Rotating machinery fault diagnosis based on parameter-optimized variational mode decomposition

Abstract

Variational Mode Decomposition (VMD) has been widely applied in rotating machinery fault diagnosis. However, VMD needs to solve two key problems in its application: parameter optimization and the selection of high-quality Intrinsic Mode Functions (IMFs). To address these issues, firstly, the Sparrow Search Algorithm (SSA) is employed to optimize VMD parameters, incorporating fitness functions based on the Hoyer measure and the square envelope spectrum negative entropy. Subsequently, by minimizing the fitness function, optimal parameters for VMD are obtained. Secondly, a novel IMF selection method is proposed, evaluating IMF quality through sample entropy, energy ratio, and Spearman rank correlation coefficient to identify IMFs sensitive to fault features. Finally, the proposed method is validated using gear and bearing datasets containing noise of varying intensities, demonstrating its outstanding capability for extracting sensitive information in different noise environments.