A Computationally Efficient Fea–surrogate–based Framework for Multi-Objective Optimization of Ipm Motors Considering Rotor Notch Geometry Sensitivity

This study proposes a computationally efficient framework for the design and multi-objective optimization of an Interior Permanent Magnet (IPM) motor, with a particular focus on rotor notch geometry refinement to enhance both electromagnetic and thermal performance. A systematic sensitivity analysis of the notch depth (ND), notch inner arc (NIA), and notch outer arc (NOA) is performed using finite element analysis (FEA) to quantify their influence on torque ripple, back-electromotive force (BEMF) harmonic distortion, and peak winding temperature. To reduce the prohibitive computational cost associated with direct FEA-based optimization, Gaussian Process Regression (GPR) surrogate models are trained on the generated dataset of FEA simulations. These surrogate models are subsequently coupled with the Non-dominated Sorting Genetic Algorithm II (NSGA-II) to efficiently explore trade-offs among competing performance objectives and to determine Pareto-optimal design solutions. Further performance improvements are obtained through rotor skewing, which contributes to the attenuation of torque ripples and electromagnetic harmonics. The optimized motor configuration is finally validated via high-fidelity FEA, demonstrating notable enhancements in torque smoothness, thermal robustness, and BEMF waveform quality. The proposed methodology provides an effective and robust design tool for the development of high-performance IPM motors intended for electric vehicle applications.