Multi-physics optimization design of the canned permanent magnet synchronous motor based on NGO-BP: Considering motor manufacturing errors

The canned permanent magnet synchronous motor (CPMSM) has attracted attention due to its high efficiency, energy saving, and high reliability. With the development of vacuum technology, high-end special motor applications have raised higher and more comprehensive performance requirements for the CPMSM, driving the demand for design optimization in various directions of this type of motor. In response to this demand, this paper proposes a multi-objective optimization method to improve the performance of the CPMSM in multiple physical fields, considering motor manufacturing errors. The optimization objectives in this paper include increasing the maximum torque, suppressing torque ripple, and reducing the temperature of the permanent magnet. Furthermore, the optimization process uses an NGO-BP neural network model to fit the numerical model solved by the finite element method. In addition, during the motor design process, the machining errors of the motor design were considered, and a stability test and comparative study were conducted on the final optimized results, with the optimal solution selected. The results show that the optimized structure of the CPMSM exhibits enhanced torque, reduced torque ripple, and lower permanent magnet temperature.