The Solar Energy Forecasting Using LSTM Deep Learning Technique

Abstract

The future power systems will be characterized by high penetration or even dominance of renewable energy sources. This will present a challenge for the power system operators responsible for the coordination and distribution of electricity due to the direct dependence of the renewable energy sources on climatic changes. For its unpredictable and intermittent nature, the integration of the renewable energy sources imposes significant problems on the management of the grid and the balance between electricity consumption and production. Deep learning techniques are used to predict an accurate generated power of solar photovoltaic (PV) power plants. The various meteorological conditions are related to each other in terms of influence. In this paper, a wide range of features are considered in the forecasting process. From the results, the effect of the atmospheric factors on the forecasting process is not the same. Attention should be paid to these differences to increase the accuracy of the prediction. The root mean squared error (RMSE) and mean square error (MSE) are used for evaluation.