Temporal Evolution-Based Risk Assessment for Fault Correlations in Catenary Systems

As the primary power source for high-speed railways, understanding the fault propagation mechanisms among the components of the catenary system is crucial for developing proactive maintenance strategies. To quantitatively characterize fault propagation risk among catenary components, this paper proposes a temporal evolution-based fault correlation risk assessment model from a data-driven perspective. First, fault correlations are defined based on temporal evolution, and a risk assessment model integrating credibility and discredibility is developed based on the MYCIN model to avoid misclassifying high-frequency but low-risk fault correlations as high-risk. Second, to address potential fault correlations that are not explicitly observed in historical data, a latent path-based risk inference graph is constructed to indirectly infer their risk levels through observable fault correlations. Third, to reveal the temporal evolution characteristics of fault propagation risk, a temporal evolution series of system risk coefficients after component faults is constructed, and risk decay node and risk stabilization node are defined. A case study based on real historical fault data collected from the operation and maintenance system of a high-speed railway catenary system validates the effectiveness of the proposed method. The results demonstrate that the framework can comprehensively assess both observable and potential fault correlation risks and capture their temporal evolution characteristics, providing quantitative support for time-targeted proactive maintenance strategies of catenary systems.