Assessment of artificial neural networks for hourly solar radiation prediction

Tamer Khatib, Azah Mohamed, Kamaruzzaman Sopian, M. Mahmoud

Research output: Contribution to journalArticle

26 Citations (Scopus)

Abstract

This paper presents an assessment for the artificial neural network (ANN) based approach for hourly solar radiation prediction. The Four ANNs topologies were used including a generalized (GRNN), a feed-forward backpropagation (FFNN), a cascade-forward backpropagation (CFNN), and an Elman backpropagation (ELMNN). The three statistical values used to evaluate the efficacy of the neural networks were mean absolute percentage error (MAPE), mean bias error (MBE) and root mean square error (RMSE). Prediction results show that the GRNN exceeds the other proposed methods. The average values of the MAPE, MBE and RMSE using GRNN were 4.9, 0.29 and 5.75, respectively. FFNN and CFNN efficacies were acceptable in general, but their predictive value was degraded in poor solar radiation conditions. The average values of the MAPE, MBE and RMSE using the FFNN were 23, -.09 and 21.9, respectively, while the average values of the MAPE, MBE and RMSE using CFNN were 22.5, -19.15 and 21.9, respectively. ELMNN fared the worst among the proposed methods in predicting hourly solar radiation with average MABE, MBE and RMSE values of 34.5, -11.1 and 34.35. The use of the GRNN to predict solar radiation in all climate conditions yielded results that were highly accurate and efficient.

Original languageEnglish
Article number946890
JournalInternational Journal of Photoenergy
Volume2012
DOIs
Publication statusPublished - 2012

Fingerprint

solar radiation
Solar radiation
Neural networks
root-mean-square errors
Mean square error
predictions
Backpropagation
climate
cascades
topology
Topology

ASJC Scopus subject areas

  • Renewable Energy, Sustainability and the Environment
  • Chemistry(all)
  • Atomic and Molecular Physics, and Optics
  • Materials Science(all)

Cite this

Assessment of artificial neural networks for hourly solar radiation prediction. / Khatib, Tamer; Mohamed, Azah; Sopian, Kamaruzzaman; Mahmoud, M.

In: International Journal of Photoenergy, Vol. 2012, 946890, 2012.

Research output: Contribution to journalArticle

@article{5d6aea9bedcd435793bde29140f2a84f,
title = "Assessment of artificial neural networks for hourly solar radiation prediction",
abstract = "This paper presents an assessment for the artificial neural network (ANN) based approach for hourly solar radiation prediction. The Four ANNs topologies were used including a generalized (GRNN), a feed-forward backpropagation (FFNN), a cascade-forward backpropagation (CFNN), and an Elman backpropagation (ELMNN). The three statistical values used to evaluate the efficacy of the neural networks were mean absolute percentage error (MAPE), mean bias error (MBE) and root mean square error (RMSE). Prediction results show that the GRNN exceeds the other proposed methods. The average values of the MAPE, MBE and RMSE using GRNN were 4.9, 0.29 and 5.75, respectively. FFNN and CFNN efficacies were acceptable in general, but their predictive value was degraded in poor solar radiation conditions. The average values of the MAPE, MBE and RMSE using the FFNN were 23, -.09 and 21.9, respectively, while the average values of the MAPE, MBE and RMSE using CFNN were 22.5, -19.15 and 21.9, respectively. ELMNN fared the worst among the proposed methods in predicting hourly solar radiation with average MABE, MBE and RMSE values of 34.5, -11.1 and 34.35. The use of the GRNN to predict solar radiation in all climate conditions yielded results that were highly accurate and efficient.",
author = "Tamer Khatib and Azah Mohamed and Kamaruzzaman Sopian and M. Mahmoud",
year = "2012",
doi = "10.1155/2012/946890",
language = "English",
volume = "2012",
journal = "International Journal of Photoenergy",
issn = "1110-662X",
publisher = "Hindawi Publishing Corporation",

}

TY - JOUR

T1 - Assessment of artificial neural networks for hourly solar radiation prediction

AU - Khatib, Tamer

AU - Mohamed, Azah

AU - Sopian, Kamaruzzaman

AU - Mahmoud, M.

PY - 2012

Y1 - 2012

N2 - This paper presents an assessment for the artificial neural network (ANN) based approach for hourly solar radiation prediction. The Four ANNs topologies were used including a generalized (GRNN), a feed-forward backpropagation (FFNN), a cascade-forward backpropagation (CFNN), and an Elman backpropagation (ELMNN). The three statistical values used to evaluate the efficacy of the neural networks were mean absolute percentage error (MAPE), mean bias error (MBE) and root mean square error (RMSE). Prediction results show that the GRNN exceeds the other proposed methods. The average values of the MAPE, MBE and RMSE using GRNN were 4.9, 0.29 and 5.75, respectively. FFNN and CFNN efficacies were acceptable in general, but their predictive value was degraded in poor solar radiation conditions. The average values of the MAPE, MBE and RMSE using the FFNN were 23, -.09 and 21.9, respectively, while the average values of the MAPE, MBE and RMSE using CFNN were 22.5, -19.15 and 21.9, respectively. ELMNN fared the worst among the proposed methods in predicting hourly solar radiation with average MABE, MBE and RMSE values of 34.5, -11.1 and 34.35. The use of the GRNN to predict solar radiation in all climate conditions yielded results that were highly accurate and efficient.

AB - This paper presents an assessment for the artificial neural network (ANN) based approach for hourly solar radiation prediction. The Four ANNs topologies were used including a generalized (GRNN), a feed-forward backpropagation (FFNN), a cascade-forward backpropagation (CFNN), and an Elman backpropagation (ELMNN). The three statistical values used to evaluate the efficacy of the neural networks were mean absolute percentage error (MAPE), mean bias error (MBE) and root mean square error (RMSE). Prediction results show that the GRNN exceeds the other proposed methods. The average values of the MAPE, MBE and RMSE using GRNN were 4.9, 0.29 and 5.75, respectively. FFNN and CFNN efficacies were acceptable in general, but their predictive value was degraded in poor solar radiation conditions. The average values of the MAPE, MBE and RMSE using the FFNN were 23, -.09 and 21.9, respectively, while the average values of the MAPE, MBE and RMSE using CFNN were 22.5, -19.15 and 21.9, respectively. ELMNN fared the worst among the proposed methods in predicting hourly solar radiation with average MABE, MBE and RMSE values of 34.5, -11.1 and 34.35. The use of the GRNN to predict solar radiation in all climate conditions yielded results that were highly accurate and efficient.

UR - http://www.scopus.com/inward/record.url?scp=84864936175&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84864936175&partnerID=8YFLogxK

U2 - 10.1155/2012/946890

DO - 10.1155/2012/946890

M3 - Article

VL - 2012

JO - International Journal of Photoenergy

JF - International Journal of Photoenergy

SN - 1110-662X

M1 - 946890

ER -