基于分层模糊符号有向图的故障诊断方法研究综述

A review of fault diagnosis methods based on hierarchical fuzzy sign directed graph model

本文简述了国内外故障诊断领域的研究现状,分类整理了各类方法优缺点,并提出基于定性模型的诊断方法具有不需要建立数学模型、故障推理能力强的优势,更适合于小卫星的部分分系统.选择定性模型法中的符号有向图法研究了建模原理,将分层结构与符号有向图法结合,得到了改进的分层符号有向图模型,从而减少了搜索范围,提高诊断效率.之后又将定性模型与隶属度函数这一定量信息结合,通过数字化的信息实现对故障程度的精确描述.最后提出了基于改进故障模型的诊断流程.本文旨在通过改进现有的符号有向图建模方法与诊断流程,为不易建模的小卫星系统如何进行自主故障诊断提出参考,从而为日后实现自主检测故障、隔离并恢复系统打下理论基础.

This paper describes the research status of fault diagnosis at home and abroad,classifies the advantages and disadvantages of various methods,and puts forward that the diagnosis method based on qualitative model has the ad-vantages of not needing to establish mathematical model and strong fault reasoning ability.By combining the hierarchical structure with the symbolic directed graph method,an improved hierarchical symbolic directed graph model is obtained which reduces the search range and improves the diagnostic efficiency.Then the qualitative model is combined with the membership function to describe the fault degree accurately.Finally,a diagnosis process based on improved fault model is proposed.The purpose of this paper is to improve the existing symbolic directed graph modeling method and diagnosis process to provide a reference for autonomous fault diagnosis of complex large-scale systems that are difficult to model.Thus,it lays a theoretical foundation for the realization of fault detection,isolation and recovery system in the future.