基于HCNAM-BN的化工过程爆炸事故致因链分析  

Causation chain analysis on explosion accident of chemical process based on HCNAM-BN

作　　者：田震[1] 宋博达 

机构地区：[1]华南理工大学安全科学与工程研究所,广东广州510640

出　　处：《中国安全生产科学技术》2017年第5期145-150,共6页Journal of Safety Science and Technology

基　　金：2016年广东省安全生产专项资金项目(2016-125)

摘　　要：为了从源头上预防化工过程爆炸事故,依据风险耦合理论,探讨了各风险因子非线性耦合演化为爆炸事故的机理,构建了层次耦合网络分析模型(HCNAM);从多因素风险耦合角度分析了国内外44起典型化工过程爆炸事故,统计了各风险因子之间的耦合概率并进行了耦合致因重要度分级;采用耦合概率与二态分布相结合的条件概率分布,将层次耦合网络分析模型转化为贝叶斯网络,并对氯乙烯单体槽爆炸性混合气体爆炸事故进行了应用研究。结果表明:91种双因子耦合风险状态中,47种呈现弱耦合致因特性;7种因子双耦合形成风险的概率较大;基于HCNAM-BN模型分析事故,可有效辨识事故最可能致因因素,获取各事故致因链的发生概率并确定事故网络关键节点。In order to prevent the occurrence of explosion accidents in chemical process from the source, the mechanism of nonlinear coupling between risk factors and the evolvement to explosion accident were discussed based on the risk coupling theory, and a hierarchical coupling network analytical model （HCNAM） was constructed. 44 typical explosion accidents of chemical process at home and abroad were analyzed from the view of multiple factors risk coupling, then the statistics on the coupling possibility among the factors was carried out, and the importance degree of coupling causation was graded. The HC- NAM was converted into the Bayesian network by using the conditional probability distribution which combined with coupling possibility and two - states distribution, and the application study on the explosion accident of explosive gas mixtures in vinyl chloride monomer tank was conducted. The results showed that among the 91 types of coupling risk status with two factors, 47 types presented the characteristic of weak coupling causation. 7 factors with double coupling had higher probability to form risk. The analysis on the accident based on the HCNAM - BN model can effectively identify the most likely causative factors, obtain the occurrence probability of each accident causation chain, and determine the key nodes of the accident network.

关 键 词：化工过程 爆炸 事故致因链 风险耦合 贝叶斯网络(BN) 

分 类 号：X937[环境科学与工程—安全科学]