关键词: 故障诊断, 同步压缩变换, 故障信号提取, 自适应阈值去噪

Abstract: In order to overcome the difficulties of feature extraction of non-stationary faulty signals in rolling bearing fault diagnosis, this paper proposed a fault feature extraction method by using the synchrosqueezing wavelet transform (SWT). Firstly, the measured vibration signals were processed with the continuous wavelet transform (CWT), and the wavelet transform coefficients were subjected to synchrosqueezing transform (SST). Moreover, an adaptive threshold denoising technology was presented to cancel noises of the SST coefficients, and the effective signal data near the center of the frequency were extracted by integrating. Finally, the signal reconstruction was carried out by utilizing the extracted effective signals. The simulation and the equipment tests were designed to verify the effectiveness proposed methods herein. The test results show that SWT has a high signal extraction accuracy and noise reduction capability. SWT also has higher time-frequency resolution, which may convert the fault signals into high-resolution time-frequency spectrum and make up for the lacks of CWT.

Key words: fault diagnosis, synchrosqueezing transform, faulty signal extraction, adaptive threshold denoising