基于卡尔曼滤波与子带选取的轴承声信号增强方法.

Aiming at the problem that the sound signal of gearbox bearing of belt conveyor was seriously disturbed by reverberation and running noise of other components, which led to the difficulty of acoustic diagnosis, the composition of acoustic signal, the cause of reverberation, the signal transmission path and the characteristics of each interference component were analyzed, and an acoustic signal feature enhancement method combining Kalman filter and optimal sub-band selection was proposed. Firstly, the Kalman filter hyperparameters were set according to the maximum kurtosis criterion, and Kalman filtering was used to reduce the interference of the reverberant noise to the fault acoustic signal of the bearing. Then, the wavelet packet noise reduction algorithm was used to process the signal after de-reverberation, the energy difference was compared between the sub-bands of the signal in the state to be measured and the normal state, the optimal sub-band containing the most fault information was selected, and the bearing fault features were extracted by the envelope spectrum analysis. Finally, the effectiveness of the proposed method was verified by bearing fault simulation tests and compared with the combination of improving singular value decomposition (ISVD) and resonance-based sparse signal decomposition (RSSD). The results of the study show that the proposed method is effective in de-reverberation and noise reduction, and the envelope spectrum contains obvious fault features and their associated components. The proposed method can realize the enhancement of bearing acoustic signal in indoor measurement environment and accurately extract the bearing fault characteristics. [ABSTRACT FROM AUTHOR]