Solar Irradiation Forecast for Solar Photovoltaic Systems Using Machine Learning. A Supervised Learning Approach to Forecast Solar Irradiation in the State of São Paulo, Brazil

• Grant number: 20/09607-9
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: September 01, 2020
• End date: August 31, 2021
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Renato Fernandes Cantão
• Grantee: Enzo Laragnoit Fernandes
• Host Institution: Centro de Ciências e Tecnologias para a Sustentabilidade (CCTS). Universidade Federal de São Carlos (UFSCAR). Sorocaba , SP, Brazil

Abstract

With the increasing demand for energy different forms of electrial energy yield arise. Among them there is the photovoltaic solar energy, a renewable, abundant and low cost energy source whose applications deppend directly on the solar irradiation intensity. In this project one proposes the use of Supervised Machine Learning techniques in order to implement models capable of predicting solar irradiation from a known dataset in especific sites within the State of São Paulo, Brazil. Once such dataset is obtained different techniques will be applied aiming to carry the preprocessing and Feature Engineering steps to optimize the Supervised Machine Learning algorithms. Models' capability of prediction will be evaluated in two moments: during the adjustment stage and, subsequently, in the validation stage. Results will be tabulated and analised regarding the models' hyperparameters, the difference between predicted and observed values and error metrics. Implementation will be carried out using the Python programming language in addition to diverse machine learning, data manipulation and visualization libraries.