p.739
p.743
p.752
p.758
p.762
p.770
p.785
p.790
p.795
Intellectual Gear Fault Detection Based on Wavelet Time-Frequency Analysis
Abstract:
In this paper, wavelet packet transform and support vector machines are used to detect gear system faults. Testing signals were obtained by measuring the vibration signals of gear system at different rotating speed for different faults. Vibration feature signals were analyzed using wavelet de-noising. By using wavelet packet transform (WPT), signals were decomposed into different frequency bands. the fault detection is used for calculation of energy percents of every frequency. All these were used for fault recognition using Support vector machine (SVM). SVM and neural network transform results were compared. The research indicates that the de-noised signal is superior to the original one. When dealing with various signals, such as Multi-Faults, the diagnosis identification rates are over 92%. This method can be effectively used not only in engineering diagnosis of different faults of gear system, but also for other machinery fault style classification.
Info:
Periodical:
Pages:
762-769
Citation:
Online since:
August 2013
Authors:
Price:
Сopyright:
© 2013 Trans Tech Publications Ltd. All Rights Reserved
Share:
Citation: