Optimal Design and Experimental Test of Surface-Mounted Permanent Magnet Motors with Cost-Effective Magnet Utilization to Suppress Torqu Pulsations

Abstract

The surface-mounted permanent magnet (SPM) motors with NdFeB magnets, offering high torque density and high efficiency, have been widely applied to various domestic and industrial applications [1]. However, the adoption of NdFeB magnets not only brings high torque density, but also leads to high torque pulsations and high magnet cost. The approaches to suppress torque pulsations in SPM motors have been heavily investigated through drive control or motor design methods [2], [3]. In particular, the suppression of cogging torque has received great attention along with numerous methods, such as skewing [4], auxiliary slots [5], teeth notching [6], and slot-opening shifting [7], etc. However, most of the methods to suppress cogging torque may not result in low torque ripple due to the effects of armature reaction fields. Hence, the approaches to suppress both cogging torque and torque ripple simultaneously are more desired. It is well known that the magnet flux density distribution has a significant effect on torque performance. Accordingly, extensive magnet shaping methods, such as magnet pole shape optimization [8] and sinusoidal magnet poles [9], have been reported to obtain a sinusoidal magnet flux density distribution, thus to reduce torque pulsations. However, these reported methods inevitably lead to manufacturing difficulty and performance degradation. Regarding to the magnet cost saving, the approaches by using ferrite magnets, or hybrid ferrite and NdFeB magnets are investigated, but generally neglecting the issues on torque pulsations [10]. In this paper, an optimal design is proposed for the SPM motor to reduce the cogging torque and toque ripple, and save the magnet cost using multi-grade NdFeB and ferrite magnets. Based on a conventional SPM motor with single-grade NdFeB magnets, the proposed SPM motor is designed with three-grade NdFeB and ferrite magnets, and then optimized to further reduce torque pulsations and save the magnet cost by maintaining the high average torque using the Kriging method and a genetic algorithm. All the motor characteristics are first predicted using the finite element method (FEM) at the same operating conditions. Then the experimental test is performed for the optimized model to validate the optimal design and analysis results.