Traffic flow-oriented reliability assessment and enhancement of urban transportation networks

System reliability is a critical aspect of urban transport networks. It ensures that modern cities can provide safe, efficient, and resilient mobility services. However, existing reliability studies primarily focus on network topology and connectivity, overlooking the impact of disruptions on passenger travel quality and the temporal heterogeneity inherent in traffic patterns. Here, we propose a traffic flow-oriented reliability analysis method. This method integrates real-world passenger flow data into a percolation-based reliability evaluation. Specifically, we construct a weighted network model that accounts for commuting time costs and transfer penalties. We also introduce traffic-aware centralities to identify critical links and propose reliability metrics that consider both traffic flow preservation and travel time variation in the event of disruptions. We further evaluate two protection strategies to enhance system reliability. We apply our approach to the Shanghai metro system and conduct extensive numerical simulations across different traffic patterns. Our results show that failing to consider travel quality can lead to an underestimation of network vulnerability. We also demonstrate that a centrality-based protection strategy improves the effectiveness and repeatability of protected links under similar traffic patterns. This study offers a data-driven, temporally adaptive methodology for evaluating and enhancing the reliability of urban transportation systems, providing insights for infrastructure planning and risk management in smart cities.