As a central driver of global trade, the safety of navigation on ocean-going voyages holds paramount importance. This study improved dynamic Bayesian network model for comprehensive ocean-going ship risk assessment. Firstly, addressing the risk posed by harsh natural environments in ship navigation, information nodes from the three dimensions were selected, and a dynamic Bayesian network was incorporated to capture the dynamic information. Second, for the uncertainty problem of unstructured data, the dynamic Bayesian network was improved to effectively integrate multi-source information through the hesitant cloud model. Subsequently, a multi-level risk assessment framework was constructed to achieve a refined assessment for the risk changes under different human behaviors, ship vulnerability and navigational environment conditions. An empirical study of container ship accidents in the North Pacific Ocean verifies that the experimental model effectively captures dynamic information, enabling a more accurate determination of depression locations and consequently achieving a more precise navigation risk assessment. Furthermore, the model is capable of capturing seasonal variations in marine environmental risks within the study area. Consequently, the improved model furnishes a scientific foundation for devising targeted risk response strategies.
Read full abstract7-days of FREE Audio papers, translation & more with Prime
7-days of FREE Prime access