Gearbox fault diagnosis using independent component analysis in the frequency domain and wavelet filtering

We combine independent component analysis in the frequency domain (ICA-FD) and Morlet wavelet filtering for gearbox fault diagnosis. Collected vibration signals from a gearbox are separated into two components with ICA-FD. Morlet wavelet filtering is then applied to the separated components. The opt...

Full description

Saved in:
Bibliographic Details
Main Authors: Xinhao Tian, Jing Lin, Fyfe, K.R., Zuo, M.J.
Format: Conference Proceeding
Language:English
Subjects:
Blind source separation
Fault diagnosis
Filtering
Frequency domain analysis
Independent component analysis
Signal processing
Source separation
Vibrations
Wavelet analysis
Wavelet domain
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We combine independent component analysis in the frequency domain (ICA-FD) and Morlet wavelet filtering for gearbox fault diagnosis. Collected vibration signals from a gearbox are separated into two components with ICA-FD. Morlet wavelet filtering is then applied to the separated components. The optimal shape parameter, /spl beta/, of the basic Morlet wavelet is obtained by minimizing the wavelet entropy. Better diagnosis results are obtained with this combination than using wavelet filtering alone.
ISSN:1520-6149
2379-190X
DOI:10.1109/ICASSP.2003.1202340