An AI-Enabled Control for Dynamic Wireless Power Transfer

Abstract

Dynamic wireless power transfer (DWPT) is a nascent technology that brings forth great flexibility for charging electric vehicles (EV). For it to achieve fruition and to avoid power wastage, certain measures need to be in place to improve the power transmission. One of such measures is to use multiple transmitters and make the charging system to be aware of the location of the receiver thereby adapting power transmission to the receiver location. This is because, the change in relative location affects the mutual inductance which in turn, affects the amount of power that gets transmitted. This research presents a scheme where the charging system adapts to EV location and tunes its internal parameters to optimize the transmit power. Maximum power point tracking is employed to ensure optimal power transmission. First, the system uses a trained machine learning neural network to estimate the coupling coefficient between the transmitter and receiver. This parameter is then fed into two optimization algorithms; Jaya and Crow search algorithm. These algorithms along-side some predefined parameters would determine how much tuning certain circuit components need to go through to operate the circuit at the maximum power point. Two transmit coils were used and preliminary results show a transmit efficiency of 95% and at a distance of 200 mm. Results were also validated using LTSpice simulation software and they show the prospect of implementation for DWPT.