Optimization of VMD and SVM for Gear Fault Diagnosis Based on Crested Porcupine Optimization Algorithm

Aiming at the difficulty of feature extraction in gearbox fault diagnosis, a new classification method based on crested porcupine optimization (CPO) algorithm, variational mode decomposition (VMD) and SVM model is proposed. Because the parameter setting of VMD method has great influence on the decomposition effect of gear vibration signal, CPO is proposed to optimize the VMD method. Decompose the vibration signal using parameter optimized VMD, and the obtained components are reconstructed by correlation analysis. Input the reconstructed signal into the support vector machine model for fault classification. Through experimental analysis, the accuracy of gear fault diagnosis by the proposed method is 95%.