Deep reinforcement learning based solution for sustainable energy management in photovoltaic systems

Abstract

Photovoltaic solar cells (PSCs) have emerged as a practical solution to meet the growing energy demands, offering advantages such as high efficiency, flexibility, and adaptability to diverse materials and manufacturing methods. However, fully realizing the potential of solar energy requires the development of affordable and widely applicable solar cells. This article underscores the crucial role of PSCs in reliably addressing future energy needs. Despite their inherent benefits, there remains a significant challenge in improving the affordability and applicability of these solar cells. To tackle these challenges, the article introduces a novel approach called Sustainable Energy Management Architecture in Photovoltaic systems (SEMA), which integrates the robust capabilities of Deep Reinforcement Learning (DRL) and the Deep Q-Network (DQN) technique with Proximal Policy Optimization (PPO) methods. By employing this comprehensive approach, the goal is to enhance cost-effectiveness and promote broader adoption of PSCs. The novelty of the research lies in the integration of DQN with PPO. This underscores the importance of exploring innovative manufacturing techniques and materials to overcome current limitations and fully unlock the potential of solar energy. Sustained efforts in this direction are essential for advancing and implementing PSCs as a reliable and sustainable energy solution for the future. Continuous research and development will drive improvements in solar cells' affordability, applicability, and efficiency. By addressing these challenges and harnessing advanced techniques like DRL-DQN and PPO, the objective is to pave the way for widespread and cost-effective adoption of PSCs, ultimately contributing to a greener and more sustainable energy future.