Stator Magnetic Field Consistency Fault Detection of MMMT-PMSLM Based on Bayesian Optimization Convolutional Neural Network

In response to the problem of inconsistent air gap magnetic fields caused by different stator winding heights in modular moving magnet type permanent magnet synchronous linear motors (MMMT-PMSLM), a Bayesian optimization convolutional neural network (BO-CNN) based method for detecting stator magnetic field consistency is proposed. This method enables accurate identification of stator winding Magnetic field inconsistency fault in MMMT-PMSLM. First, based on the stator winding airgap magnetic density signal, new fault feature information is reconstructed through envelope reconstruction. Second, the impact rate of faults on motor thrust performance is used as a criterion for evaluating the severity of faults, and a fault sample library is established. Then, using the BO-CNN classification method, the fault magnetic density feature information is used as the input of the CNN network, and the fault severity level is used as the output of the CNN to train the data in the sample library. Construct a fault classifier for stator winding magnetic field inconsistency. Finally, the simulation experiment results demonstrate the accuracy of the proposed method.