With the rapid growth of the electric vehicle (EV) market, electric motors have become key components that must achieve both high power density and high efficiency within a limited installation space. Accordingly, research on novel motor topologies an...
With the rapid growth of the electric vehicle (EV) market, electric motors have become key components that must achieve both high power density and high efficiency within a limited installation space. Accordingly, research on novel motor topologies and advanced optimal design techniques has been actively conducted. Among various candidates, axial-flux permanent-magnet motors (AFPMMs) have attracted significant attention as promising alternatives for next-generation EV drive systems, owing to their shorter magnetic flux paths and higher torque density compared to conventional radial-flux permanent-magnet motors (RFPMMs). However, AFPMMs inherently exhibit complex three-dimensional (3D) electromagnetic behaviors and structural constraints, which necessitate 3D finite element analysis (FEA) for accurate performance evaluation. However, 3D FEA requires significant computational resources and long simulation times, which limits its direct application in optimal design processes. To address these challenges, this study proposes a novel analysis and optimization framework that significantly reduces computation time while maintaining analytical reliability. First, a point-specific analysis method (PSAM), combining quasi-3D and static 3D FEA, is introduced to predict key performance indices with minimal 3D FEA. This approach enables a substantial reduction in simulation time while maintaining high accuracy even in the early design stage. Furthermore, to minimize computational cost during the optimization process, latin hypercube sampling (LHS) and a newly proposed level-up dynamic area sampling (LEDAS) technique are employed. Based on these sampling strategies, a LEDAS-based kriging surrogate model is constructed to efficiently approximate the nonlinear relationship between design variables and motor performance. By integrating the proposed PSAM and LEDAS-based kriging surrogate model, a multi-objective optimal design of the AFPMM for EV traction applications is performed. The objectives are to maximize the average torque and minimize torque ripple while satisfying structural constraints such as cooling channel thickness and support configuration. The results demonstrate that the proposed approach dramatically reduces computational time compared to conventional 3D FEA-based optimization, without compromising prediction accuracy or reliability. Consequently, the optimized model achieves higher average torque and reduced torque ripple, thereby realizing both high power density and smoother torque characteristics suitable for EV applications. The proposed methodology is not limited to AFPMMs and can be readily extended to other electrical machines and energy conversion systems that involve complex 3D physical phenomena and high-dimensional design spaces. In particular, the combination of PSAM-based minimal analysis and LEDAS-based kriging surrogate optimization is expected to provide an effective and general framework for computationally efficient and accurate optimal design in future electric machine development.