基于云贝叶斯网络盾构下穿铁路近接施工安全风险动态评价

Dynamic Evaluation of Construction Safety Risk for Approaching Excavation Near Connection of Shield Tunnel Undercrossing High-Speed Railway Under Shield Based on Cloud Bayesian Network

提出云模型和贝叶斯网络相结合的安全风险动态评价方法,用于盾构下穿铁路近接施工复杂系统的安全研究。基于文献和大量工程经验建立盾构下穿铁路施工风险评价指标体系和评价标准。通过云模型实现连续节点状态的离散化,获取先验概率。构建云贝叶斯网络推理模型预测盾构下穿铁路施工安全风险的聚合概率分布,并生成综合云获取定量风险值,有效实现风险定性概念与定量数值大小的转化。结合敏感性分析和风险诊断,实现事前、事中、事后阶段的实时安全动态管控。以某下穿铁路工程为例,利用该方法有效评估了系统风险水平,并结合反向诊断确定重点管控因素,验证了该方法为实现盾构近接施工安全管理预警决策提供有力支持。

In this paper,A dynamic safety risk evaluation method combining cloud model and Bayesian network was proposed to study the safety of complex approaching excavation of shield tunnel undercrossing railway close connection construction.Based on literature and a lot of engineering experience,the risk evaluation index system and evaluation standard of shield undercrossing railway construction were established.The state of continuous nodes was discretized through cloud model and the prior probability was obtained.The cloud bayesian network reasoning model was established to predict the aggregation probability distribution of the construction safety risk of shield tunneling railway,and the comprehensive cloud was generated to obtain quantitative risk values,which effectively realized the transformation of qualitative concept of risk and quantitative values.Combined with sensitivity analysis and risk diagnosis,real-time security dynamic control was realized in the stage before,during and after the event.Taking an undercrossing underpass railway project as an example,the method was used to effectively evaluate the system risk level,and the key control factors were determined by reverse diagnosis.It was verified that the method could provide a strong support for the safety management and early warning decision of shield close construction.