Classificação do tipo de sombreamento e de sujidade em módulos fotovoltaicos utilizando LGBM e curvas I-V

Abstract

With the growing global demand for electricity, it becomes essential to invest in clean and renewable energy production. Brazil, with its vast grographic potential, has the opportunity to diversify its energy matrix with renewable sources, bringing economic, social, and environmental benefits. Investing in renewable energy production is crucial for creating a sustainable matrix that can efficiently meet consumption needs. Reducing costs for solar energy operation and maintenance is critical to ensure the economic viability of photovoltaic projects. Efficient management of these costs can make the difference between a successful project and an unprofitable installation, especially considering the increase in the installed solar system capacity. This work proposes the structuring of a dataset using I-V curves from a photovoltaic array under different shading and dirtiness conditions, aiming for reducing maintenance costs. The dataset used contains measurements of current, voltage, module temperature and irradiation at the time of sampling. The model used for evaluation and comparison is the Light Gradient-Boosting Machine, which achieved an accuracy of 99.1% in dirtiness evaluation and 96.5% in shading evaluation. Results from other models were also compared, reaching an accuracy of 100% in dirtiness evaluation.