Feature extraction and intelligent diagnosis for ball bearing early faults

In the study on ball bearing fault diagnosis based on wavelet transform, the parameter selection of wavelet transform and computation of fault features cannot be carried out automatically at present. Aiming at these problems, a new ball bearing fault feature auto-extracting method based on binary discrete wavelet transform is proposed in this article, which can select automatically wavelet function parameters and extract the fault features. In addition, an intelligent diagnosis model based on the neural network with self-adaptive structure is established to implement the intelligent diagnosis of ball bearing faults. Finally, practical ball bearing experiment data is used to verify the new method put forward in this article, and the results fully validate its application.