Multi-objective shape optimization of Permanent Magnet Synchronous Motor based on Kriging surrogate model and design domain reduction

Permanent Magnet Synchronous Motor (PMSM) is a nonlinear, multi-physics coupled system that makes it difficult to build an accurate mathematical model to optimize design parameters. Traditional design for PMSM always relies on the experience of engineers. Although some works have been done for the size optimization of motors, the performances of PMSM still need to be improved. In this paper, a multi-objective shape optimization method is proposed for the optimal design of PMSMs. In the optimization model, the shape of slot and the size of permanent magnets are considered as design variables. The objective is to minimize the torque ripple and the loss under the constraint of average torque of motors. To obtain the accurate global solution, the Kriging surrogate model algorithm with an effective design domain reduction is used. Several novel designs that can obviously reduce the torque ripple and loss of PMSM are obtained by using the proposed method. The optimization results also indicate that using the proposed shape optimization algorithm is more effective in the optimal performance design of PMSM than using size optimization methods.