基于数据驱动贝叶斯网络的桥梁运营事故风险分析  

Risk Analysis of Bridge Operation Accidents Based on Data-Driven Bayesian Network

作　　者：申建红 张静[1] 张宇馨 尹琪 SHEN Jianhong;ZHANG Jing;ZHANG Yuxin;YIN Qi(School of Management Engineering,Qingdao University of Technology,Qingdao 266525,China;Research Center for Smart City Construction Management,Qingdao University of Technology,Qingdao 266525,China)

机构地区：[1]青岛理工大学管理工程学院,山东青岛266525 [2]青岛理工大学智慧城市建设管理研究中心,山东青岛266525

出　　处：《沈阳大学学报（自然科学版）》2023年第6期535-543,550,共10页Journal of Shenyang University：Natural Science

基　　金：国家自然科学基金资助项目(71471094)。

摘　　要：为深入探究近年来我国南方地区桥梁运营事故频发的原因,提出了一种将故障树和贝叶斯网络方法相结合的可靠性风险分析方法。首先,采用故障树法识别提取出桥梁运营事故的各个风险因素,借助GENIE 2.0软件构建基于故障树的贝叶斯网络模型;其次,结合本文收集的我国2000—2021年的南方地区桥梁运营事故案例,从人为、自然和管理3个风险因素分析运营阶段事故的发生机理,利用贝叶斯网络的双向推理能力识别出最容易引发目标事件的关键风险因素和最大致因路径;最后,针对上述分析推理结果提出了风险防控策略,为南方地区桥梁运营阶段的安全作业提供参考。In order to explore the causes of frequent bridge operation accidents in southern China in recent years,a reliability risk analysis method combining fault tree and Bayesian network method was proposed.Firstly,the fault tree method was used to identify and extract each risk element of bridge operation accident,and the Bayesian network model based on fault tree was constructed with GENIE 2.0 software.Secondly,combined with the collected bridge operation accident cases in southern China from 2000 to 2021,the occurrence mechanism of the accident in operation stage was analyzed from the aspects of human,natural and management risk factors,and the bidirectional reasoning ability of Bayesian network was used to identify the key risk factors and the most approximate cause path that were most likely to cause the target event.Finally,based on the above analysis and reasoning results,the risk prevention and control strategies were proposed to provide a reference for the safe operation of Bridges in southern China.

关 键 词：贝叶斯网络 数据驱动 桥梁运营 事故原因 风险分析 

分 类 号：U447[建筑科学—桥梁与隧道工程]