Application of Auto-Regulative Sparse Variational Mode Decomposition in Mechanical Fault Diagnosis

The variational mode decomposition (VMD) method has been widely applied in the field of mechanical fault diagnosis as an excellent non-recursive signal processing tool. performance VMD depends on its inherent prior parameters. Searching for key parameters using intelligent optimization algorithms poses challenges internal essence and fitness function selection algorithm. Moreover, computational complexity is high. Meanwhile, such methods are not competitive evaluating orthogonality between intrinsic functions reconstruction error a joint indictor termination decomposition. Therefore, this paper proposes new auto-regulative sparse (ASparse–VMD) to achieve accurate feature extraction. regularization term handles sparsification by constructing L2-norm with damping coefficient ε, number K set adaptively recursive manner ensure appropriateness. penalty parameter α dynamically selected according modes sampling frequency. update step τ algorithm signal-to-noise ratio singleness modal components suppress aliasing. experimental results simulation measured demonstrate effectiveness proposed strategies improving defects VMD. Extensive comparisons state-of-the-art show that more effective practical hybrid extraction faults.