改进贝叶斯网络模型在起重作业人机交互差错风险分析中的应用  

An improved Bayesian Network model for human⁃machine interaction error risk analysis in lifting operations

作　　者：晋良海[1,2,3] 闫月蓉 陈颖 邵波 陈述[1] 陈云[1] JIN Lianghai;YAN Yuerong;CHEN Ying;SHAO Bo;CHEN Shu;CHEN Yun(College of Hydraulic&Environmental Engineering,China Three Gorges University,Yichang 443002,Hubei,China;Safety Production Standardization Evaluation Center of China Three Gorges University,Yichang 443002,Hubei,China;Hubei Anhuan Technology Co.,Ltd.,Yichang 443002,Hubei,China)

机构地区：[1]三峡大学水利与环境学院,湖北宜昌443002 [2]三峡大学安全生产标准化评审中心,湖北宜昌443002 [3]湖北安环科技有限公司,湖北宜昌443002

出　　处：《安全与环境学报》2024年第1期213-220,共8页Journal of Safety and Environment

基　　金：国家自然科学基金面上项目(52179136)。

摘　　要：为量化分析起重作业人机交互差错风险,根据安全工效学原理及安全技术规范将起重作业人、机、环相关影响因素作为根节点,按照事故致因层次关联关系确定子节点,构建起重作业人机交互差错的3层级贝叶斯网络模型(Bayesian Network, BN);基于模糊集理论,采用认知可靠性与失误分析方法(Cognitive Reliability and Error Analysis Method, CREAM),厘定贝叶斯网络父节点失效概率以及中间节点条件概率;利用逆向推理仿真技术分析起重作业人机交互差错发生的因果链,探究起重伤害事故发生的人机交互差错风险。结果表明:起重作业人机交互差错最可能致因链为起重设备安全检查不到位→管理人员失误→人员操作失误→起重伤害事故发生;单因素失效条件下,起重作业人机交互差错风险概率呈线性增长趋势;在多因素失效条件下,一级节点因素失效概率愈大则人机交互差错效应愈显著,且呈现非线性增长态势。To quantify the risk of human⁃machine interaction errors in lifting operations,the Cognitive Reliability and Error Analysis Method⁃Bayesian Network(CREAM BN)method is used for inference analysis.First,according to the principles of safety ergonomics and the Technical Code for Safety of Lifting in Construction(JGJ 276—2012),the factors related to the man,machine,and environment for lifting operations as root nodes.And then we determined the child nodes according to the accident⁃causing hierarchical correlations and constructed a three⁃level BN model of human⁃machine interaction errors in lifting operations.Secondly,based on fuzzy set theory,the CREAM was used to determine the probability of failure of the parent node and the conditional probability of the intermediate nodes of the Bayesian network.Finally,the inverse reasoning simulation technique was used to solve for the posterior probability of each node in the event of a human⁃machine interaction error in the lifting operation.The results were analyzed to obtain the causal chain of human⁃machine interaction errors in lifting operations.To analyze the trend of the influence of the change in the value of each factor on the human⁃machine interaction errors in lifting operations,a sensitivity analysis of the probability of the nodes at all levels in the Bayesian network was carried out and the key nodes of the probability change of the Bayesian network were identified to explore the risk of human⁃machine interaction errors in lifting injuries.The results show that the most likely cause chain of human⁃machine interaction errors in lifting operations is"lifting equipment safety inspection is not in place→managerial error→personnel operation error→lifting injury accident".Under the condition of single⁃factor failure,the probability of human⁃computer interaction error in lifting operation shows a linear growth trend.Under the multi⁃factor failure condition,the greater the probability of failure of the first node factor,the more signif

关 键 词：安全工程 起重作业 人机交互差错 贝叶斯网络(BN) 认知可靠性与失误分析方法(CREAM) 

分 类 号：X943[环境科学与工程—安全科学]