Model cascading overload failure and dynamic vulnerability analysis of facility network of metro station

The existing studies mainly focused on the cascading failure of the whole metro but rarely on stations. In this paper, we map a metro station into a directed and hierarchical network, and propose a cascading overload failure model to study the network dynamic vulnerability. The passenger behavior and node load update are considered based on impedance function and mutual flow redistribution. Without changing the network topology, a flow optimization strategy is proposed to improve the network performance against cascading failure. Yuantong station of Nanjing Metro is studied. The results show the network is vulnerable and global failure occurs easily under small node tolerance. When the node tolerance excesses a critical value, the network can transit abruptly from global failure to balance state. We obtain that the critical node tolerances are 0.17 for the largest degree-based attack(LDA) and 0.19 for the largest load-based attack(LLA). The ranges of the optimal flow redistribution coefficient are identified by the flow optimization strategy, which are [0.15 0.2] for LDA and [0.2 0.25] for LLA. The result indicates that a proper guidance of passenger flow can be beneficial to the network performance while a strong intervention to passenger flow can instead increase the cascading failure risk.