关键词: 滚动轴承, 特征提取, 微弱故障诊断, 双时域变换, 稀疏编码收缩

Abstract: Aiming at the weak early fault characteristics of rolling bearings which were difficult to extract under interference of noise and harmonics, a fault diagnosis method was proposed based on DTD transform and SCS. Firstly, original signals were transformed by DTD transform, and the diagonal series of DTD spectrum was used as reconstructed signals. Secondly, the reconstructed signals were handled by SCS method to suppress the interference of noises and low-frequency harmonics. Finally, denoised signals were analyzed by envelope spectrum to extract fault characteristic frequency, judge fault type, and realize fault diagnosis. Analyzed results of simulation signals, experimental signals, and engineering signals show that the proposed method may effectively extract weak fault characteristics in the early fault signals of rolling bearings and accurately judge the fault type.

Key words: rolling bearing, feature extraction, weak fault diagnosis, double time-domain(DTD) transform, sparse coding shrinkage(SCS)