The Application of Deep Learning Techniques for Solar Power Forecasting

The future forecasting of the solar irradiation is becoming very essential. There are other elements (features) contribute to the forecasting process of the solar irradiation. These features are closely related to the amount of solar irradiation arriving from the sun. As the weather factors are related to each other in terms of influence, a wide range of features that are necessary to enter the i process are considered in this research. This paper investigate the effect of some atmospheric factors like Evapotranspiration and soil temperature using deep learning techniques, like LSTM algorithm. The results show that higher accuracy is achieved when new features related to solar irradiation were included in the forecasting process. To the best of authors' knowledge, these new features have not previously considered in literature. The RMSE value is 0.34 obtained with 16 features used in forecasting process.