摘要
为预防深水井控事故,保障深水钻井安全高效运行,运用贝叶斯网络识别深水井控主要风险诱因。首先以事故树模型为基础,从一级井控和二级井控2个阶段构建深水井控贝叶斯网络模型;然后在井喷失控情景下,利用贝叶斯网络的逆向分析和共因失效分析能力,求解各风险诱因的后验概率;最后按诱因发生概率和对整个系统的影响,识别深水井控主要风险诱因。分析结果表明,深水井控共包含10个主要风险诱因,且后验概率趋势与先验概率趋势基本一致,计算结果符合客观实际。
In order to prevent deepwater well control accidents and guarantee deepwater drilling under safe and efficient conditions,Bayesian network was used to identify principal risk factors in deepwater well control. Firstly,based on fault tree model,a Bayesian network was built from primary and secondary well control phases. Then reverse analysis and common cause failure capacities of Bayesian network were utilized to calculate the posterior probabilities of risk factors for a blowout scenario. Finally,principal risk factors were identified,according their probabilities and influences. The results show that there are10 principal risk factors in deepwater well control,that the posterior probabilities follow the same trend as prior probabilities,and that the calculation results conform with the reality.
出处
《中国安全科学学报》
CAS
CSCD
北大核心
2015年第5期157-163,共7页
China Safety Science Journal
基金
国家重点基础研究发展("973")计划项目(2015CB251206)
关键词
深水井控
贝叶斯网络
风险识别
主要风险诱因
事故树
deepwater well control
risk identification
Bayesian network
principal risk factors
fault tree
