基于UHBs的输气管道高后果区泄漏燃爆事故分析

为探究输气管道高后果区中人的不安全行为（Unsafe Human Behaviors,UHBs）对输气管道泄漏燃爆事故发生的影响，结合模糊Bow-tie模型和贝叶斯网络对输气管道泄漏燃爆事故进行分析。构建基于T-S模糊故障树的输气管道泄漏燃爆模糊Bow-tie模型，并转化为贝叶斯网络；从人的不安全行为发生的可能性出发，将不同等级高后果区划分为不同等级人口敏感区；利用专家经验评判法得到不同等级人口敏感区基本事件的先验概率和中间事件的条件概率表；运用贝叶斯网络双向推理算法求解模糊Bow-tie模型。结果表明:随着地区人口敏感等级的提高，输气管道泄漏燃爆事故发生的概率随之增大，发现导致输气管道失效泄漏事故发生的最主要原因为施工破坏，失效原因与EGIG分析的结果基本相符，验证该方法在高后果区输气管道泄漏燃爆事故分析上的可行性，可为输气管道高后果区的安全管理提供决策依据。

Analysis on leakage and explosion accidents in high consequence areas of gas pipeline based on UHBs

In order to explore the influence of unsafe human behaviors (UHBs) in the high consequence areas of gas pipelines on the leakage and explosion accidents of gas pipelines,the fuzzy Bow-tie model and Bayesian network were combined to analyze the leakage and explosion accidents of gas pipeline.A fuzzy Bow-tie model of the leakage and explosion accidents of gas pipeline based on T-S fuzzy fault tree was established and transformed into a Bayesian network.According to the possibility of UHBs,the high consequence areas with different levels were divided into the population sensitive areas with different levels,then the prior probabilities of basic events and the conditional probabilities of intermediate events for the population sensitive areas with different levels were determined by using the expert experience evaluation method,and the fuzzy Bow-tie model was solved by the bidirectional inference algorithm of Bayesian network.The results showed that as the sensitivity level of the population in the area increased,the probability of the leakage and explosion accidents of gas pipeline increased.It was found that the most important cause of the leakage and explosion accidents of gas pipeline was the construction damage,and the failure causes were basically consistent with the results of EGIG analysis,which verified the feasibility of this method in the analysis of the leakage and explosion accidents of gas pipeline in the high consequence areas.It can provide decision-making basis for the safety management of high consequence areas of gas pipelines.