Wind Speed Prediction Study: Practical Application of ANN to Energy Production In Brazilian Territory

Tipo
Artigo de evento
Data de publicação
2023
Periódico
2023 IEEE PES Innovative Smart Grid Technologies Latin America, ISGT-LA 2023
Citações (Scopus)
0
Autores
Oliveira M.B.
Vega-Garcia V.
Orientador
Título da Revista
ISSN da Revista
Título de Volume
Membros da banca
Programa
Resumo
© 2023 IEEE.This article provides a quick guide for users, using online tools, focusing on wind power generation and choosing the best region for installing wind turbines (WT). Highlighting topographic-map tools for terrain feasibility in the installation of WT and artificial neural networks (ANN) to understand climate data from the region under study and predict future values of the same data. To carry out the proof-of-concept, Python was used in this study with several neural network models. For this study, a climate database extracted from the National Institute of Meteorology and altitude records from the Brazilian Wind Atlas was used for the period from 2015 to 2020. The present study found that the best correlation results for climatic variables and forecasts were proportional to the increase in the database. Several ANN models were tested for wind speed predictions, highlighting a dense and convolutional model (multiple-output method).
Descrição
Palavras-chave
Assuntos Scopus
Climate data , Energy productions , On-line tools , Proof of concept , Time series. , Times series , Topographic map , Wind energy. , Wind power generation , Wind speed prediction
Citação