3 - Forecasting of solar radiation using machine learning and deep learning algorithms

Abstract

This chapter covers four applications: the first application is about one-step ahead forecasting of the daily global horizontal irradiation (GHI), using machine learning methods. The second application is related to one-step ahead forecasting of in-plane solar irradiance using deep learning neural networks. The third application concerns the investigation of deep learning natural networks for multistep ahead forecasting of in-plane solar irradiance. In the last application, emphasis is given to the use of neural networks for one-step ahead forecasting of daily and hourly GHI using meteorological parameters. In all, 16 examples have been described in order to help readers understand how to apply machine learning and deep learning for the forecasting of solar radiation. Four databases have been used to develop different models and all models presented have been written and developed using Python programming language.