Reference wind farm selection for regional wind power prediction models - Mines Paris Accéder directement au contenu
Communication Dans Un Congrès Année : 2006

Reference wind farm selection for regional wind power prediction models


Short-term wind power forecasting is recognized today as a major requirement for a secure and economic integration of wind generation in power systems. This paper deals with the case of regional forecasting of wind power with a large number of wind farms involved. Due to the large amount of potentially available information and also because part of the wind farms may not be "observable", forecasting systems use input from selected “reference” wind farms to predict the total wind power. The paper studies the influence of the reference farms on the prediction accuracy and proposes a methodology for their selection, based on advanced statistical analysis of the spatial-temporal characteristics of wind generation. Keywords: regional forecasting, upscaling, reference farm selection, information, clustering the final objective. At a primary level the problem of variables selection can be simplified to a problem of wind farms selection. In this paper a study is conducted to evaluate the impact of input selection on regional forecasting model performance, and several input selection methods that can help in model setup are examined. The results of the proposed methodology are evaluated on a Danish case study of regional forecasting using a non-linear prediction model.
Fichier principal
Vignette du fichier
EWEC_2006_SIEBERT_KARINIOTAKIS.pdf (429.97 Ko) Télécharger le fichier
Origine : Fichiers éditeurs autorisés sur une archive ouverte

Dates et versions

hal-00526690 , version 1 (15-10-2010)


  • HAL Id : hal-00526690 , version 1


Nils Siebert, Georges Kariniotakis. Reference wind farm selection for regional wind power prediction models. European Wind energy conference, EWEC 2006, Feb 2006, Athènes, Greece. 10 p. ⟨hal-00526690⟩
166 Consultations
176 Téléchargements


Gmail Facebook X LinkedIn More