Short-Term load forecasting using a neuro-fuzzy model based on entropy maximisation

Abstract : The paper presents a new short-term load forecasting approach based on dynamic fuzzy logic modelling. The developed model produces forecasts for the next 48 hours, which are updated every hour. Such a sliding window scheme is different than conventional models that operate usually once a day. The paper emphasizes on developing appropriate learning and on-line adaptation schemes based on the maximal entropy principle. In contrast to the traditional approach, such schemes permit to avoid overfitting of the model to the data. Thus, the ability of the model to predict new data (generalisation) is maximized. The architecture of the model is selected using non-linear optimisation techniques such the non-linear Simplex. The model has been developed in the frame of the EU research project More-Care and implemented for on-line use at the islands of Crete and Madeira. Results from the case studies are presented showing the efficiency of the approach.
Type de document :
Communication dans un congrès
Liste complète des métadonnées

Littérature citée [4 références]  Voir  Masquer  Télécharger

https://hal-mines-paristech.archives-ouvertes.fr/hal-00534191
Contributeur : Corinne Matarasso <>
Soumis le : vendredi 4 mai 2018 - 10:22:40
Dernière modification le : jeudi 30 août 2018 - 18:54:02

Fichier

MED02-071a.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-00534191, version 1

Citation

Georges Kariniotakis. Short-Term load forecasting using a neuro-fuzzy model based on entropy maximisation. Med Power 2002, Nov 2002, Athènes, Greece. ⟨hal-00534191⟩

Partager

Métriques

Consultations de la notice

67

Téléchargements de fichiers

18