AUTOREGRESSIVE NEURAL NETWORK MODELS FOR SOLAR POWER FORECASTING OVER NIGERIA

Document Type : Original Article

Authors

1 Federal University of Technology Akure

2 Federal University of Technology, Akure, Nigeria

Abstract

In this study, the nonlinear autoregressive neural network with exogenous input (NARX) model was employed to predict solar power in different geoclimatic zones of Nigeria using six solar radiation parameters. The solar power was first deduced using the surface direct and diffuse solar radiation data obtained from the archives of the Modern-Era Retrospective Analysis for Research and Application, Version 2, over 20 stations spread across Nigeria. NARX model was then created and trained using Levenberg-Marquardt (LM), Bayesian regularization (BR), scaled conjugate gradient (SCG), and Broyden–Fletcher–Goldfarb–Shanno (BFGS) algorithms, and the values were compared to the calculated values of the solar power. The performance of the four algorithms were assessed using standard evaluation metrics. Error analyses showed that all the algorithms had desirable performances with root mean square error (RMSE) values ranging from 0.162 to 0.544 W/m2. Regionally, the NARX-BFGS model had the best performance in the Coastal and Guinea Savanna zones, whereas the NARX-LM and NARX-BR models had the best performances in the Sahel and Derived Savanna zones, respectively. The results of this study will assist solar engineers in calibrating the performance of solar conversion systems for the future planning of sustainable renewable energy policies.

Keywords

Journal of Solar Energy Research
Volume 7, Issue 1
January 2022
Pages 983-996

Files

Share

How to cite

Statistics