基于概率神经网络和层次分析法的硐室群施工风险评估  

Construction Risk Assessment of Cavern Group Based on Probabilistic Neural Network and Analytic Hierarchy Process

作　　者：宗志栓 张逸飞 林作忠[3] 陈晨 杨航 邱泽刚 申玉生[2] ZONG Zhishuan;ZHANG Yifei;LIN Zuozhong;CHEN Chen;YANG Hang;QIU Zegang;SHEN Yusheng(Urban Rail Transit Engineering Co.,Ltd.,China Railway 15 Bureau Group Co.,Ltd.,Guangzhou 510540,China;School of Civil Engineering,Southwest Jiaotong University,Chengdu 610031,China;China Railway Siyuan Survey and Design Group Co.,Ltd.,Wuhan 430063,China)

机构地区：[1]中铁十五局集团城市轨道交通工程有限公司,广州510540 [2]西南交通大学土木工程学院,成都610031 [3]中铁第四勘察设计院集团有限公司,武汉430063

出　　处：《铁道标准设计》2024年第3期177-185,共9页Railway Standard Design

基　　金：国家重点研发计划项目(2019YFC0605104);中铁第四勘察设计院集团有限公司科研课题(2022K005,2022K006)。

摘　　要：地下硐室群施工风险研究尚处于起步阶段,为快速准确评判风险因素,预防施工安全事故,利用概率神经网络(PNN)和层次分析法(AHP)建立风险评估模型,并研发硐室群施工风险评估软件。通过统计分析硐室群施工风险因素,设置工程地质、自然、设计施工和管理4个一级风险因素,23个风险控制指标,建立针对硐室群施工的风险指标体系。收集典型样本数据后,基于PNN对施工风险等级进行评判,同时采用AHP定量分析风险因素权重,迅速捕捉风险点,采取风险控制措施并优化施工方案。运用研发软件对重庆轨道交通18号线歇台子站硐室群施工进行风险评价,得到风险概率等级为Ⅳ,在施工过程中需要重点监测和控制地下水、围岩等级和支护及时性等带来的影响,实例评价结果与现场情况相吻合,验证了该评估软件的有效性和实用性。研究表明:针对硐室群施工建立的指标体系和评估方法能有效预测风险级别,实时指导施工过程,确保地下硐室群施工安全。The research on construction risk of underground cavern group is still in infancy.In order to quickly assess the risk factors and prevent construction accidents of tunnels,the construction risk assessment software is created based on probabilistic neural network(PNN)and analytic hierarchy process(AHP)to establish risk assessment model.Through statistical analysis of risk factors and references to relevant documents,a risk indicator system for cavern groups construction is established.There are four first-class factors which include engineering geology,nature,design and management and 23 risk control indexes.Firstly typical sample data are collected,then construction risk grade is evaluated based on PNN and the weights of risk factors are analyzed by AHP to quickly capture the risk points,take risk control measures and optimize the construction scheme.The developed software is used to evaluate the risk of Xietaizi station cavern group project of Chongqing rail transit line 18,and the risk level is set to beⅣ.The impact of groundwater,surrounding rock grade and support timeliness need to be monitored and controlled during the construction process.And the evaluation results are consistent with the field situation,which verifies the effectiveness and practicability of the evaluation software.The research shows that the index system and evaluation method established for the construction of underground caverns can effectively predict the risk level,guide the construction process in real time,and ensure the construction safety of underground caverns.

关 键 词：硐室群 概率神经网络 层次分析法 风险评价 软件开发 

分 类 号：TV554[水利工程—水利水电工程] TP183[自动化与计算机技术—控制理论与控制工程]