Torque Quality Optimization of Combined-Halbach Array AFPMM with Asymmetric Skewed Poles Based on SMPSO Algorithm

Due to its high torque density and compact structure, axial-flux permanent magnet machine (AFPMM) shows obvious advantages in direct drive and miniaturization. In order to further adapt the machine to the requirements and improve the performance, an asymmetric bidirectional skewed poles combined-Halbach array rotor structure is proposed and applied to the AFPMM to improve the torque density, suppress the torque ripple and improve the stability of the machine. Firstly, a 3D finite element method (FEM) was used to establish the machine. Then, the XGBoost (eXtreme Gradient Boosting) model was used to predict the machine parameters, and establish high-precision surrogate model. Finally, the speed-constrained multi-objective particle swarm optimization (SMPSO) algorithm was used to optimize the rotor skewing angle, phase shifting angle, permanent magnet (PM) thickness and pole-arc coefficient, aiming at average torque, torque ripple and unit volume magnet torque. Compared with the machine without skewed pole, the machine proposed in this paper showed obvious advantages, the average torque increased by 5%, the torque ripple decreased by 50%, and the unit volume magnet torque increased by 23.5%.