动态贝叶斯网络在管廊电缆舱火灾风险评估中的应用研究  

Application study of dynamic Bayesian network in fire risk assessment of pipe gallery cable cabin

作　　者：陈雍君[1] 关鸿浩 李晓健 CHEN Yongjun;GUAN Honghao;LI Xiaojian(School of Urban Economics and Management,Beijing University of Civil Engineering and Architecture,Beijing 102616,China)

机构地区：[1]北京建筑大学城市经济与管理学院,北京102616

出　　处：《安全与环境学报》2025年第1期11-20,共10页Journal of Safety and Environment

摘　　要：电缆舱是城市综合管廊必不可少的舱体,担负着至关重要的电力传输功能。电缆舱火灾不仅会导致能源供应中断,还可能对周边生态环境和公共安全构成严重威胁。为了对这一特殊空间内的火灾风险进行系统的分析,构建蝴蝶结(Bow Tie,BT)模型并映射到动态贝叶斯网络(Dynamic Bayesian Network,DBN),通过实例分析确定某地下管廊电缆舱起火的先验概率与后验概率,识别主要风险源及对电缆起火的敏感度,探究在具备不同灭火屏障条件下的电缆舱火灾发展态势。结果表明,单相接地故障、制造缺陷、绝缘损坏、连接缺陷等是导致电缆舱起火的关键因素,并且随时间推进,各种不利后果发生概率呈上升趋势。这一发现凸显了及时有效的预防与早期干预措施对于遏制电缆舱火灾扩散的重要性。The cable cabin of the underground pipe gallery is the nerve center of urban infrastructure.In the process of operation and maintenance,the consequences of cable cabin fires are very serious.In order to reasonably evaluate the risk of underground pipeline cable,this paper developed a Dynamic Bayesian Network(DBN)model of cable compartment fire.A representative section of the cable compartment was selected for empirical analysis.24 risk factors were summarized through field studies and expert consultation.Firstly,the logical relationship between these risk factors was sorted out,and the bow model was constructed with an event tree and fault tree.Then,after obtaining the fuzzy scoring of risk factors by 5 experts,the conditional probability table of risk factors and the fuzzy evaluation results were obtained through the probability solution.According to the mapping logic,the network topology network was transformed into a Dynamic Bayesian Network,and the fire risk evaluation model of the underground pipe gallery based on a Dynamic Bayesian Network was established.The analysis results show that the fire risk probability of underground pipeline cable cabins changes from 0.1511 to 0.3006 with time,and the risk changes greatly,among which the risk probability of corrosion changes the fastest and the risk of human behavior with the smallest change.The reverse diagnosis results show that the probability of basic events“electrical corrosion”,“small cable spacing”,“connection defect”and“chemical corrosion”are 2.411%,1.559%,1.258%,0.976%,respectively,indicating a high occurrence probability,which is more likely to be the cause of cable cabin fire.The sensitivity analysis shows that single-phase ground fault(0.2587),manufacturing defect(0.2442),insulation damage(0.2043),connection defect(0.2016),short circuit between cables(0.1835)is the key factor leading to the cable compartment fire.In the accident scenario after the cable fire,the likelihood of the accident result increases with time,where the probability

关 键 词：安全工程 动态贝叶斯网络 电缆舱起火 风险评价 

分 类 号：X913.4[环境科学与工程—安全科学]