Solar Radiation Prediction on Photovoltaic Systems Using Machine Learning Techniques

Estimation of solar radiation is essential to help decision-makers in the planning of isolated solar energy farms or connected to electricity distribution networks to take advantage of renewable energy sources, reduce the impact produced by climate change, and increase coverage rates in electricity service. The number of existing measurement stations is insufficient to cover the entire geography of a region, and many of them are not capturing solar radiation data. Therefore, it is important to use mathematical, statistical, and artificial intelligence models, which allow predicting solar radiation from meteorological data available. In this work, datasets taken from measurement stations located in the cities of Cali and Villavicencio were used, in addition to a dataset generated by the World Weather Online API for the town of Mocoa, to carry out solar radiation estimations using different machine learning techniques for regression and classification to evaluate their performance. Although in most related works researchers used deep learning to predict solar radiation, this work showed that, while artificial neural networks are the most widely used technique, other machine learning algorithms such as Random Forest, Vector Support Machines and AdaBoost, also provide estimates with sufficient precision to be used in this field of study.

• Fecha de publicación: 2020
• Institución: Universidad Pedagógica y Tecnológica de Colombia
• Repositorio: RiUPTC: Repositorio Institucional UPTC
• Idioma: spa
• OAI Identifier: oai:repositorio.uptc.edu.co:001/14284
• Acceso en línea: https://revistas.uptc.edu.co/index.php/ingenieria/article/view/11751; https://repositorio.uptc.edu.co/handle/001/14284
• Palabra clave: deep learning; machine learning; photovoltaic systems; prediction model; solar radiation; supervised learning; aprendizaje automático; aprendizaje profundo; aprendizaje supervisado; modelo de predicción; radiación solar; sistemas fotovoltaicos
• License: http://purl.org/coar/access_right/c_abf153