煤炭工程 ›› 2020, Vol. 52 ›› Issue (4): 130-136.doi: 10.11799/ce202004025

• 研究探讨 • 上一篇    下一篇

基于事故树和贝叶斯网络的隧道塌陷风险概率估计方法研究

李泽荃,孙景来,陈磊   

  1. 1. 应急管理部华北科技学院
    2. 煤炭科学研究总院
  • 收稿日期:2020-01-16 修回日期:2020-03-24 出版日期:2020-04-20 发布日期:2020-07-22
  • 通讯作者: 孙景来 E-mail:sunjinglai83@163.com

Risk assessment method combined by Fault Tree Analysis and Bayesian Network and its engineering application in the tunnel collapse

  • Received:2020-01-16 Revised:2020-03-24 Online:2020-04-20 Published:2020-07-22

摘要: 为解决隧道风险评估中存在的主观性大、结果不准确的问题,提出基于事故树和贝叶斯网络的区间概率等级、权重信心指标与置信区间相结合的综合风险概率估计法,并将其应用于渔寮隧道的坍塌风险评估中。首先利用事故树构建贝叶斯网络,并利用案例中因素之间的依赖关系得出节点的条件概率(联合概率)。然后利用提出的区间概率等级划分与权重信心指标法进行调查,得出专家j对于基本事件xi出现概率的估计值Pij,由所有专家的估计值构建样本空间Ui及其统计量,通过引入置信区间的方法得出基本事件xi出现的概率范围。获得所有事件的概率范围后与贝叶斯网络的条件概率相结合进行风险推断,保证了风险评估的科学性和准确性,同时可利用该模型进行事故原因诊断。

关键词: 事故树, 贝叶斯网络, 隧道塌陷, 模糊数学, 条件概率

Abstract: To solve the problem of subjectivity and inaccurate result in the risk assessment, this paper presents a method which is applied in Yu Liao Tunnel based on Fault Tree Analysis(FTA) and Bayesian Network(BN. This risk probability assessment method includes three methods, such as the probability intervals, the weight and confidence index and confidence interval. First, FTA was used to construct the BN, and the dependent relationships between the factors in the cases were used to calculate the Conditional Probability Table (CPT) of every node. Second, the professor’s investigations were carried out to get the survey data which is based on the small probability interval hierarchies and the weight and confidence index method. The survey data gained from professor j about factor will be presented by . All the survey data will be used to estate the sample space and the statistic, introducing the confidence interval to get the probable range of factor . After finished all the data processing, the probable range and CPT will be used to calculate the risk, it also can be used to diagnosis the accident.

中图分类号: