基于贝叶斯网络的城市爆炸事故风险概率研究

Study on risk probability of urban explosion accidents based on Bayesian network

针对频发的各类城市爆炸事故,采用事故树和贝叶斯网络相结合的综合风险评价方法构建概率模型,在分类溯源的基础上进行风险的概率测算、原因诊断和敏感节点分析。首先,按爆炸源属性将研究对象划分为粉尘爆炸、可燃性气体或液体爆炸、可燃性固体爆炸和物理爆炸四类;其次,运用专家信心权重法评估细化的28个根节点事件的先验概率,结合条件概率表计算顶事件发生的风险概率;再次,通过逆向推理法分析风险的形成原因。结果表明:在城市爆炸事故中可燃性气体或液体爆炸发生的概率最大,人员监管不到位、防爆知识欠缺、违规操作生产等是城市爆炸事故的关键风险因素和敏感节点;最后,根据研究结论提出风险管控预防措施,以期减少城市爆炸事故的发生。

A probability model of city explosion accidents is constructed using the comprehensive risk assessment method combining fault tree and Bayesian network.The risk probability calculation,cause diagnosis,and sensitive node analysis are carried out based on classification and traceability.Firstly,according to the attribute of explosion sources,the research objects are divided into four categories:dust explosion,combustible gas or liquid explosion,combustible solid explosion,and physical explosion.Secondly,the expert confidence weight method is used to evaluate the a priori probability of 28 root node events and then combined with the conditional probability table to calculate the risk probability of top events.Thirdly,the causes of risk are analyzed by reverse reasoning.The results show that the probability of combustible gas or liquid explosion is the highest in urban explosion accidents.Inadequate supervision of personnel,lack of explosion-proof knowledge,and illegal operation and production are the key risk factors and sensitive nodes of common urban explosion accidents.Finally,according to the research conclusions,risk control and preventive measures are put forward to reduce urban explosion accidents.