主管:中华人民共和国应急管理部
主办:应急管理部天津消防研究所
ISSN 1009-0029  CN 12-1311/TU

消防科学与技术 ›› 2022, Vol. 41 ›› Issue (5): 634-638.

• • 上一篇    下一篇

基于EWM-AHP模糊贝叶斯网络的储罐区风险分析

彭玉杰1,董绍华1,谢书懿2,段宇航1   

  1. (1.中国石油大学(北京) 安全与海洋工程学院,北京 102249;2.中国石油集团工程材料研究院 石油管材及装备材料服役行为与结构安全国家重点实验室,陕西 西安 710077)
  • 出版日期:2022-05-15 发布日期:2022-05-15
  • 作者简介:彭玉杰(1998-),女,中国石油大学(北京)安全与海洋工程学院硕士研究生,主要从事油气设备完整性管理、安全科学与工程方面的研究,北京市昌平区府学路18号,102249。

Risk analysis of tank farm based on EWM-AHP and fuzzy Bayesian network

PENG Yu-jie1, DONG Shao-hua1, XIE Shu-yi2, DUAN Yu-hang1   

  1. (1. College of Safety and Ocean Engineering, China University of Petroleum in Beijing, Beijing 102249, China; 2. State Key Laboratory for Performance and Structure Safety of Petroleum Tubular Goods and Equipment Materials, CNPC Tubular Goods Research Institute, Shaanxi Xi'an 710077, China)
  • Online:2022-05-15 Published:2022-05-15

摘要: 储罐区作为储库中的重大危险源,风险因素复杂繁多,有效的风险分析对于储罐区的安全运行至关重要。结合储罐区实际情况,用熵权法对层次分析所得各危险因素的风险权重值进行了修正,增加了权重数据的科学性,形成了多层次风险分析指标体系。针对储罐区数据来源少,各风险因素相关性复杂的特性,通过模糊分析进行风险因素先验概率的计算,构建贝叶斯网络实现了事故类型及概率的动态预测。该方法能很好地实现具有数据来源少、事故后果严重等事故特性的储罐区风险分析。

关键词: 熵权法, 层次分析法, 模糊分析, 贝叶斯网络, 储罐区, 风险分析

Abstract: Abstract: As a major hazard source in the reservoir, the tank farm has complex risk factors. Effective risk analysis is crucial for the safe operation of tank farm. Combined with the actual situation of the tank farm, the entropy weight method is used to modify the risk weight value of the risk factors obtained from the hierarchical analysis, which increases the scientificity of weight data and forms a multi-level risk analysis index system. In view of the characteristics of few data sources and complex correlation of risk factors in tank farm, the prior probability of risk factors was calculated by fuzzy analysis, and the dynamic prediction of accident types and probabilities was realized by constructing Bayesian network. This method can realize the risk analysis of tank farm with few data sources and serious accident consequences.

Key words: Key words: entropy weight method, hierarchical analysis, fuzzy analysis, Bayesian network, tank farm, risk analysis