Multi-Temporal-Spatial-Scale Electromagnetic-Thermal Coupling Analytical Modeling and Multi-Fidelity Optimization of Permanent-Magnet Vernier Machines

Permanent magnet vernier machines (PMVMs), owing to their simple structure and high torque density, exhibit significant advantages for the actuation of humanoid robot joints. However, the rich magnetic field harmonics within PMVMs can lead to rapid magnetic saturation and severe heating under overload conditions, complicating the modeling and optimization of PMVMs in practical operational scenarios and thus restricting their broader adoption in humanoid robots. This project aims to incorporate “ multi-temporal-spatial-scale electromagnetic-thermal coupling analytical modeling” and “multi-fidelity surrogate model-assisted optimization” into the modeling and optimization process of PMVMs, establishing an efficient electromagnetic-thermal coupled computation and optimization method to meet the high torque density and high overload performance requirements of humanoid robots’ joint motors. The project will: (1) research high-precision analytical models of PMVMs’ electromagnetic field affected by enriched magnetic field harmonics; (2) elucidate the electromagnetic-thermal coupling mechanism of PMVMs and study the corresponding multi-temporal-spatial-scale electromagnetic-thermal coupling computing method; (3) develop a multi-fidelity surrogate model based on electromagnetic-thermal coupling computation and investigate surrogate model-assisted parameter optimization methods under complex operating conditions; (4) complete the prototype manufacturing and testing to validate the effectiveness of the proposed methods. The research of this project is of significant value for advancing the application of PMVMs in humanoid robots with high-load and high-explosive task requirements.