Fault Diagnosis of Bearing Based on Variational Mode Decomposition and Deep Learning

Abstract

Aiming at the problem of difficult fault diagnosis caused by serious noise pollution and weak fault characteristic information in the rolling bearing vibration signal, a fault diagnosis method based on the combination of variational mode decomposition (VMD) and deep learning is proposed. First, VMD is performed on the original bearing vibration signal to obtain several Intrinsic Mode Functions (IMF). Then, the envelope spectral entropy of each IMF can be obtained by calculating. The IMF with the smallest envelope spectral entropy is selected as the main analysis IMF. Secondly, a stacked auto encoder (SAE) network initial model is built according to the data characteristics, and the initial values of the model parameters can be obtained by performing unsupervised pre-training on the network model; then the supervised backpropagation algorithm is used to fine-tune the network parameters to obtain the model of optimal parameter. Finally, the model is use to perform pattern recognition on the test set. Validation of examples and comparative experiments show that this method has higher diagnostic accuracy, better diagnostic effect, and better engineering application value.