Spatiotemporal dynamic risk analysis of railway line equipment driven by data and experience

Safe railway line equipment operation is crucial for ensuring railway transportation safety. In this study, cloud models (CMs), Bayesian networks (BNs), and risk matrices are integrated to construct a risk assessment model for railway line equipment, enabling quantitative evaluation and classification of risk occurrence likelihoods. First, based on engineering practices, technical reports, accident cases, and expert knowledge, risk events related to railway line equipment and their corresponding risk factors are identified. Leveraging the grid method to consider the spatiotemporal dynamic characteristics of railway line equipment, the system is divided into grid units, and risk factor state data within each grid are dynamically collected across different time periods. Subsequently, drawing on expertise, CMs are applied to handle uncertainties in qualitative knowledge, while BNs are used to calculate the probability of risk events occurring in each grid unit over various time intervals. Finally, risk levels are determined by integrating likelihood and consequence levels through a risk matrix. The model was validated using a rail break case study, demonstrating its ability to accurately identify high-risk grid areas and times. These findings provide a scientific basis for railway authorities to develop refined and targeted risk mitigation strategies, thereby enhancing railway transportation safety and stability.