基于T-S模糊故障树与贝叶斯网络的作业风险评估  

Risk Evaluation for Engineering Work Based on T-S Fuzzy Fault Tree and Bayesian Network

作　　者：熊娜[1] XIONG Na(Nanchang Institute of Technology,Nanchang 330044,China)

机构地区：[1]南昌理工学院,南昌330044

出　　处：《火力与指挥控制》2024年第12期173-179,共7页Fire Control & Command Control

基　　金：军队预研基金资助项目。

摘　　要：针对传统故障树分析法存在一定局限性的问题,提出一种基于T-S模糊故障树与贝叶斯网络的作业风险评估方法。构建工程作业风险评估T-S模糊故障树模型,引入四边形隶属函数表征底事件风险概率和风险程度;根据底事件与中间事件、中间事件与顶事件之间的逻辑关系确定T-S模糊门规则;然后,根据T-S模糊故障树模型向有向无环图转换规则建立贝叶斯网络模型,根据T-S模糊门规则确定节点之间的条件概率关系。最后,以安排导弹战备值班为例进行了作业风险评估分析。仿真结果表明:T-S模糊故障树计算结果与贝叶斯网络正向推理一致,验证了评估方法的准确性与合理性;诊断推理确定了顶事件发生严重风险、一般风险情况下各底事件出现相应风险的概率大小,从而为确定风险隐患优先排查对象和重点排查对象提供决策参考。Aiming at the limitation of traditional fault tree analysis,a risk evaluation method for engineering work is put forward based on T-S fuzzy fault tree and Bayesian network.Firstly,a risk evaluation model for engi-neering work is established based on T-S fuzzy fault tree,and a trapezoidal membership function is introduced to describe the risk probability and degree of bottom events.Secondly,T-S fuzzy gate rules are determined by the relationships between the bottom events and the intermediate events,and also the top event.Thirdly,Bayesian network model is constructed according to the rules of converting T-S fuzzy fault tree into directed acyclic graph,and T-S fuzzy gate rules are converted into conditional probability relationships.At last,arrangement for missiles in combat readiness duty is taken as an example to perform risk evaluation.The result shows that the computa-tion result of the T-S fuzzy fault tree equals to that of Bayesian network which verifies the correctness and ratio-nality of the evaluation method.The probabilities of bottom events in different levels of risk are detemined through diagnostic reasoning of Bayesian network when the top event in the state of serious risk and general risk.And the results can be used to find out the key factors in risk identification and risk assessment.

关 键 词：T-S模糊故障树 贝叶斯网络 风险评估 四边形隶属函数 条件概率关系 

分 类 号：O232[理学—运筹学与控制论] TP393[理学—数学]