Adaptive Suppression of Mode Mixing in CEEMD Based on Genetic Algorithm for Motor Bearing Fault Diagnosis

Fault signal analysis and processing are critical issues in bearing fault diagnosis, and complete ensemble empirical mode decomposition (CEEMD) is an effective means to dispose of fault vibration signals. In this article, a CEEMD optimization method based on the genetic algorithm (GA) is proposed, which suppresses mode mixing by adaptive matching of the Gaussian white noise amplitude. The proposed optimization method utilizes the GA’s global optimality to optimize the white noise amplitude in ensemble empirical mode decomposition (EEMD) and then processes signal by CEEMD with optimized Gaussian white noise. The results of selecting the amplitude of the Gaussian white noise by practical experience are compared. Compared with the empirical method results, this optimization method can adaptively match the appropriate white noise amplitude for different signals, which further suppresses the mode mixing phenomenon, and the fault frequency can be found in the spectrum diagram. It is shown that the proposed optimization method can process different signals adaptively and can be used for bearing fault diagnosis.