IMPLEMENTATION OF THE RADIAL BASIS FUNCTION NEURAL NETWORK METHOD IN PREDICTING SHORT-TERM POWER GENERATED BY PHOTOVOLTAIC

Abstract

The adoption of photovoltaic (PV) technology in Indonesia offers significant economic and environmental benefits, such as job creation and greenhouse gas reduction. Nevertheless, there are major technological hurdles to overcome, particularly the inherent variability and intermittency of solar energy, as well as severe financial limitations and insufficient policy frameworks, which prevent widespread adoption. Accurate forecasting of PV power is necessary in order to overcome these technical obstacles and maintain grid stability. The Radial Basis Function Neural Networks (RBF-NN) are the primary focus of this study, which assesses cutting-edge prediction techniques using neural networks. The complicated nonlinear mappings used in predicting PV power production are efficiently approximated by this approach using localized receptive fields. With a root mean square error (RMSE) of 3.244, this method produced the best forecasting simulation results in this analysis.