基于贝叶斯网络的运营隧道渗漏水病害风险评估

RISK ASSESSMENT OF THE OPERATIONAL TUNNEL LEAKAGE BASED ON BAYESIAN NETWORKS

为了实现对隧道工程运营阶段渗漏水病害的定量化研究,解决传统致因分析与诊断方法在复杂系统不确定性处理方面的不足,在全面分析运营阶段隧道渗漏水病害成因机理的基础上,提出一种基于贝叶斯网络(Bayesian Networks,BN)的隧道渗漏水灾害风险定量评估与实时决策方法。利用BN的正向推理技术定量预测其渗漏水灾害发生概率,运用BN反向推理技术进行动态循环诊断,实时判定渗漏水病害形成原因,并将其运用到武汉某已竣工运营的地铁隧道工程中,结果论证了该模型与方法的科学性和可行性。

In order to realize the quantitative study on the tunnel leakage water in operation phase and solve the deficiencies of traditional cause analysis and diagnosis methods in complex system uncertainty processing,a quantitative evaluation and real-time decision-making method based on Bayesian networks is proposed after overall analysis of the formation mechanism of the tunnel leakage water in operation phase.Forward causal reasoning is used to predict the probability of the tunnel leakage water under different conditions.Dynamic cycle diagnosis is took to determine the leakage water reason in real time with reverse diagnosis reasoning.This method was applied to a completed metro tunnel in Wuhan,China.The results demonstrated that this method is scientific and feasible.