A Parameterized Linear 3D Magnetic Equivalent Circuit for Analysis and Design of Radial Flux Magnetic Gears–Part II: Evaluation

This is Part II of a two-part paper on the extension of a previously developed parameterized linear 2D magnetic equivalent circuit (MEC) for radial flux magnetic gears with surface permanent magnets to a 3D model. Part I explains the implementation of the 3D MEC. This section, Part II, evaluates the 3D MEC model's accuracy relative to 2D and 3D nonlinear finite element analysis (FEA) models. First, analysis of three base designs illustrates the impacts of the axial discretization parameters and stack length. This leads to the development of guidelines for selecting the axial discretization parameters. Additionally, the MEC is shown to accurately match 3D FEA in predicting the axial variation of flux densities. Finally, 3D MEC, 2D FEA, and 3D FEA models are deployed for a 144,000 case parametric study. The 2D FEA model significantly overpredicts the torques of designs with short stack lengths. However, the 3D MEC provides a much better agreement with 3D FEA, with less than 1% average absolute discrepancy. Additionally, the 3D MEC model is much faster than the 3D FEA model, with average 3D MEC model evaluation speeds ranging from about 100 times to 300 times faster than the 3D FEA for the base designs.