Performance Enhancement of Buck Converter Using Reinforcement Learning Control
P S V Kishore, Jayaram Nakka, Jami Rajesh
Abstract
Closed-loop control of DC-DC converter is necessary for many applications such as EV charging circuits, energy storage systems, and household applications. DC-DC converters are extremely sensitive to changing load conditions. Advanced control system promises the enhanced performance of these converters. In this paper, a reinforcement learning based Deep Deterministic Policy Gradient (DDPG) algorithm has been proposed to control the output voltage of the buck converter. The duty cycle is considered as an output of the agent in DDPG. This duty cycle is used to generate the pulse to the switch of the buck converter. The proposed control scheme is designed and simulated in a MATLAB/Simulink environment and the effectiveness of this method is verified by changing the reference voltage of buck converter.