The Research of Machinery Fault Feature Extraction Methods Based On Vibration Signal

The vibration signal of machinery fault is complex, performing non-stationary characteristic. The classical filter can not extract low frequency impulse from the vibration signal. In this paper the machinery fault feature extraction methods will be discussed. These methods includes Fourier transform frequency spectrum feature extraction, envelopment analysis and local mean decomposition. Envelopment analysis can extract the low frequency envelopment from single amplitude modulated pure mode function. Local mean decomposition is an adaptive time-frequency analysis method, which can decompose a non-stationary and complex signal into a number of amplitude modulated pure mode functions and extract the low frequency envelopment from signal mixed with strong noise. The mathematical theory of every feature extraction method is elaborated. In each method the vibration signal of common machinery fault is designed. It extracts the signal feature with these methods and analyzes the simulation output. The simple vibration experiment is done with land leveler experiment system. It proves these feature extraction methods are correct. (C) 2018, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.