Research on rolling bearing fault diagnosis method based on SBOA-Optimized FMD

Abstract To address the challenges in extracting rolling bearing fault features and the parameter sensitivity of Feature Mode Decomposition (FMD), this study introduces a novel approach that integrates the Secretary Bird Optimization Algorithm (SBOA) with FMD. The proposed method leverages SBOA to identify the most suitable parameter settings for FMD, employing the minimum arrangement entropy as the objective function. The optimized FMD is then employed to process the raw signals. The most informative components are selected based on the correlation coefficient-cliffiness index. Their time-domain characteristics and multiscale dispersion entropy (MDE) are computed to construct the fault feature vectors. Finally, SVM with SBOA optimization is used for classification. Empirical results show that SBOA-FMD achieves better fitness values than other methods with the same number of iterations. The method achieves an average recognition accuracy of 99.45%, demonstrating its effectiveness.