摘要
故障诊断一直是工业领域研究的热点,国内外对此问题提出了很多相关的方法。针对故障诊断方法的研究角度和发展阶段不同,分类的方法也出现了很多种。目前常用的分类方法把故障诊断的方法分为基于模型的方法和基于数据驱动的方法。基于模型的方法是工业界比较早的一类方法,最近随着机器学习和数据挖掘技术的兴起,基于数据驱动的故障诊断方法已经成为工业领域的研究热点和发展方向,本文根据工业数据本身的特性,对基于数据驱动的故障诊断方法提出了一种新的分类方法,并对各类方法进行了详细阐述。
Fault diagnosis in the industrial field has always been hot,and a number of related methods were proposed on this issue at home and abroad.There are a variety of classification methods according to the research angle and the development stage of fault diagnosis methods.Now commonly the fault diagnosis method is divided into model-based approach and data-driven-based approach.Model-based approach is a earlier method of industrial fault diagnosis.With the development of the recent machine learning and data mining technology,the fault diagnosis based on data-driven approach has become the industry＇s research focus and direction of development.According to industrial data characteristics,a new classification method was proposed based on data-driven fault diagnosis method,and all kinds of methods were elaborated.
出处
《微计算机信息》
2010年第28期104-106,共3页
Control & Automation
基金
基金申请人:阎威武
项目名称:半监督回归学习理论和方法及其在工业过程软测量建模中的应用研究
基金颁发部门:国家自然科学基金委(60974119)
关键词
故障诊断
数据驱动
机器学习
Fault diagnosis
Data-driven
Machine learning
