An event-based probabilistic model of disruption risk to urban metro networks

Abstract

Metro systems serve large populations and form extensive networks. Incidents such as signal failure, train failure, and power failure, pose great challenges to the reliable operation of metro systems around the world. For the Beijing metro system, incidents caused 408 disruptions of train services from 2014 to 2018. These incidents are investigated in detail, and a Monte Carlo approach and incident parameter functions are used to generate stochastically simulated incident events. Based on the simulated incidents, combined passenger flow data and an input-output model, we estimate the risk associated with the Beijing metro system in terms of disrupted passenger flows by considering risk propagation in the network, where both direct passenger loss and indirect passenger flow loss considering interchanges between different lines are considered. Lines at high risk are identified, and a sensitivity analysis is performed to investigate the effects of risk mitigation measures. This study provides a generic risk modeling method for urban metro systems and can improve decision making to manage metro system risk.