基于粗糙集的逻辑故障树方法及其应用

故障树是一种分析复杂系统可靠性、诊断系统故障的有效方法。在传统的故障树方法中 ,故障树的生成通常靠人工完成。故障知识的获取以及故障树结构的确定一直是有待解决的瓶颈问题。这里结合粗糙集、专家系统及人工智能等理论 ,提出了构造逻辑故障树进行故障诊断的方法并给出了相应的故障树评价标准。这种方法利用粗糙集对知识系统的知识发现和知识提取能力 ,从系统运行状态样本中建立基于知识的故障树模型。通过实例讨论了如何运用该方法对工业监控过程进行故障建模 ,检测系统运行过程中所发生的故障。

Rough Set-based Logical Fault Tree and Its Application

Fault tree analysis (FTA) is very effective in analyzing the reliability of control systems. In conventional FTA, fault trees are always generated by manual work. How to get knowledge about system's faults and construct the structure of fault tree automatically are key problems. Rough set theory (RST) is a tool for knowledge discovery. This paper proposes an approach for fault detection and diagnosis (FDD) based on RST and logical fault tree (LFT). Knowledge is extracted from samples by rough set reduction for LFT generation. Some definitions of criterion of LFT are given. Case study shows how to use this approach to model process system faults and to find system failure.