Rapid Hotspot Temperature Prediction of Magnetic Couplers in WPT Systems via a Simulation-Based Data-Driven Model

The magnetic coupler is crucial for energy transfer in wireless power transfer (WPT) systems. With increasing demand for miniaturization and lightweight designs, higher power densities are required, which intensify thermal management challenges and raise the risk of thermal runaway. Therefore, accurately and rapidly predicting the hotspot temperature is essential for assessing cooling system performance and ensuring the safe operation of WPT systems. This paper proposes a method for predicting hotspot temperatures in magnetic couplers using a simulation-based data-driven model. By analyzing the nonlinear coupling effects of power losses, misalignment, and cooling conditions on the hotspot temperature, an electromagnetic–thermal coupled simulation model is developed to generate a dataset. A simulation-based data-driven model for temperature prediction is then created using support vector regression (SVR) optimized by a genetic algorithm. The proposed method enables the fast and accurate calculation of hotspot temperatures, demonstrating its broad applicability. An experimental prototype is constructed to validate the effectiveness of the method.