Bayesian Updating of Vulnerability Assessment for a Shield Tunnel Under Adjacent Excavation Disturbances

Vulnerability is a fundamental component of risk and its understanding is essential for evaluating infrastructure reliability and mitigating risks associated with their damages.The work focuses on refining the disturbance vulnerability framework established for specific geological formations, and enhancing its adaptability to transfer to other geological environments. This endeavor contributes to avoiding substantial computational costs. A novel methodology is developed that utilizes field measurements to update vulnerability model of shield tunnels. It incorporates the numerical simulation-based vulnerability analysis and a surrogate modeling technique into the Bayesian updating framework, where statistical characteristics of relevant design parameters are accordingly considered.Anillustrative scenariois herein employed to illustrate and verify the proposed framework. Several indicators for excavation disturbances are herein provided, specifically emphasizing the horizontal displacement of tunnels as a measure of their performance when confronted with nearby foundation excavations. Understanding such vulnerability is essential for evaluating infrastructure reliability and mitigating risks associated with their damages. To characterize the uncertainty of disturbance intensities, the lateral deflection of diaphragm walls is employed as an intermediate variable to indirectly condition tunnel displacement. The present work achieves a significant update of the excavation-induced vulnerability model based on a local damage database, facilitating probabilistic assessment of damage states in existing tunnels caused by deep foundation excavations across diverse geological environments.

Keywords: Bayesian updating, Excavation disturbances, Vulnerability assessment, Surrogate model.