Solar radiation forecasting based on artificial neural networks optimized by genetic algorithm for energy management in smart grids

An accurate evaluation of solar energy production in power grids is a crucial aspect for the optimal exploitation of the available resources and for the effective integration of photovoltaic (PV) generators. This is especially advantageous in a smart grid context where a higher penetration of renewable energy sources is expected combined with distributed intelligence and information and communication technology (ICT) infrastructures. In this paper the nonlinear autoregressive network with exogenous input (NARX) is used to perform hourly solar radiation forecasting, according to a multi-step ahead approach. Temperature has been considered as the exogenous variable in the analysis. The NARX topology selection is supported by a combined use of two techniques: 1. a genetic algorithm (GA)-based optimization technique for the determination of the best weight set and 2. a method that determines the optimal network architecture by pruning it according to the Optimal Brain Surgeon (OBS) strategy. The considered variables are observed at hourly scale in a seven year dataset and the forecasting is done for several time horizons in the range from 8 to 24 hours-ahead.