Article (Scientific journals)
Fault detection based on Kernel Principal Component Analysis
Nguyen, Viet Ha; Golinval, Jean-Claude
2010In Engineering Structures, 32, p. 3683-3691
Peer Reviewed verified by ORBi
 

Files


Full Text
KPCA_ES.pdf
Author preprint (350.71 kB)
Request a copy

All documents in ORBi are protected by a user license.

Send to



Details



Keywords :
PCA; KPCA; subspaces; nonlinearity; detection
Abstract :
[en] In the field of structural health monitoring or machine condition monitoring, the activation of nonlinear dynamic behavior may render the procedure of damage or fault detection more difficult. Principal Component Analysis (PCA) is known as a popular method for diagnosis but as it is basically a linear method, it may pass over some useful nonlinear features of the system behavior. One possible extension of PCA is Kernel PCA (KPCA), owing to the use of nonlinear kernel functions that allow to introduce nonlinear dependences between variables. The objective of this paper is to address the problem of fault detection (in terms of nonlinear activation) in mechanical systems using a KPCA-based method. The detection is achieved by comparing the subspaces between the reference and a current state of the system through the concept of subspace angle. It is shown in this work that the exploitation of the measurements in the form of block Hankel matrices can improve effectively the detection results. The method is illustrated on an experimental example consisting of a beam with a geometric nonlinearity.
Disciplines :
Mechanical engineering
Author, co-author :
Nguyen, Viet Ha ;  Université de Liège - ULiège > Doct. sc. ingé. (aérosp. & méca. - Bologne)
Golinval, Jean-Claude  ;  Université de Liège - ULiège > Département d'aérospatiale et mécanique > LTAS - Vibrations et identification des structures
Language :
English
Title :
Fault detection based on Kernel Principal Component Analysis
Publication date :
September 2010
Journal title :
Engineering Structures
ISSN :
0141-0296
eISSN :
1873-7323
Publisher :
Elsevier Science, Oxford, United Kingdom
Volume :
32
Pages :
3683-3691
Peer reviewed :
Peer Reviewed verified by ORBi
Available on ORBi :
since 01 October 2010

Statistics


Number of views
108 (5 by ULiège)
Number of downloads
2 (0 by ULiège)

Scopus citations®
 
146
Scopus citations®
without self-citations
145
OpenCitations
 
121

Bibliography


Similar publications



Contact ORBi