论文部分内容阅读
火灾概率分析是海洋平台火灾定量风险评估的重要组成部分,考虑到传统事故树和事件树方法存在一定的局限性,提出了基于逻辑树和贝叶斯网络的火灾概率分析模型。首先采用数理统计方法对墨西哥湾地区2 837起火灾事故进行统计分析,依据事故情况构建逻辑树,然后将逻辑树转化为贝叶斯网络,根据历史数据确定贝叶斯网络各节点的先验概率和条件概率。结果表明:海洋平台火灾事故是设备、人因和组织管理多因素耦合作用的结果;基于贝叶斯网络模型得到海洋平台火灾概率约为1.0×10~(-5),为海洋平台火灾定量风险评估提供了基础数据;由贝叶斯网络模型分析得出,人因操作失误与缺乏作业安全分析的后验概率分别达到0.471和0.119,表明人因组织因素对海洋平台火灾事故具有重要影响。
Fire probability analysis is an important part of quantitative risk assessment of offshore platform fire. Considering the limitations of traditional accident tree and event tree method, a fire probability analysis model based on logical tree and Bayesian network is proposed. First of all, mathematical statistics were used to conduct statistical analysis of 2 837 fire accidents in the Gulf of Mexico. The logic tree was constructed according to the accident situation, then the logic tree was transformed into a Bayesian network, and the prior probability of each node in the Bayesian network was determined according to historical data And conditional probability. The results show that the fire disaster of offshore platform is the result of multi-factor coupling of equipment, human factors and organizational management. The probability of fire on offshore platform is about 1.0 × 10 -5 based on Bayesian network model, which is the quantitative risk of offshore platform fire The analysis provided the basic data. According to the Bayesian network model, the posterior probabilities of human error and lack of operational safety analysis reached 0.471 and 0.119 respectively, indicating that human factors have an important impact on the offshore platform fire accidents.