Rolling Bearing Fault Diagnosis Based on SABO–VMD–IBES–SVM

In order to enhance the diagnostic accuracy and model generalization capability for rolling bearing faults, this paper proposes a fault diagnosis method based on an improved bald eagle search (IBES) algorithm, which is optimized using multiple strategies and combined with support vector machine (SVM). First, the original data are preprocessed using the variational mode decomposition (VMD) optimized by the subtractive average–based optimizer (SABO). Next, the bald eagle search algorithm is enhanced by incorporating scrambled Halton initialization, Lévy flight strategy, dynamic elite retention mechanism, adaptive inertia weight, and differential mutation–based collaborative optimization. Finally, a rolling bearing fault diagnosis model based on the SABO–VMD–IBES–SVM approach is established, and a series of simulation experiments are conducted to compare this model with traditional SVM, GA–SVM, PSO–SVM, and BES–SVM models. The results demonstrate that the SABO–VMD–IBES–SVM model achieves a diagnostic accuracy of 99.167%.