安全与环境工程2025，Vol.32Issue(5)：158-169,12.DOI:10.13578/j.cnki.issn.1671-1556.20240548
山区交汇水域船舶航行风险分析与致因推理
Risk analysis and causal reasoning of vessel navigation in mountainous confluence waters
摘要
Abstract
To enhance the safety of inland waterway vessels,it is essential to investigate the causal mechanisms behind navigation risks in mountainous confluence waters.The study first established a risk indicator system based on accident data from the confluence areas in the upper Yangtze River(2007-2021)using grey relational analysis.A data-driven tree-augmented naive Bayesian network(TAN-BN)was then constructed.Relationships between various factors and"accident types"were analyzed using mutual information and joint probability theory,with the model's reliability verified.Finally,the TAN-BN was subsequently employed to identify the most probable causes of different accident types and their key influencing factors.A method for quantifying the degree of influence between factors was proposed,and the effects of environmental and management factors on human and vessel risks were examined.The results show that traffic density is the factor most strongly linked to"accident types"in mountainous confluence waters;significant water level fluctuations increase the likelihood of crew errors in emergency response;the effective implementation of safety management by shipping companies significantly reduces human and vessel risk probabilities,with a particularly strong effect on human factor risk control.The research findings can provide some reference for ship enterprises and maritime authorities to enhance the safety management of ship navigation in mountainous confluence waters.关键词
船舶航行风险/致因推理/树状贝叶斯网络(TAN-BN)/山区交汇水域Key words
risk of vessel navigation/causal reasoning/tree-augmented naive Bayesian network(TAN-BN)/mountainous confluence water分类
资源环境引用本文复制引用
张丹,王平安,盛进路,盛小曼,康莉..山区交汇水域船舶航行风险分析与致因推理[J].安全与环境工程,2025,32(5):158-169,12.基金项目
教育部产学合作协同育人项目(202102104008) （202102104008）
重庆市研究生导师团队建设项目(JDDSTD2022009) （JDDSTD2022009）
