Artificial intelligence technique for estimating PV modules performance ratio under outdoor operating conditions - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue Journal of Renewable and Sustainable Energy Année : 2018

Artificial intelligence technique for estimating PV modules performance ratio under outdoor operating conditions

Résumé

In this paper, artificial neural networks (ANNs) have been used for the performance ratio modelling of four photovoltaic (PV) modules. The PV modules are selected from three different silicon technologies including one monocrystalline, two polycrystalline, and one micromorph (a-Si/μc-Si) modules. The adopted ANN architecture is a multilayer perceptron (MLP). The inputs of the ANN models are the solar irradiance on the PV module plane and air ambient temperature, while the output is the PV module performance ratio. It is shown that ANN models with three layers and five hidden neurons accurately model the performance ratio regardless of PV module technology. The results obtained from the ANN model are compared with those obtained from the five parameter model (L5P). The model comparison is done through two widely used forecasting errors: the root mean square error (RMSE) and the mean absolute percentage of error (MAPE). The values of both RMSE and MAPE are less than 0.02 for MLP based models and are about three to nine times lower than those obtained from the electrical model. It is also shown that the poor fit of the L5P model is due to the bad estimation of series and shunt resistances.
Fichier non déposé

Dates et versions

hal-01942718 , version 1 (03-12-2018)

Identifiants

Citer

Alain Tossa, Y. Soro, Y. Coulibaly, Y. Azoumah, Anne Migan-Dubois, et al.. Artificial intelligence technique for estimating PV modules performance ratio under outdoor operating conditions. Journal of Renewable and Sustainable Energy, 2018, 10 (5), pp.053505. ⟨10.1063/1.5042217⟩. ⟨hal-01942718⟩
71 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More