Navigation Risk Assessment of LNG Ships in Port Waters Based on Bayesian Networks

Due to the dangerous characteristics of LNG and the particularity of LNG ships, it is difficult to ensure the safety of ship navigation. Studying the navigation risks of LNG ships is beneficial for reducing the probability of accidents and various losses caused by accidents, which plays a crucial role in maritime traffic safety. In this paper, literature analysis, expert evaluation, Bayesian networks and other methods are used to study the navigation risks of LNG ships in port waters. Firstly, by studying the characteristics and risks of LNG and LNG ships, the risk factors are analyzed from four aspects as "human", "machine", "environment" and "management". Then a Bayesian network structure diagram of risk factors is established, and a conditional probability table is obtained by analyzing historical data and expert evaluation. The sensitivity analysis and posterior probability inference of risk factors are carried out. Finally, the risk factors with the highest sensitivity are identified, and corresponding countermeasures and suggestions are proposed for the risk factors that have a significant impact.