Design of an Intelligent Layer for Improving Performance of Solar Energy System

Abstract

The escalating demand for sustainable and renewable energy sources has positioned so- newlinelar energy at the forefront of global initiatives to combat climate change and foster a clean energy future. Significantly, the inexhaustible and environmentally friendly nature of solar energy makes it a key player in addressing the growing energy demand while mitigating the adverse effects of conventional fossil fuel sources. The significance of harnessing solar power newlinecan not be overstated as the world shifts towards a more sustainable sources of energy. India, recognizing the potential of solar energy, has set ambitious targets to optimize its solar resources, striving to achieve 175 GW of solar capacity by 2023 and emphasizing solar pivotal role in the nation s energy mix. The thesis main focus is to design an intelligent layer for improving the performance of solar energy systems. The research delves into the application of newlinemachine learning models to enhance the efficiency and performance of solar energy systems. newlineBy leveraging bio-inspired algorithms and predictive analytic, The proposed model aims to optimize duty cycle and passive component values, thus enhancing the Solar energy systems overall efficiency. The insights gained from this research can pave the way for innovations in solar technology, fostering a sustainable energy transition and contributing to the global quest for cleaner and more resilient energy solutions. The thesis proposes the design of a ma- newlinechine learning-based single ended primary inductor Converter (SEPIC) model that produces high efficiency by optimizing inherent factors, duty cycle, and inductor values using genetic algorithm (GA) and particle swarm optimization (PSO). Additionally, a Q-learning model con- newlinetinually monitors the SEPIC circuit performance for enhanced efficiency. Introducing a novel hybrid soft computing model for passive components selection in Zeta converters. The model employs PSO for initial component ratings and enhances them using grey wolf optimization (GWO) for impro