Bearing Fault Diagnosis Based on VMD-KPCA Fusion Algorithm for Vibration and Sound Signals

Traditional signal processing methods are often unable to fully reflect the operating status of bearings when dealing with vibration and sound signals, resulting in insufficient accuracy of fault detection. The single signal is easily disturbed by the surrounding environment, which affects the judgment of equipment reliability. In this paper, a fusion algorithm of sound and vibration signals based on variational modal decomposition (VMD) and kernel principal component analysis (KPCA) is proposed for bearing fault diagnosis. The VMD algorithm decomposes the signals efficiently to reduce the phenomenon of aliasing, while the KPCA enhances the feature extraction capability through the kernel function. The algorithm fuses the features of vibration and acoustic signals, which can realize complementary information and improve diagnostic accuracy and robustness. In this study, by analyzing the signals of normal and multiple fault states, efficient processing of data and accurate diagnosis are realized. The experimental results show that the VMD-KPCA algorithm proposed in this multi-proposed paper demonstrates superior robustness while improving the fault recognition rate, and has good prospects for practical application. Future research can further explore the optimization of this algorithm to meet the demands of industrial practical applications.