Abstract: The coupling of risk factors between risks due to propagation effects and interdependence can lead to serious accidents. To analyze the fire risk coupling of ship carrying vehicles, this paper proposes a novel quantitative analysis method that integrates the N-K model with the Bayesian Network (BN) model. Firstly, through the statistical analysis of historical accidents, the fire causes of ship carrying vehicles accidents are investigated, and the types of human, mechanical, electrical, environmental factors and risk coupling are clarified. Second, the N-K model is constructed to study the coupling relationship between factors and the risk coupling mechanism. Finally, the BN is constructed and verified based on the above results. Sensitivity analysis and backward inference analysis are carried out. The results indicate that the coupling causing the highest probability of fire is “human-mechanical” with a probability of 14.795%; the coupling causing the lowest probability of fire is “human-mechanical-electrical-environmental” with a probability of 1.354%, yet it has the highest risk coupling value, approximately 0.43. It is of great importance to focus on the risk factors related to human and machinery to prevent the co-existence of the two or more, with the goal of reducing the occurrence of fire accidents. The results of this study will not only provide guidance for improving the fire safety of ship-carrying vehicles, but also provide an important reference for risk assessment and firefighting.

Key words: fire of ship carrying vehicles, risk coupling analysis, BN model, N-K model