Forecasting of Photovoltaic Power Generation by RBF Neural Networks

Abstract

Recent studies suggest that in order to facilitate higher market and grid penetration of solar power, the users need accurate forecasts of generating power from photovoltaic (PV) plants on multiple time horizons. Despite the large number of forecasting methods, the comparison of results and evaluation of relative advantages between models has been evasive. The general purpose of the paper is to explore the way of performing accurate forecasts of generating power from renewable energy sources so that independent system operators can act consequently. Different aspects of radial basis functions (RBF) neural networks (NNs) are discussed and an illustration of the proposed predictor software interface is given.