Cloud Effects on Photovoltaic Power Forecasting: Initial Analysis of a Single Power Plant Based on Satellite Images and Weather Forecasts

Solar power is becoming increasingly important as a source of renewable energy, and photovoltaic (PV) technology has become the key method of harnessing solar energy. Unlike conventional power plants, scheduling PV power production is not possible as it relies on the unpredictable nature of sunlight and local weather conditions. It is why accurate forecasts of power production are crucial for ensuring a stable and reliable supply of electricity. This study is an initial analysis of the effect of local cloud cover on solar production forecasting for Vis power plant. It was shown that even a crude representation of cloud mask images from EUMETSAT can greatly improve production forecasting in a best-case scenario. The model that integrated both solar irradiance and cloud images exhibited superior performance (up to 24%) compared to the model that solely relied on solar irradiance in the one-hour ahead forecast. These results show great promise in further expanding this research with more advanced algorithms and EUMETSAT products.