Smart Tuning of Wireless Power Transfer Systems with Neural Networks

This study presents a novel approach utilizing Neural Networks (NN) to estimate the distance between coils in a Wireless Power Transfer (WPT) system while implementing an automatic impedance matching mechanism. The methodology is validated using experimental data and simulation of the experimental setup to cover a wider range of configurations, achieving low predictive error rates. The NN is trained on impedance measurements and scattering parameters collected from the transmitter coil, enabling accurate distance estimation between the two coils. Furthermore, this framework facilitates the identification of optimal capacitance values to enhance circuit performance and achieve ideal impedance matching. The results indicate significant improvements in WPT efficiency, underscoring the potential of NN applications in optimizing wireless energy transfer systems.