A Novel Rolling Bearing Fault Diagnosis Method Based on MFO-Optimized VMD and DE-OSELM

Rolling bearings are critical in maintaining smooth operation of rotating machinery and considerably influence its reliability. The signals collected from rolling field conditions often subjected to noise, creating a challenge extract weaker fault features. This paper proposes bearing diagnosis method that addresses the above-mentioned problem through moth-flame optimization algorithm optimized variational mode decomposition (MFO-optimized VMD) an ensemble differential evolution online sequential extreme learning machine (DE-OSELM). By using dynamic adaptive weight factor genetic cross operator, accuracy global ability (MFO) improved, two basic parameters VMD level quadratic penalty selected. Since vibration characteristics signal cannot be fully interpreted by single index, effective weighted correlation sparsity index (EWCS) is utilized relevant intrinsic functions (IMF) their energies as In order improve classification accuracy, energy feature set subsequently inputted into DE-OSELM for training purposes, proposed assessed via sample with four different health states actual bearings. Our results compared other methods, proving feasibility diagnose faults higher accuracy.