A Robust Primary-Side Hybrid Data-driven Load Monitoring Strategy for Wireless Power Transfer Systems

Conventional front-end load monitoring strategies for wireless power transfer (WPT) systems are designed based on circuit models, which may not be accurate when the system parameters drift or measurement errors occur. In this paper, a primary-side hybrid data-driven load monitoring strategy is proposed to address this issue, while still maintaining the merit of non-communication between the transmitter and receiver. The hybrid data includes both simulation and experimental data. The simulation data can be rapidly acquired in large numbers and the experimental data can reflect disturbance. The hybrid data are used for training a Multilayer Perceptron (MLP) with one input layer, six hidden layers, and one output layer. The proposed MLP is validated to be more accurate than the conventional circuit models for monitoring load conditions of practical WPT systems with disturbance.