Research on Rolling Bearing Fault Diagnosis Based on Variational Modal Decomposition Parameter Optimization and an Improved Support Vector Machine

Aiming at the problems of modal aliasing and poor noise resistance when processing vibration acceleration signal rolling bearings by empirical decomposition (EMD), a variational (VMD) method based on parameter optimization is proposed. Combined with improved particle swarm algorithm (IPSO) envelope entropy, VMD layers penalty parameters were optimized. The components high correlation coefficients original screened out, fault characteristics extracted combining sample entropy. low classification accuracy support vector machine fixed in diagnosis stage defects gray wolf algorithm, such as insufficient population diversity large influence initial effect, an (IGWO) multistrategy improvement IGWO was combined to obtain (IGWO-SVM). bearing test bench established collect signals under different states. experimental results show that strong can be effectively realized applying above methods, average rate reaches 98.875%.