Fault Diagnosis Method of Servo Motor Bearing Installation Misalignment Based on CSFF-CNN

As an important part of servo motor, bearing misalignment fault diagnosis has attracted more and more attention. In order to improve the fault diagnosis performance of servo motor gear, a fault diagnosis method based on compressed sensing feature fusion is proposed. In this paper, the alpha phase current is collected and converted into the image by the image transformation method. In order to effectively improve the feature extraction ability of the model for key information, this paper proposes to use the compressed sensing of the signal and the compressed sensing of the signal transformed into the image and the neural network to extract the feature of the original signal transformed into the image. In order to improve the feature extraction ability of the model for multi-gradient information flow, this paper proposed a module named multi-gradient information flow module. The performance of the model proposed in this paper has been verified by the collected servo motor gear data. Under different working conditions, the effect of the method proposed in this paper is improved compared with that of convolution neural network, and the model has strong generalization ability, which proves the superiority of the model performance.