基于因果图最小割集的故障分析方法研究

复杂系统的最小割集很多,对故障诊断很重要,虽然因果图能够进行故障诊断,但缺乏对最小割集的分析.通过最小割集可以进行故障诊断,本文打算研究基于最小割集的故障诊断方法,并将这种新方法成功运用于煤矿瓦斯爆炸的故障分析中.由于根据故障发生的概率大小依次进行诊断,利用了正常事件的剪枝信息和多故障联合诊断,因此这种方法的诊断速度较快,效率较高.研究表明,基于因果图最小割集的故障分析方法,充分结合多种信息,可以快速有效地进行故障诊断.

Research on Diagnosis Approach Based on MCS of Causality Diagram

There are many Minimal Cut Set (MCS) which are important to fault diagnosis in complex system. Causality diagram methodology is can be applied in fault analysis, however lack of analysis on MCS which can help to diagnose. In this paper, the diagnosis approach based on MCS was proposed to develop causality diagram methodology and was successfully applied in gas explosion. The research shows that diagnosis approach based on MCS of causality diagram, is an effective and fast method to diagnose, because of combination multifold information, such as posterior probability of MCS, normal event and multifold fault.