基于贝叶斯网络的燃油控制系统人因安全性定量分析  

Quantitative Human Safety Analysis of Fuel Control System based on Bayesian Network

作　　者：侯政良[1] 杨自春[1] 陈国宾[1] 

机构地区：[1]海军工程大学船舶与动力学院,武汉430033

出　　处：《燃气轮机技术》2010年第4期38-43,共6页Gas Turbine Technology

基　　金：总装武器装备预研基金项目(9140A27050106JB11);总装"十一五"装备预先研究项目(513270301)

摘　　要：通过分析概率安全评价对人因安全性分析的本质需求,比较第一代和第二代人因可靠性分析方法,运用基于人的认知过程分析的安全性分析方法对概率安全评价中的人因事件进行计算分析,并针对传统事件树、故障树分析方法的局限性,采用了贝叶斯网络方法使得整个分析评价过程方便、有效。最后对燃气轮机燃油控制系统进行建模,并选取三种情境进行模拟和灵敏度分析,对相应的结果进行分析讨论,找出了安全性薄弱环节并提出相应的对策。Considering the fundamental demand of the PSA（Probabilistic Safety Assessment） upon HSA（Human Safety Analysis）,compare the first generation（Human Reliability Analysis） with the second generation HRA,the human errors in the PSA were analyzed and calculated by the HAS method based on Human Cognitive.Because of the deficiency of event tree and fault tree,Bayesian network was used to make the whole analysis progress more convenient and efficient.At last,a model was built up to analyze the fuel control system,three cases were simulated and sensitivity analysis was made.The weak parts of system safety were found and corresponding disposals were also proposed.

关 键 词：概率安全评价 人因安全性 人的认知过程 贝叶斯网络 灵敏度分析 

分 类 号：TK413-8[动力工程及工程热物理—动力机械及工程]