Design Optimization of a Brushless PM Outer Rotor Motor for Electric Scooters: A Surrogate-Based Multi-Objective Approach

This study presents a multi-objective design optimization of a Brushless Permanent Magnet Outer Rotor (BPMOR) motor intended for electric scooter applications, focusing on the impact of rotor-mounted magnet geometry. Specifically, the permanent magnet (PM) thickness, magnet arc, and magnet reduction—key design variables—were investigated for their influence on three critical performance metrics: torque ripple, back-electromotive force total harmonic distortion (BEMF THD), and torque per rotor volume (TPRV). A limited set of finite element analysis (FEA) simulations was used to generate a sensitivity dataset, which served to train Gaussian Process Regression (GPR) surrogate models. These models enabled rapid motor performance prediction during optimization—without rerunning computationally intensive FEA—using Particle Swarm Optimization (PSO). The resulting Pareto front revealed the trade-offs between conflicting objectives and identified an optimal design region that satisfies practical engineering constraints: torque ripple ≤ 10% to reduce noise, vibration, and harshness (NVH), BEMF THD ≤ 5% to ensure smoother inverter operation and better control accuracy, and maximized TPRV to achieve high torque density while minimizing magnet material cost. The final design, validated through high-fidelity FEA, demonstrates a marked reduction in torque ripple and BEMF THD, along with a notable increase in TPRV. The proposed approach provides a computationally efficient design methodology for exploring rotor topology configurations, contributing to the design of compact, cost-effective, and high-performance electric traction motors.