Short-term renewable electric energy generation forecast in the state of Ceará using prophet regression

• Autor(a) principal: Silva, Francisco Eduardo Mendes da
• Data de Publicação: 2022
• Outros Autores: Oliveira, Lincoln Moura de, Antunes, Fernando Luiz Marcelo, Sá Junior, Edilson Mineiro
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Research, Society and Development
• Texto Completo: https://rsdjournal.org/index.php/rsd/article/view/29579
• topic: Geração de Energia Elétrica; Energia Renovável; Modelos de previsão.; Generación de Electricidad; Energías Renovables; Modelos de pronóstico.; Electric Power Generation; Renewable Energy; Forecast models.

Brazil went through a period of energy crisis in the last year of 2021, due to low rivers that supply hydroelectric plants, being forced to activate thermal plants to supply electricity to the Brazilian population. This energy crisis brings several negative aspects, which can be avoided or partially avoided with the use of forecasts that can help in the decision making by the Electric Energy System Operators. Within this perspective, this work has as main objective to predict the generation of renewable electricity in the state of Ceará (CE) in a period of three days ahead, through the Prophet prediction model, an algorithm used on a large scale by the social network Facebook, using electricity generation data extracted from the website of the National System Operator (ONS). Data were collected from November 1, 2018 to March 1, 2021, totaling 852 measurements considering daily intervals. The forecasts were evaluated by the model evaluation metrics: RMSE, MSE and MAPE. The data was divided into 75% training data and 25% testing data. As a result, it was observed that the model obtained an error of 5.5% taking into account the MAPE metric.