Short-term wind forecasting for energy resources scheduling

• Autor(es): Ramos, Sérgio ; Soares, João ; Silva, Marco ; Vale, Zita
• Data: 2012
• Identificador Persistente: http://hdl.handle.net/10400.22/1444
• Origem: Repositório Científico do Instituto Politécnico do Porto
• Assunto(s): Artificial neural network; Wind power and speed forecasting; Short - term forecasting; Short - term scheduling
• Tipo de Documento: Documento de conferência
• Idioma: Inglês

Descrição

This paper proposes a wind power forecasting methodology based on two methods: direct wind power forecasting and wind speed forecasting in the first phase followed by wind power forecasting using turbines characteristics and the aforementioned wind speed forecast. The proposed forecasting methodology aims to support the operation in the scope of the intraday resources scheduling model, namely with a time horizon of 5 minutes. This intraday model supports distribution network operators in the short-term scheduling problem, in the smart grid context. A case study using a real database of 12 months recorded from a Portuguese wind power farm was used. The results show that the straightforward methodology can be applied in the intraday model with high wind speed and wind power accuracy. The wind power forecast direct method shows better performance than wind power forecast using turbine characteristics and wind speed forecast obtained in first phase.