文献类型：期刊文献

英文题名：Feather extraction of equipment fault diagnosis

作者：Zhang, Jun Liu, Fan

第一作者：张均

通信作者：Zhang, Jun

机构：[1] Guizhou Institute of Technology, Guiyang, Guizhou, 550004, China

第一机构：贵州理工学院

年份：2017

卷号：55

期号：12

起止页码：653-659

外文期刊名：Boletin Tecnico/Technical Bulletin

收录：EI(收录号：20174504380712);Scopus(收录号：2-s2.0-85032914657)

基金：This work was supported by Guizhou education department project: Embedded university innovation and entrepreneurship center.

语种：英文

外文关键词：Attitude control - Automation - Classification (of information) - Control systems - Extraction - Failure analysis - Principal component analysis - Process control

摘要：In this paper, principal component analysis (PCA) is used to extract the features of fault primitive data of the automatic control subsystem of spacecraft, and a few new features are obtained. Compared with primitive characteristics, the new characteristics quantity can better preserve the classification information of the original features, and the separation of the main faults of the 16 automatic control systems can be realized. Some parameters will change if there is a fault in the operation process of equipment. The cause and location of the fault can be gradually determined based on careful analysis of the changes in these parameters, that is, the fault diagnosis is completed. However, there may be dozens or hundreds of parameters to be accurately measured in highly complex systems, which consume a great deal of manpower, time and financial resources. The same parameter may have the same or similar trend of change under different failure modes.