Data-Driven Electromagnetic–Thermal Coupling Optimization of Permanent Magnet Machines for UAVs

This paper presents a data-driven approach for achieving the optimal design of external permanent magnet synchronous machines (PMSMs) utilized in unmanned aerial vehicles (UAVs). The analysis is grounded in the subdomain (SD) model, which enables precise prediction of electromagnetic performances. In particular, this approach integrates thermal management to meet the rigorous thermal requirements of UAVs. The methodology is based on previously collected experimental data from a prototype machine, which is then employed to develop a mathematical model that correlates with temperature rise under varying loading conditions. Consequently, the proposed approach enables both electromagnetic and thermal analyses with remarkable precision and efficiency. Following that, the proposed model is subjected to social group optimization (SGO), an evolutionary algorithm that has been developed recently, to search for optimal solutions while satisfying the thermal constraints. This methodology, therefore, offers a dependable and practical solution for the development of lightweight, high-performance external PMSMs for UAVs, which are highly favored in the final design phase for mass production. Finally, the efficacy of the proposed model is verified by finite element (FE) and experimental results.