改进DDET的分支生成机制——设备失效概率阈值法

概率阈值法是离散动态事件树(dynamic discrete event tree,DDET)常用的分支生成方法,此方法存在没有考虑实际设备失效的问题,建立的DDET具有较大不确定性,本文针对概率阈值法存在的不足提出相应的改进措施——设备失效概率阈值法.通过动态水箱实例分析发现,设备失效概率阈值法的总节点数仅是概率阈值法50%左右,计算时间节省了近70%,说明设备失效概率阈值法具有重要的工程应用价值;设备失效概率阈值法得到的总事故概率比概率阈值法的保守一些,及其建立的DDET更能贴近动态水箱的真实情况来反应系统的运行动态特性,利于研究动态水箱更为真实的事故演化过程.

Improvement of generating branch for DDET——Equipment failure probabilistic threshold value method

Probabilistic threshold value method (PTVM) always were used to generate branch for dynamic discrete event tree (DDET) that had two deficiencies:didn't take failure of equipment into account, more uncertainty of building DDET, this paper presented equipment failure probabilistic threshold value method (EFPTVM) to make up for the deficiencies. Through analyzing the case of dynamic cistern, EFPTVM got less 50 percent node, less 70 percent counting time than PTVM, showing that EFPTVM had important value of engineering application; EFPTVM counted accident probability more conservative than PTVM and its DDET reflected system operational characters closer to real situation of dynamic cistern, be fitter for studying operational dynamic characters of dynamic cistern.