Analysis Method for Passenger Congestion Propagation and Network Resilience Changes of Urban Rail Transit under Failures Based on the Improved Susceptible Exposed Infected Recovered Model

Abstract

Operational failures in urban rail transit (URT) systems can cause congestion propagation in passenger flow, and the networked operation mode makes the propagation characteristics more complex. Most studies have focused on the state of stations under congestion propagation or examined the issue from a network perspective, but ignored the differences in station spacing, section running times, and the remaining train capacity among different initial failure stations. Few researchers have considered the bidirectional propagation characteristics along the line and integrated congestion propagation with network resilience. An improved susceptible exposed infected recovered (SEIR) congestion propagation model was proposed, which considers the bidirectional propagation in both forward and backward directions. Specifically, the model quantified congestion propagation rates integrating train speed, station capacity, and section passenger flow. The differences in congestion propagation between workdays and holidays were discussed using the Nanjing URT network as a case study. Network resilience under two operation adjustment schemes was compared based on the transfer station failure propagation scenario. The congestion propagation rate and propagation range at transfer stations vary between different lines, and can propagate to up to four adjacent stations along one line. Compared with the scheme of adding standby trains, the scheme of the overtaking lines can improve the resilience performance and resilience index of unweighted network by 5.2% and 5.3%, and improve by 7.9% and 6.7% in weighted network.