Research on Automotive Bearing Fault Diagnosis Based on the Improved SSA-VMD Algorithm

Due to the complexity and variability of automotive bearing fault vibration signals, as well as the challenges in feature extraction, a novel automotive bearing fault diagnosis method based on an improved SSA-VMD algorithm is proposed in this paper. Firstly, the number of modes (K) and the penalty factor (α) in the variational mode decomposition (VMD) algorithm are optimized by the Sparrow Search Algorithm (SSA). Secondly, the VMD using optimized parameters is used to decompose the automotive bearing fault vibration signal into a series of modal components. Modal components with larger kurtosis values are selected for reconstruction and feature extraction. Lastly, the feature vectors are input into a kernel extreme learning machine (KELM) model for fault recognition. A comparative analysis is conducted between the improved SSA-VMD method and the original SSA-VMD method in automotive bearing fault diagnosis. Experimental results demonstrate that the improved SSA-VMD method is satisfactory in extracting characteristic information from automotive bearing fault vibration signals and achieves higher fault diagnosis accuracy.