基于灰色模糊贝叶斯网络的隧道坍塌风险评估方法

Risk assessment method of tunnel collapse based on grey fuzzy Bayesian network

为了准确评估隧道施工期间的坍塌风险,采取及时有效的控制措施,针对公路隧道施工的复杂性和风险状态的模糊不确定性,提出一种灰色模糊贝叶斯网络的隧道坍塌风险评估方法。该方法将灰色系统理论引入风险状态分析中,用包含区间灰数的梯形模糊隶属函数表征节点风险状态,用区间灰数描述风险因素与坍塌风险之间逻辑关系的不确定性。构建隧道坍塌风险特征参量规划模型,采用全局优化算法计算特征参量的最终取值范围。基于区间灰数运算规则和贝叶斯网络推理,计算出风险因素或风险事件处于不同状态的概率和节点重要度,提升了贝叶斯网络解决不确定问题的能力。工程实例应用结果表明,该方法能够准确评估隧道坍塌的可能性,可作为隧道施工安全风险分析的决策工具。

In order to assess accurately the collapse risk during tunnel construction and take timely and effective control measures,a grey fuzzy Bayesian network for tunnel collapse risk assessment method was proposed for the complexity of highway tunnel construction and the fuzzy uncertainty of risk state.The method introduced gray system theory into risk state analysis,characterized the node risk state with a fuzzy trapezoidal membership function containing interval gray numbers,and described the uncertainty of logical relationship between risk factors and collapse risk with interval gray numbers.A planning model of tunnel collapse risk characteristic parameters was constructed,and a global optimization algorithm was used to calculate the final value range of characteristic parameters.Based on the interval gray number operation rules and Bayesian network inference,the probability of the risk factors or risk events in different states and the node importance were calculated.The ability of Bayesian network was improved to solve uncertainty problems.The application of engineering example shows that the method can accurately assess the possibility of tunnel collapse and can be used as a decision tool for tunnel construction safety risk analysis.