基于高斯Copula贝叶斯模型的盾构下穿既有隧道施工风险的分析

Construction with Tunnel Boring Machines Passing Under Shield Tunnels Passing Below Existing Tunnels Based on Gaussian Copula Bayesian Network Model

为对盾构下穿既有隧道施工工程安全风险进行分析和管控,提出一种基于高斯Copula贝叶斯(GCBN)模型的盾构下穿既有隧道施工风险分析方法。基于故障树建立了一套包括12个因素的施工安全风险指标体系,将贝叶斯网络的动态推理诊断与Copula理论的依赖性表达相结合,在不确定和不完全信息下构建盾构下穿既有隧道施工风险分析的GCBN模型。以武汉轨道交通12号线下穿既有7号线工程为例,利用高斯Copula识别各因素边际分布类型,计算各因素间相关系数并连接网络中结点。通过模型推理进行定性和定量分析,识别盾构下穿施工安全风险状态,分析各致险因素对风险结果的影响。最后,对敏感性高的因素采取措施进行防控。通过防控前、后模型计算结果对比,实现盾构下穿施工过程实时动态安全预警管控。实践表明:GCBN模型预测结果与专家评价结果吻合,验证了所构建GCBN风险分析模型的可靠性。

In order To analyze,manage and control and manage the safety risks of construction with tunnel boring machines passing under existing tunnels,shield tunnel underpass construction project,a risk analysis method based on the Gaussian Copula Bayesian Network(GCBN)model was proposed.Based on the fault tree,a set of the construction safety risk index system including 12 factors was established,and the dynamic inferential diagnosis of Bayesian Network was combined with the dependence expression of Copula Theory.The GCBN model for risk analysis of construction with tunnel boring machines passing under existing tunnels was the shield underpass existing line is constructed under in the condition of uncertain and incomplete information.Taking the project of Wuhan Rail Transit Line 12 passing under underpass the existing Line 7 as an example,the Gaussian Copula was used to identify marginal distribution types of each factor,calculate correlation coefficients between each factor,and connect nodes in the network.Through the model reasoning,can undertake qualitative analysis and quantitative analysis were conducted,identification of shield the risk in the construction safety state under construction with tunnel boring machines passing under existing tunnels could be identified,the impact,the analysis to the risk factors on results of risks was analyzed.Eventually,preventive and control measures for high sensitivity factors were implemented.By comparison with the calculating results of models before and after being prevented and controlled,the real-time dynamic security warning and control for construction with tummel boring machines passing under existing tunnels were achieved.,choose high sensitivity factors take measures for prevention and control,and control of the model calculation results before and after contrast,wear under the shield tunnel underpass construction process real-time dynamic security early warning and control.The application results indicated that the prediction results by the GCBN model were consistent with the evaluation results by experts,which verified the reliability of the established risk analysis model of GCBN.and provides reference and support for solving the risk analysis problems of the shield underpass existing tunnel in similar projects.