Research on Demagnetization Fault Diagnosis of Permanent Magnet Linear Synchronous Motor Based on SqueezeNet Neural Network

This paper studies the demagnetization fault diagnosis method of Permanent Magnet Linear Synchronous Motor based on SqueezeNet neural network. A new demagnetization fault signal acquisition method is proposed to adapt to the spatial topological structure constraints of the double-stator coreless motor, and to obtain effective demagnetization fault signals without invasive measurement, so as to improve the accuracy of the fault signal source. At the same time, a simple linear motor demagnetization fault diagnosis device is designed. The one-dimensional demagnetization fault signal is converted into a two-dimensional image through the Recurrence Plot, and fault feature information is effectively extracted. In addition, this paper innovatively uses the lightweight SqueezeNet model for training. After continuous adjustment of the SqueezeNet network model, it can efficiently complete the classification of permanent magnet linear synchronous motor demagnetization faults.