Rolling Bearing Fault Feature Extraction Based on Morphological Wavelet and S-transform
YANG Xian-yong,ZHOU Xiao-jun,ZHANG Wen-bin,YANG Fu-chun,LIN Yong
DOI: https://doi.org/10.3785/j.issn.1008-973x.2010.11.008
2010-01-01
Abstract:Aimed at the deficiency of traditional wavelet extracting impulse fault features from strong noise background,a fault feature extraction method for rolling bearing was proposed based on max-lifting morphological wavelet(MLMW) and S-transform.Firstly,decomposed to different levels by MLMW,signal's local maxima were mapped to scale signals and preserved over several scales,and the detail signals were denoised by soft threshold.Secondly,the signal was reconstructed,and fault features were extracted from the denoised signal's S-transform spectrum with excellent time-frequency focus characteristic.The experimental results show MLMW analysis not only reduces noise and harmonic components,but also significantly enhances fault features,and can extract nonlinear and non-stationary fault features more clearly than classical wavelet transform and envelopment analysis.Furthermore,with simple morphological operators and efficient lifting scheme adopted,the MLMW algorithm is simple and the cost is low,so it is suitable for on-line fault features analysis.