基于贝叶斯网络模型的工业生产安全事故风险评估  

Risk Assessment of Industrial Production Safety Accidents Based on Bayesian Network Model

作　　者：张城 王子云[1,2] 陈星百 向月 李久鹏 ZHANG Cheng;WANG Ziyun;CHEN Xingbai;XIANG Yue;LI Jiupeng

机构地区：[1]四川大学建筑与环境学院,四川成都610065 [2]宜宾四川大学产业技术研究院,四川宜宾644002 [3]重庆科技学院安全工程学院,重庆401331 [4]广州开发区投资控股有限公司,广东广州510770

出　　处：《煤气与热力》2023年第11期I0029-I0036,共8页Gas & Heat

基　　金：四川省重点研发项目(2021YFS0362)。

摘　　要：为了加强工业生产安全事故的预警,定量评价事故风险,以国内651起工业生产安全事故为研究对象,建立了工业生产安全事故蝴蝶结模型。根据贝叶斯网络工业生产安全事故风险评估模型预测事故计算概率。首先,计算各节点事件的条件概率,通过软件Genie3.0构建模型并计算,得到国内工业生产安全事故发生的计算概率为12.50%。其次,通过模型诊断得到导致事故发生的主要原因,生产或维修过程人为错误B1和工业生产过程故障B2是主要原因,其中违章操作C1是导致生产或维修过程中人为错误B1的主要原因,工艺条件不畅C5是导致工业生产过程故障B2的主要原因。In order to strengthen the early warning of industrial production safety accidents and quantitatively evaluate accidents risk,a bowtie model of industrial production safety accidents was established with 651 domestic industrial production safety accidents as the research object.The calculation probability of accidents is predicted based on the Bayesian network industrial production safety accident risk assessment model.First,the conditional probability of each node event is calculated,the model is constructed and calculated by Genie3.0 software,and the calculation probability of domestic industrial production safety accidents is 12.50%.Secondly,the main causes of accidents are obtained through model diagnosis.The results show that human errors B1 in production or maintenance process and malfunctions B2 in industrial production process are the main causes,among them improper operation C1 is the main cause of human errors B1 in the production or maintenance process,and poor process conditions C5 are the main cause of malfunctions B2 in the industrial production process.

关 键 词：工业生产安全事故 蝴蝶结模型 贝叶斯网络 概率预测 

分 类 号：TU996.9[建筑科学—供热、供燃气、通风及空调工程]