基于模块分解的多态故障树可靠性分析方法  

A Modular Method for Multi-State Fault Tree Reliability Analysis

作　　者：李淑敏[1] 孙树栋[1] 司书宾[1] 蔡志强[1] 

机构地区：[1]西北工业大学机电学院,陕西西安710072

出　　处：《西北工业大学学报》2014年第2期251-255,共5页Journal of Northwestern Polytechnical University

基　　金：国家自然科学基金(71271170;71101116);西北工业大学基础研究基金(JC20120228)资助

摘　　要：提出了基于模块分解的多态故障树分析方法:采用基于部件的线性时间改进算法将多态故障树分解成相互独立的子故障树,并基于多态多值决策图方法对子故障树进行建模,递归综合子故障树可靠性结果计算整个系统的可靠性;解决了底事件非统计独立的多态故障树模块分解问题。通过案例分析与传统故障树分析方法进行对比,验证了该方法的正确性和有效性。The modular analysis method based on binary decision diagrams was proposed for reducing the computa-tional cost of basic operations on binary fault tree, in which the basic events are statistically independent. To ana-lyze multi-state fault trees （ MFT） with dependent basic events, this paper proposed a modular analysis method for multi-state fault trees, which is explained in section1 and 2. Section 1 introduces the multi-state multi-valued deci-sion diagram （ MMDD） and gives the modeling method of MMDD based on multi-state fault tree analysis （ MFTA） . The core of section 2 consists of：（1） we divide the multi-state fault trees into independent subtrees, using a com-ponent-based linear-time enhanced algorithm;（2） multi-state multi-valued decision diagram is applied to each sub-tree for modeling and reliability analysis, and the results of various independent subtrees are integrated to obtain system reliability. Section 3 analyzes a case study;compared with existing MFTA approach, the proposed modular method is more computationally efficient.

关 键 词：多态故障树 模块化 线性时间算法 多态多值决策图 

分 类 号：TB114.3[理学—概率论与数理统计]