The traditional risk analysis methods of hydrogen sulfide leakage cannot carry out dynamic analysis on the develop-ment process of accident well, which causes the results with deviation from the actual situation. Based on Bayesian method, a Bow-tie model of hydrogen sulfide leakage during the well drilling process in the high temperature, high pressure and high sour gas ("three-high") gas field was established, and it was proposed to transform the model into Bayesian network. The occurrence probabilities of basic events will be updated in case of occurred accident, then by assuming the condition that the cumulative occurrence number of accident consequence within a certain time period is known, the occurrence probabilities of safety barriers and accident consequence will be updated, thus the dynamic risk analysis on hydrogen sulfide leakage can be completed. The results showed that the method overcomes the shortcomings of traditional static quantitative analysis methods well. It can dynamically assess the occurrence probabilities of basic events leading to hydrogen sulfide leakage and the impact extent on the occurrence of top event, and dynamically reflect the risk change of safety barriers and accident consequence, so as to provide reference for risk analysis and prevention measures of hydrogen sulfide leakage during well drilling process.%传统的H2 S泄漏风险分析方法不能很好地对事故发展过程进行动态分析，导致分析结果偏离实际。基于贝叶斯方法，构建了高温、高压、高含硫(“三高”)气田钻井过程中H2 S泄漏的蝴蝶结模型并提出将其转化为贝叶斯网络，在事故已发生的情况下更新基本事件发生的概率。然后，假定事故后果在确定的时间段内发生的累积次数已知的条件下，更新安全屏障及事故后果发生的概率，从而完成对H2 S泄漏的动态风险分析。结果表明，该方法克服了传统静态定量分析方法中的不足，可动态评估导致H2 S泄漏的基本事件发生的概率和对顶事件发生的影响程度，并动态反映安全屏障和事故后果的风险变化，能为钻井过程中H2 S泄漏的风险分析及防控措施提供参考。
展开▼
