Northern Cthulhu Algorithm Optimized VMD Combined with SVM for Fault Diagnosis

Abstract

For a long time, in the face of complex signal processing, such as rolling bearings signals and other complex nonlinear signals, most of them are using traditional signal processing methods to extract signal features. However, it is difficult for general signal processing strategies to extract all the signal features contained in the signal one by one. With mature signal extraction methods like variational mode decomposition (VMD), the number of layers of signal decomposition determines the effect of final fault detection. To solve this problem, this paper proposes a northern goshawk optimization (NGO) algorithm to optimize the VMD and find the optimal decomposition parameter K, which further improve the detection effect. Finally, the experimental data simulated in the MATLAB software platform shows that the detection effect achieved by the optimized VMD of the NGO algorithm is improved by 6.4814%.