基于自动搜峰和shannon熵的车辆轴承多普勒畸变故障声信号校正研究

Fault Diagnosis of Rolling Bearing Doppler DisturbanceBased on Automatic Peak Search and Shannon Entropy

摘要:

针对传声器采集的运动声源信号存在多普勒畸变问题,提出一种基于自动搜峰和shannon熵的滚动轴承多普勒畸变故障声信号校正方法。首先对所采集的声音信号进行短时傅里叶(STFT)时频分析；然后利用自动搜峰方法进行瞬时频率估计,设置shannon熵来提高瞬时频率估计精度,并得到拟合的瞬时频率曲线,进而得到信号重采样时间点；最后对原信号进行时域重采样,从而使畸变信号得以矫正。通过仿真和动态滚动轴承内外圈故障声信号的实验验证了此种方法的可行性。

Abstract:

For the doppler distortion problem for motion source signals collected by microphones. The method for correcting the acoustic signal of a rolling bearing Doppler distortion based on automatic peak search and shannon entropy. Firstly, the collected sound signal is analyzed by short time Fourier (STFT) time-frequency analysis; then, the instantaneous frequency estimation is performed using the automatic peak search method, the shannon entropy is set to improve the accuracy of the instantaneous frequency estimation, and the fitted instantaneous frequency curve is obtained. Then the signal re-sampling time point is obtained; in the end, the original signal is re-sampled in the time domain so that the distortion signal can be corrected. The feasibility of this method is verified by experiments using acoustic signals of the inner and outer rings of the simulation and dynamic rolling bearing.