Bearing works in complex environment

which makes the bearing vibration signal have a certain nonlinearity. In order to solve this problem in bearing fault diagnosis

this study proposes a bearing fault depth modeling method based on time-frequency domain feature extraction. Firstly

the descriptive statistical parameters in the time domain of the vibration signal are extracted

and then the fast Fourier transform (FFT) is used to convert the time domain signal to the frequency domain. The frequency frequency domain features and descriptive statistical parameters in the frequency domain are extracted. Use the chi-square test to select important features in the time-frequency domain combined feature set. Finally

a deep belief network (DBN) is used to classify bearing faults. By comparing the single-domain feature extraction and the fault diagnosis results of different classification algorithms on the two datasets

it is verified that the proposed method can effectively improve the fault identification accuracy.