On the Iron Loss Reduction Design Improvement of an Axial Flux Permanent Magnet Motor

Reducing iron loss in axial flux permanent magnet (AFPM) motors is critical for improving efficiency. This study proposes a design-optimization procedure that combines 3D finite-element analysis (FEA) data with an artificial neural network (ANN) surrogate. For four design variables—airgap length, rotor back-yoke thickness, stator slot width, and stator slot depth—the search bounds were defined to avoid tooth and back-yoke saturation, and the corresponding space was sampled to construct a dataset. Using this dataset, the ANN was trained and then used to explore low-iron loss solutions. On an independent validation set, ANN predictions showed high agreement with 3D-FEA reference values, enabling rapid evaluation of many design candidates. As a result of the optimization, total iron loss decreased relative to the baseline, and torque increased by 3 Nm. These results demonstrate that the ANN-based surrogate model can reliably perform geometry-dependent iron loss optimization in AFPM motors, providing a fast and accurate alternative to repetitive 3D-FEA evaluations.