To mitigate the influence brought about by the sudden decay of generation power due to the shading of the photovoltaic power station by cloud banks and improve very short-term prediction accuracy and the duration

a photovoltaic very short-term power predicting model is proposed by combining numerical weather forecast with ground-based cloud images. First

the model uses photovoltaic power very short-term predicting method under the approaching clear sky condition to predict the next four hours’ photovoltaic power generation. Then

the cloud banks data of numerical weather forecast and sudden decay due to the cloud shaded photovoltaic power plants is used to predict the sudden decay in the forthcoming four hours and for approaching correction to improve the accuracy of very short-term prediction of the photovoltaic power. The results obtained show that the proposed model is feasible and practical as an accurate method for photovoltaic power plants 0~4 h very short-term prediction. This work is supported by National High Technology Research and Development Program of China (863 Program) (No. 2011AA05A104) and State Grid Corporation of China (No. NY71-12-008).