DocumentCode
2319350
Title
Application of fault diagnosis expert system in grinding process
Author
Qin, Wei ; Yan, Wenjun ; Xu, Jing
Author_Institution
Coll. of Electr. Eng., Zhejiang Univ., Hangzhou, China
fYear
2010
fDate
16-20 Aug. 2010
Firstpage
290
Lastpage
295
Abstract
Directing at the complex structure, cockamamie production technology and great operation difficulty of vertical roller mill, this paper constructs the fault expert system to guide and optimize the grinding process. First, the expert database is established with fuzzy cluster analysis, principal component analysis (PCA) and tendency analysis. With support vector machine (SVM) regression, the multiple regression equations of main operation loops are fitted. In online, by looking up expert database and corresponding equations, the system gives out adjustment suggestions according to the variation tendency. Making use of OPC and VC++ techniques, this system has been applied in the practical field. The practical running conditions indicate that this system is effective and credible.
Keywords
condition monitoring; expert systems; fault diagnosis; fuzzy set theory; grinding; pattern clustering; powder technology; principal component analysis; production engineering computing; regression analysis; rolling mills; support vector machines; OPC techniques; VC++ techniques; cockamamie production technology; fault diagnosis expert system; fuzzy cluster analysis; grinding process; multiple regression equations; principal component analysis; support vector machine; tendency analysis; vertical roller mill; Databases; Employee welfare; Equations; Expert systems; Mathematical model; Monitoring; Principal component analysis; Expert System; Fuzzy Cluster Analysis; PCA; SVM; Vertical Roller Mill;
fLanguage
English
Publisher
ieee
Conference_Titel
Automation and Logistics (ICAL), 2010 IEEE International Conference on
Conference_Location
Hong Kong and Macau
Print_ISBN
978-1-4244-8375-4
Electronic_ISBN
978-1-4244-8374-7
Type
conf
DOI
10.1109/ICAL.2010.5585295
Filename
5585295
Link To Document