Neural Network Based Mutual Inductance Estimation for Maximum Power Point Tracking in Wireless Power Transfer Array

Dynamic Wireless Power Transfer (DWPT) requires that participating coils and circuitry respond to change in receiver location. This change directly affects the mutual inductance between the coils. In turn, the amount of power transmitted is directly affected. To maintain desired level of power transmission, it is imperative to estimate the mutual inductance and implement a maximum power point tracking scheme. Eliminating the need for communication between transmitter and receiver would reduce the overall response time of the system. In this paper, a neural network-based mutual inductance estimation scheme is presented. Location data of receiver is detected by transmitter and used to estimate the mutual inductance. The estimated value is subsequently utilized to drive the circuit to maximum power point. To do this, two optimization algorithms, Jaya and Crow search, are employed. Results are verified using LTSpice simulation software. In order to implement the presented algorithm in hardware, further optimization of algorithm hyper-parameters still needs to be carried out.