基于模块化分解的故障树分析方法

Fault Tree Analysis Method Based on Modular Decomposition

传统故障树分析算法存在诊断成本高和耗时长的问题,为此,在研究故障树结构中的特殊规律的基础上,采用深度优先最左遍历算法对故障树进行模块化分解,减小故障树分析的规模。结合if-then-else运算符,将最左底层模块子树转化为相应的二元决策图结构。运用深度优先最左遍历算法得到该二元决策图结构中的割集和最小割集,用相同故障概率的基本事件替代最左底层模块子树得到新故障树。采用自底向上、从左至右的递归综合分析思想,获得系统元件故障发生的概率,实现对故障树的分析。对故障实例的分析诊断结果表明,该方法可有效提高诊断速度,减少诊断成本。

For solving the diagnose cost and time applied to traditional fault tree analysis,based on studying the special disciplinarian of fault tree,this paper adopts the depth first left most searching algorithm to decompose the fault tree into modules,and decreases the scale of fault tree analysis.Combined with if-then-else operator,it converts the most left and bottom module binary decision diagram.It applies the depth first left most searching algorithm to acquire the cut set and the minimum cut set of the binary decision diagram,and then uses a new bottom event with the same failure probability to replace the module to generate a new fault tree.The probability that system elements occur faults is obtained by comprehensive analysis of from bottom to up and from left to right,and the fault tree analysis is finished.By analyzing fault diagnosis,it is verified that the method improves the speed of diagnose and decreases the cost of diagnose.