A noise reduction and reconstruction method of rolling bearing based on EMD-AWT-SWT

To address the challenge of diagnosing rolling bearing fault in the presence of significant noise and varying speed conditions, an EMD-AWT-SWT model that integrates empirical mode decomposition (EMD), adaptive wavelet threshold denoising algorithm (AWT), and synchrosqueezing wavelet transform (SWT) is proposed, and the tacholess order analysis is carried out based on this model. Firstly, the intrinsic mode functions (IMFs) of the EMD are denoised utilizing the AWT, and then the processed IMFs are linearly superimposed to complete the signal denoising. Then, the time-frequency analysis is completed utilizing the SWT, the instantaneous frequency is accurately identified, and the identified curve is extracted and fitted. Finally, the variable speed signal is angular resampling based on the tacholess order analysis. The simulation outcomes demonstrate that the proposed approach can achieve effective noise reduction, and retain the fault characteristics after reconstruction, with high accuracy.