Incipient fault diagnosis of rolling bearings based on adaptive variational mode decomposition and Teager energy operator

Extracting incipient fault features of rolling bearings is a hard work as the impact compositions in vibration signals are faint and disturbed by a lot of environmental noise. An adaptive variational mode decomposition and Teager energy operator method (AVMD-TEO) for diagnosing incipient fault of rolling bearings is proposed. Firstly, the minimum average envelope entropy is used as the fitness value to search the optimal parameters of VMD adaptively by grey wolf optimization algorithm. Subsequently, the efficient weighted kurtosis index is introduced to select the efficient modal components for signal reconstruction. Finally, the reconstructed signal is processed by Teager energy operator to enhance the faint transient impact compositions and identify the defect frequency. The superiority of AVMD in parameters selection compared with fixed-parameter VMD and maximum weighted kurtosis optimized VMD is verified by simulated signal analysis. Results from the cases show that the peak signal-to-noise ratio and fault characteristic coefficient obtained by the proposed method are increased by 8% to 229% and 37% to 258% respectively compared with some traditional methods. The proposed AVMD-TEO can effectively reduce signal noise and extract incipient fault feature of rolling bearings.