In this paper, we analyze metro cascading failure induced by load fluctuation and present a robust station capacity assignment method. A cascading failure model is proposed by considering system initial balance state, load matching, and capacity-based load redistribution. Based on random walks, an initial failure model is proposed to simulate different load fluctuations that represent unrealistic passenger flows for failure analysis. We study the actual case of Nanjing metro. The results show that small station tolerance and larger load fluctuation can lead to global network failure easily. The phase transition between network functional state and global failure is abrupt and first order-like, which is sensitive to station tolerance and load fluctuation strength. The coupling effects of station capacity, station tolerance, and load fluctuation strength on the network robustness are investigated and the optimal model parameters for robust station capacity assignment are obtained.

在本文中，我们分析了由负载波动引起的地铁级联故障，并提出了一种稳健的车站容量分配方法。考虑系统初始平衡状态、负载匹配和基于容量的负载再分配，提出了级联故障模型。基于随机游走，提出了一种初始故障模型来模拟不同的负载波动，这些波动代表了不切实际的客流，用于故障分析。我们研究南京地铁的实际案例。结果表明，小站容差和较大的负载波动很容易导致全局网络故障。网络功能状态和全局故障之间的相变是突然的和一阶的，对站点容限和负载波动强度敏感。站台容量、站台容差、