Resilience assessment of an urban rail transit network: A case study of Chengdu subway

Existing studies seldom consider network structure and passenger travel demand jointly, and certain impractical assumptions are generally considered for assessing the resilience of an urban rail transit (URT) network. To address the abovementioned limitations, we have proposed a performance indicator called the demand–impedance (DI) indicator, in which demand and impedance are reflected by passenger trips and travel time. By considering effective travel paths (ETPs) and passengers’ path choice behavior, we have proposed a node centrality called effective path betweenness (EPB) by modifying the betweenness centrality (BC) to evaluate the importance of stations. The performance curve of a URT network during the attack and repair processes is depicted using the DI indicator, and a modified resilience metric is formulated by referring to the resilience triangle. The model application in the Chengdu subway network demonstrates that the correlation coefficient between the EPB and BC of stations is 0.901, which indicates that stations with a higher EPB are inclined to have a higher BC. The Chengdu subway network demonstrates a higher resilience under random disturbances than it does under malicious disturbances. Disturbance duration, passengers’ tolerance time, and rescue ability on the Chengdu subway network significantly affect its resilience. Several practical suggestions involving the management of disturbances, shortening the emergency response time, providing passenger services, and improving emergency rescue ability are provided for managing the Chengdu subway system under disturbances.