Bearing fault diagnosis method based on improved approximate conjugate gradient pursuit

Ensuring the dependability, functionality, production effectiveness, and safety of mechanical systems necessitates assessing the conditions and detecting faults in rolling bearings. However, the fault features are usually hidden due to the interference of background noise and other unstable factors. To address this issue, the weak selection approximate conjugate gradient pursuit (WACGP) method and an improved sine cosine algorithm (ISCA) were introduced for more effective extraction of bearing fault features. Sine cosine algorithm(SCA) includes an inertia weight and nonlinear parameter update approach to improve the efficiency and accuracy of sparse signal representation, while the approximate conjugate gradient pursuit (ACGP) was modified to increase the speed and ability of identifying bearing fault characteristics. The validity of the method was confirmed by analyzing some bearing fault simulation signals and a certain actual vibration signals of the bearing' s inner and outer ring. The proposed method outperforms the gradient pursuit algorithm based on sine cosine optimization in terms of efficiency and accuracy. © 2024 Chinese Vibration Engineering Society. All rights reserved.