A Novel Approach of Multiscale Feature Extraction for Gearbox Condition Monitoring

Abstract

This paper proposes a novel multiscale slope feature extraction method using wavelet-based multiresolution anlaysis for gearboxes fault identification. The new method mainly includes the discrete wavelet transform (DWT), the variances calculation of multiscale detailed signals, and the wavelet-based multiscale slope features estimation. Experimental results show that the wavelet-based multiscale slope features show excellent clustering for different work conditions and have the merits of high accuracy and stability in classifying different conditions of gearbox.