EEMD-Based Feature Fusion of Acoustic and Vibration Signals for Bearing Fault Diagnosis
Aiming at the problem that the acoustic and vibration signals in rolling bearing fault diagnosis have the characteristics of non-smoothness and non-linearity,and at the same time,the fault characteristic information contained in a single sound or vibration signal is not comprehensive,a method is proposed to use the ensemble empirical modal decomposition(EEMD)to process the signal of rolling bearing and extract the characteristic quantities of the acoustic and vibration signals for the fault judgment.Firstly,using the EEMD method,the IMF components containing fault features are selected for signal reconstruction,and the fault information of rolling bearings after noise reduction is obtained by fast Fourier transform.Then,the characteristic MFCC map of each fault and the IMF component cliff value and energy entropy of the vibration signal are obtained based on the sound signal,and the acoustic and vibration signal characteristics of the rolling bearing are obtained.Finally,the fusion fault judgment is carried out based on each feature quantity.The correct rate of fault judgment of acoustic and vibration signal feature fusion shows that the method can obtain richer fault judgment information compared with the single signal fault judgment method,and the method improves the correct rate by 7.88%compared with the single sound signal fault judgment method,and improves the correct rate by 3.23%compared with the single vibration signal fault judgment method,which verifies that the EEMD-based acoustic and vibration signal feature fusion bearing fault diagnosis method is correct.
rolling bearingsEEMDacoustic and vibration signal fusionfault diagnosis
万方数据
维普
