Safety Assessment Method of High-speed Rail Interval Structure Based on Weighted Bayesian Network.

To carry out digital and intelligent research on structural safety assessment methods for rail transport infrastructure, which has an essential impact on the maintenance and management of rail transport infrastructure. This study selects 22 risk indicators, which are then used in a comprehensive evaluation of the operational context of the structure, which aims to develop a holistic risk assessment index system encompassing the interactions between structure, vehicle, and environment. Using a dynamic weight model, we compute node weights and construct a weighted Bayesian network model. This approach addresses the limitations of the conditional independence assumption typical in standard Bayesian network models., thereby addressing the restrictions associated with the conditional independence assumption inherent in the Bayesian network model. The utilization of trapezoidal fuzzy numbers and the maximum noise algorithm to compute network parameters, as well as in the determination of the safety probability of a given structure and the identification of the most probable risk factors associated with it. The findings reveal that the security level of this structure, as determined by a standard Bayesian network, is classified as level I. In contrast, the Level II security status obtained from the weighted Bayesian network corroborates with the fuzzy analytic hierarchy method's results, indicating a significant enhancement in inference accuracy with the weighted approach. [ABSTRACT FROM AUTHOR]