Application of GSABO-VMD-KELM in rolling bearing fault diagnosis

To address the difficulties in extracting fault features of rolling bearings and the low diagnostic accuracy, a fault diagnosis method for rolling bearings is proposed. This method integrates the Golden Sine Algorithm (GSA) with the Subtraction-Average-Based Optimizer (SABO) to form a Golden Sine Improved SABO Optimization Algorithm (GSABO). The GSABO algorithm is used for parameter optimization of Variational Mode Decomposition (VMD) and Kernel Extreme Learning Machine (KELM) in the fault diagnosis process. Firstly, the chaotic mapping strategy is used to optimize the population initialization of the Subtractive Clustering-Based Adaptive Optimization (SCAO) algorithm, enhancing population diversity. Secondly, the Golden Sine Algorithm (GSA) is integrated to improve the displacement algorithm, enhancing global search capability and effectively avoiding getting trapped in local optima. Then, the GSABO-VMD (Golden Sine Algorithm-Based Optimized Variational Mode Decomposition) is employed to decompose the rolling bearing fault signals, and the envelope entropy minimum criterion is used to select the effective modal components. Finally, time-frequency domain indicators of the selected modal components are computed to form a feature matrix, which is then input into GSABO-KELM (Golden Sine Algorithm-Based Optimized Kernel Extreme Learning Machine) for fault classification and recognition. Experimental analysis shows that compared to the unmodified SABO algorithm, GSABO has significant advantages in terms of escaping local optima, convergence speed, and accuracy. When compared with other traditional algorithms, GSABO-VMD-KELM achieves recognition accuracies of 99.3333 % and 99.0476 % on bearing data from Case Western Reserve University (CWRU) and Xi'an Jiao tong University (XJTU), respectively. This demonstrates the accuracy and superiority of the algorithm and provides valuable insights for engineering applications in rolling bearing fault diagnosis.