Controller Design For Dc-Dc Boost Converter Using PI, State Feedback and Q Learning

Authors : Murat Erhan Çimen

Pages : 30-46

View : 70 | Download : 88

Publication Date : 2024-12-31

Article Type : Research Paper

Abstract :In recent years, the global decline in fossil fuel reserves and the alarming rise in greenhouse gas emissions have significantly heightened the need for renewable energy sources. This urgent shift towards sustainability has made the development and optimization of efficient energy systems a top priority for countries and communities worldwide. This study focuses on the modeling and control of a DC-DC boost converter, a critical component widely utilized in renewable energy applications such as solar panels, battery systems, and fuel cells. The research explores three control strategies: the conventional Proportional-Integral (PI) controller, state feedback controller with integral action, and Q-learning-based controller, which employs reinforcement learning principles. Comparative experiments were conducted in the Matlab/Simulink environment to evaluate the performance of each controller. Results demonstrate that the Q-learning controller outperformed the traditional methods in terms of performance metrics, including Integral Squared Error (ISE), Integral Absolute Error (IAE), and settling time, showcasing its potential for enhancing the efficiency and stability of renewable energy systems.
Keywords : Boost Converter, Q learning, PI Control, State Feedback