Solar Radiation Forecasting Using Ad-Hoc Time Series Preprocessing and Neural Networks.

Christophe Paoli (1) , Cyril Voyant (1) , Marc Muselli (1, 2) , Marie Laure Nivet (1)

1 SPE - Sciences pour l'environnement
2 International Organization for Dew Utilization

In this paper, we present an application of neural networks in the renewable energy domain. We have developed a methodology for the daily prediction of global solar radiation on a horizontal surface. We use an ad-hoc time series preprocessing and a Multi-Layer Perceptron (MLP) in order to predict solar radiation at daily horizon. First results are promising with nRMSE < 21% and RMSE < 998 Wh/m². Our optimized MLP presents prediction similar to or even better than conventional methods. Moreover we found that our data preprocessing approach can reduce significantly forecasting errors.

Mots clés: Time Series; Preprocessing; Seasonality; Multi-Layer; Perceptron

Domaines: Informatique; Intelligence artificielle [cs.AI]; Réseau de neurones [cs.NE]; Energie électrique

Christophe Paoli, Cyril Voyant, Marc Muselli, Marie Laure Nivet. Solar Radiation Forecasting Using Ad-Hoc Time Series Preprocessing and Neural Networks.. International Conference on Intelligent Computing (ICIC 2009), Sep 2009, Ulsan, North Korea. pp.898-907, ⟨10.1007/978-3-642-04070-2_95⟩. ⟨hal-00438781⟩