No abstract

No abstract

No abstract

Solar energy is one of the longest-standing renewable in the world. It is available at no charge and could easily be exploited to reduce dependence on hydrocarbon-based energy. Solar radiation data plays an important role in the design, size, and efficiency of energy systems and renewable. However, these data are not available in all cases, particularly in isolated areas. Therefore, prediction of solar radiation values is often the only practical way to obtain such data. In fact, the measured sequences of radiation values are only available for some locations or regions in each country. Neural networks are classified among the techniques of artificial intelligence, its simulation of human reasoning characterizes it and neural networks have contributed to the development of several areas. In prediction, neural networks are used to solve complex forecasting problems. In this work, fifteen neural models are implemented for the modeling and prediction of global solar radiation over a horizontal surface using neural networks. The meteorological parameters are: duration of sunshine (S), hours of light (S0), total extraterrestrial solar radiation, temperature and humidity. Batna has been selected for this study, with its ten years (1996–2005) meteorological datasets collected from the HelioClim1 (HC1) database. The obtained results show that the news neuronal models are efficient for predicting daily global solar radiation with good measurement precision in this town.