Fault diagnosis method study in roller bearing based on wavelet transform and stacked auto-encoder

Considering the nonlinear and non-stationary characteristics of fault vibration signal in the roller bearing system, an intelligent fault diagnosis model based on wavelet transform and stacked auto-encoder is proposed. This paper firstly uses the combination of digital wavelet frame (DWF) and nonlinear soft threshold method to de-noise fault vibration signal. Then stacked auto-encoder is taken to extract the fault signal feature, which is regarded as the input of BP network classifier. The output results of BP network classifier represent fault categories. In addition, neural network ensemble method is also adopted to greatly improve the recognition rate of fault diagnosis.