The state-of-the-art in short term prediction of wind power from a danish perspective - Mines Paris Accéder directement au contenu
Communication Dans Un Congrès Année : 2003

The state-of-the-art in short term prediction of wind power from a danish perspective

Résumé

This presentation is based on a longer report trying to summarise more than 100 papers written in the field over the last decades. Many regions have nowadays such high penetrations of wind energy (the host region Western Denmark has in the order of 25%) that without good short-term prediction tools, an economic and secure integration of wind power with maximal ecological benefits of the wind power could not be realised. A historical perspective will lead to an account of the current crop of models, including to a high degree the experiences made in Denmark with operative use of the tools since 1994. Three horizons are interesting for utilities: a short horizon determined by the ramping and start-up times of conventional power plants for the scheduling (4-8 hours ahead), a longer horizon dealing with the trading on the different electricity exchanges (in the case of NordPool, 13-37 hours ahead), and a long horizon where the models could be used for maintenance planning (all the way to weeks ahead). For the first case, one could get away with using a time-series analysis model coupled to climatology, but for even medium horizons, the accuracy of the model is getting much better by using numerical weather prediction, typically from the local meteorological institute. The most prominent source of error is the numerical weather prediction model used, and in that it is phase errors (timing errors) that have a decisive impact on the traditional error scores, and on the financial bottom line. Current work is trying to estimate the uncertainty of the forecasts. The longer report has been prepared in the framework of work for the EU Commission, especially a Marie-Curie-Fellowship and the ANEMOS project.
Fichier principal
Vignette du fichier
paper_Billund_2003.pdf (579.94 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-00529986 , version 1 (06-02-2018)

Identifiants

  • HAL Id : hal-00529986 , version 1

Citer

Gregor Giebel, Georges Kariniotakis, Richard Brownsword. The state-of-the-art in short term prediction of wind power from a danish perspective. 4th International Workshop on large scale integration of wind power and transmission networks for offshore wind farms, Oct 2003, Billund, Denmark. ⟨hal-00529986⟩
290 Consultations
1002 Téléchargements

Partager

Gmail Facebook X LinkedIn More