A novel analytical method using virtual PM blocks to optimize magnet segmentations in surface-mounted PM synchronous machines

Infinite numbers of virtual permanent magnet (PM) blocks can be used to represent the segmented magnets per magnet pole in any combination of slot and pole numbers of the semi-closed surface-mounted permanent magnet synchronous machines (PMSMs). This concept is exploited and developed in the two-dimensional (2D) analytical subdomain model to predict the motor performances. The model is able to facilitate the analytical investigation to determine the optimum magnet pole-arcs of each magnet segment and the optimum spacing between the magnet segments. This novel method is then applied to a three-phase, 12s/8p surface-mounted PMSM. It is found that for two segmented magnets (2SM) per magnet pole machine, the optimum settings are 147.6° elect. in each magnet pole and 11.2° elect. of airgap spacing between the magnet segments. By comparing the results to the optimum magnet pole-arc of one magnet segment (1SM) per magnet pole machine, the cogging torque and the total harmonic distortion of the phase back-EMF are significantly reduced by 89 % and 25 %, respectively, where both machines utilize the same PM volume. These analytical results are also validated by the 2D finite element method (FEM), where good agreement has been achieved.