunnel in a dependent system in which risk factors interact with each other. The reliability analysis method used in present studies demonstrates two major deficiencies. First, the failure criteria of structural systems, which cannot describe the complex interactions among these failure criteria, are considered independent of one another. Second, a traditional reliability analysis is typically confined to static process reasoning, which cannot reflect the time-varying effects of adjacent environments. This study develops a binary discrete time-varying Bayesian copula model for precisely evaluating the structural reliability of operational tunnels to overcome the deficiencies. The Wuhan Metro tunnel, which was excavated by an earth pressure balance shield machine, is selected as a case study to investigate the applicability of this Bayesian copula model. Results show that the fitting effect is poorer in the Gaussian copula function than in the Clayton, Gumbel, and Frank copula functions. Convergence deformation (CD) and differential settlement (DS) values present non-Gaussian dependent structures, which indicates that traditional assumptions of the independence or Gaussian-dependent structures were not applicable to this case. Therefore, the Bayesian copula model can sensitively capture variations in the performance states of shield tunnel structures, thus verifying the accuracy and practicability of this model.