Optimized Inductance Method Based on Neural Networks for Wireless Power Transfer Applications in Implantable Medical Devices

Wireless Power Transfer (WPT) is a trendy technology that enables the charging of batteries without physical connection. Medical implants have benefited from this technology, and consequently rechargeable devices can be implemented, improving the patient's quality of life. In this work, a method to optimize multi-layer planar squared coils for a WPT application is proposed. This method involves the utilization of Artificial Neural Networks (ANN), which predict the efficiency of a system from its geometrical parameters. This ANN is trained by a dataset, which is obtained through Finite Elements Analysis (FEA). The proposed method reduces the computational cost of a pure FEA optimization, providing accurate results.