AI-Based Approach for Exploring Power Electronics Employing Reinforcement Learning

Abstract

We discuss a reinforcement learning (RL) algorithm to create a general current control structure for a converter based on a Deep Neural Network (DNN). After an introduction of artificial intelligence (AI) methods in general and especially in power electronics research, we discuss an RL algorithm which learns the design of a current controller for an AC/DC PFC rectifier without any user-provided input and knowledge. The RL algorithm tries to optimize the DNN-based controller for minimum switching losses, minimum low-order harmonics, and THD of the grid current. We implement this DNN-based controller in a circuit simulator, analyze controller performance in the time domain, compare with state-of-the-art current controllers, and evaluate converter losses.