基于STPA-BN的船舶航行人为风险因素分析与评估

Analysis and Assessment of Human Risk Factors in Ship Navigation Based on STPA-BN

人为因素是引发船舶事故的最主要因素之一,为了研究船舶人为风险因素的因果关系,从中国海事局发布的船舶事故报告出发,引入系统理论过程分析-贝叶斯网络(STPA-BN)模型对船舶航行人为风险因素进行分析和评估。采用系统理论过程分析(STPA)方法识别出船舶航行中存在的不安全控制行为,结合事故报告内容提取出12种人为风险因素,利用风险因素的内在因果关系和结构学习功能构建贝叶斯网络拓扑结构;将事故报告量化,并对网络进行参数学习,对模型进行验证。在此基础上,利用贝叶斯网络(BN)的推理功能得到船舶航行中7种突出的人为风险因素和3条事故核心致因链,为保障船舶安全航行与船员培训提供数据支持。

Human factors are the most important factors causing marine accidents.In order to investigate the cause and effect of human risk factors,the"STPA-BN"model is introduced based on the ship accident report issued by the China Maritime Safety Administration.First,the STPA method is used to identify the unsafe control behavior in ship navigation,and twelve human risk factors are extracted based on the accident report.In the Bayesian visualization software Netica,taking human risk factors as nodes,using the internal causal relationship of risk factors and structural learning function to build the Bayesian network topology.Then the Bayesian network model of human risk factors in ship navigation after the quantification of accident report and the network parameter learning is obtained.After verifying the accuracy of the model,seven prominent human risk factors and three core cause chains of accidents by using the inference function of Bayesian network has been obtained.