基于BN的模糊系统事故风险管理辅助分析  

Study on the Assistant Analysis for Accident Risk Management Based on Fuzzy System and BN

作　　者：易玉枚[1] 廖可兵[1] 易灿南[1] 

机构地区：[1]湖南工学院安全与环境工程学院,湖南衡阳421002

出　　处：《安全与环境工程》2015年第3期105-110,共6页Safety and Environmental Engineering

基　　金：教育部安全工程第一类特色专业建设点项目(TS12328);湖南工学院安全技术及工程重点建设学科项目(HXK11001)

摘　　要：为提高模糊系统事故定量风险管理水平,将三角模糊数与贝叶斯网络相结合,提出基于BN的模糊系统事故风险管理辅助分析方法。首先通过致险机理分析确定风险事件各致险因子,在此基础上构建贝叶斯网络(BN),致险因子的先验概率用三角模糊数进行处理;然后利用BN的推理技术对风险事件进行概率预测、致因诊断和致险因子重要度分析;最后以某化工企业储油罐体火灾爆炸事故为例,应用该方法进行实证研究。结果表明:该方法能定量预测风险事件发生概率以界定风险等级,致因诊断能快速查明最可能的致因因素及其组合,重要度分析能找出系统薄弱环节以明确风险管理控制要点,从而为模糊系统事故风险管理提供辅助决策支持。For the purpose of improving the quantitative accident risk management level of fuzzy system, this paper proposes an assistant analysis method of accident risk management based on the combination of trian- gular fuzzy number and Bayesian Networks（BN）. In the process of the analysis,the method determines the risk factors based on the risk mechanism analysis for the risk event, establishes the Bayesian Networks, processes the prior probability of the risk factors with triangular fuzzy number, inferences the probability of risk event, diagnoses the causes and analyzes the importance degree of risk factors. Then the paper takes the assistant analysis of a fire and explosion accident of an oil tank in a chemical enterprise as example to research the application process of the method in detail. The results show that the method can predict the risk probability to define the risk level and prevent the accident early, find out the most possible causal fac- tors and combinations quickly through cause diagnosis, and find the weak points through importance degree analysis to determine the control points for risk management. So the method provides an assistant decision support for the accident risk management based on fuzzy system.

关 键 词：模糊系统 贝叶斯网络(BN) 三角模糊数 事故风险管理 辅助分析 

分 类 号：X913[环境科学与工程—安全科学]