Risk Dynamic Regulation: Time-Varying Analysis of Risk Coupling in Hazardous Materials Road Transportation

Abstract The road transportation system of hazardous materials is a complex, giant system. The comprehensive effect of various risk factors significantly affects the safety of transportation and leads to accidents. To address this issue, this study proposes a hybrid model that combines N - K analysis and the dynamic Bayesian network (DBN) method. This model quantitatively analyzes the coupling effect of risk factors by integrating the coupling degree values from N - K analysis and the occurrence probabilities inferred by DBN. First, define the hazardous materials road transportation system and identify the accident risk factors and coupling types. Second, a DBN structure is established based on coupling types, and the node parameters are determined using diverse databases and existing knowledge. Additionally, the coupling degrees of different risk coupling types are obtained from the accident data. Finally, the temporal characteristics of different risk coupling types are revealed by using risk level indicators, and risk prevention strategies are proposed. Results indicate that the occurrence probability of risk coupling nodes gradually increases over time, and driver behavior and environmental factors belong to a very strong coupling type. This hybrid model effectively captures the dynamic risk evolution, providing a comprehensive approach for understanding and mitigating transportation risks.