基于贝叶斯网络的核电站应急电力系统安全评价  

Application of Bayesian networks in safety assessment of a nuclear power plant＇s emergency power systems

作　　者：魏利强[1] 郑恒[2] 

机构地区：[1]清华大学核能与新能源技术研究院,北京102201 [2]北京理工大学管理与经济学院,北京100081

出　　处：《高技术通讯》2007年第6期628-632,共5页Chinese High Technology Letters

基　　金：863计划（863-614-02）资助项目.

摘　　要：在故障树分析法（FTA）基础上提出了一种基于贝叶斯网络（BN）的核电站应急电力系统安全评价方法,比较了FTA和BN在建立安全评价模型和评价能力上的不同。该方法在应对众多影响因素上有很大优势,能进行更多有意义的分析：既能进行前向的预测推理,又能进行后向的诊断推理,可以找出影响故障的组合模式,从而能够找出系统的薄弱环节。同时采用基于Matlab的BNT软件包,大大简化了计算过程。通过对10MW高温气冷堆（HTR-10）应急电力系统的安全评价实例的分析,证明该方法是对传统的基于故障树分析的安全评价方法的有益改进。This paper presents a Bayesian networks （BN） approach for safety assessment of the emergency power supply system of nuclear power plant on the basis of fault tree analysis （FTA）. Compared with FTA, the modeling process and the assessing ability of BN has many advantages when dealing numerous factors. It can be used to make more significant analysis, such as forward inference （prediction）, backward inference （diagnosis）, and finding the integrated mode of inducing faults, so as to find the bottleneck of the system. Adopting the Bayes net toolbox （BNT） software based on Matlab, the modeling and analyzing process is greatly facilitated. Finally, an example on the emergency powder supply system of the 101VP~ high temperature gas-cooled reactor （HTR-10） was provided to illustrate that the BN medaod was an improvement for FTA in the safety assessment.

关 键 词：应急电力系统 安全评价 贝叶斯网络 故障树分析 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]