摘要
为减小时间开销,动态故障树经常被模块化分解为独立的静态子树和独立的动态子树,然后分别使用二叉决策图和马尔卡夫模型求解;其中的一个关键问题便是识别具有相互依赖基本事件和重复事件的动态故障树中所有的独立模块和最小独立模块.文中提出了一个基于亲戚依赖关系的独立模块识别方法IIMKDR,该方法将故障树按照其事件之间的依赖关系转换为依赖树;基于面向对象思想,为依赖树及其节点构建对象;通过对各节点对象属性的分析,得出独立模块集并进而求出最小独立模块集.最后对IIMKDR方法进行了理论分析和实验验证,分析表明该方法适用于具有相互依赖基本事件和重复事件的动态故障树独立模块识别.与其它方法从多个角度所进行的实验对比显示,当故障树没有相互依赖基本事件和重复事件时,该方法的开销略高于其它方法;当故障树具有相互依赖基本事件和重复事件时,该方法能够正确识别其中所有的独立模块,而其它方法不具备这样的功能.
In order to reduce the time cost, dynamic fault tree is always modularized into independent static modules and independent dynamic modules, then solved by BDD and Markov model separately. One of the key questions is how to identify all the independent modules and minimum independent modules in the dynamic fault tree with interdependent basic events and repeated events. An identification method, called IIMKDR, is proposed based on kinship dependency relation. IIMKDR converts fault tree to dependent tree according to the dependence relations among the events. Both dependent tree and nodes are then constructed based on the object-oriented idea, and all the independent modules and minimum independent modules are identified through the a- nalysis of the properties of nodes. Finally theoretical analysis and experimental verification are carried out on this method. The analysis indicates that this method can be applied to the dynamic fault tree with interdependent basic events and repeated events. The comparisons with other methods from various aspects show that the time cost of this method is slightly high than other methods given without interdependent basic events or repeated events, and this method can exactly identify all the independent modules while other methods cannot do given with interdependent basic events and repeated events.
出处
《计算机学报》
EI
CSCD
北大核心
2012年第2期229-243,共15页
Chinese Journal of Computers
基金
国家自然科学基金(60904082)
国家预研项目(51320010201)资助~~
关键词
动态故障树
独立模块识别
相互依赖基本事件
重复事件
可靠性分析
dynamic fault tree identification of independent module interdependent basic events
repeated events reliability analysis
