Intelligent Fault Severity Detection of Rotating Machines Based on VMD-WVD and Parameter-Optimized DBN

An intelligent fault severity detection method based on variational mode decomposition- (VMD-) Wigner-Ville distribution (WVD) and sparrow search algorithm- (SSA-) optimized deep belief network (DBN) is suggested to address the problem that typical diagnostic algorithms are inappropriate for extremely comparable vibration signals when samples insufficient. VMD used process original signal obtain band intrinsic functions (BIMFs) with different frequencies. WVD produces two-dimensional spectrum of key BIMF highest variance contribution rate. The input sample DBN composed a characteristic matrix formed by multiple signals. DBN’s learning rate batch size both tuned globally SSA, which has significant influence error. fitness function in parameter optimization network’s root mean square error (RMSE). Finally, loaded into best structure detecting severity. Experiments show that, VMD-WVD SSA-DBN, accuracy model rotating machines, good generalization ability robustness, can reach 98%. Compared BPNN, traditional DBN, VMD-DBN, VMD-PSO-DBN, other methods, proposed algorithm strong adaptive feature extraction application.