Predição da velocidade do vento a curto prazo via modelos matemáticos

• Autor(a) principal: Ferreira, Moniki Dara de Melo
• Data de Publicação: 2017
• Tipo de documento: Trabalho de conclusão de curso
• Idioma: por
• Título da fonte: Repositório Institucional da UFRN
• Texto Completo: https://repositorio.ufrn.br/handle/123456789/40279
• topic: Energia eólica; Wind power; Previsão híbrida; Wind speed; Velocidade do vento; Hybrid forecasting; CNPQ::CIENCIAS EXATAS E DA TERRA::GEOCIENCIAS::METEOROLOGIA::METEOROLOGIA APLICADA

This present work aims to predict wind speed hourly averaging from anemometric towers (at a height of 50 meters) located in Belo Jardim/PE and Camocim/CE. For this, the time-series models used are: Holt-Winters (HW), Artificial Neural Networks (ANN) and Hybrid. Observational data were also evaluated by the MERRA-2 reanalysis (state-of-the-art) at the same height of the towers. The results show that the hybrid model presented a better performance in relation to the others, including compared to evaluation with the MERRA-2. For example, in terms of statistical residues, RMSE and MAE were 0.91 and 0.62 m/s, respectively. In this way, hybrid models are a good method of forecasting wind speed data for wind generation.