A Two-Modal Weather Classification Method and Its Application in Photovoltaic Power Probability Prediction
Weather classification is an indispensable preprocessing step in photovoltaic power prediction.A new photovoltaic power clustering based two-modal weather classification method was proposed to finely depict the uncertainty of photovoltaic power.Both photovoltaic power data and meteorological data were considered for weather classification,which provided a novel and effective path for weather classification based photovoltaic power prediction.In addition,data fusion technology was used to mine relevant information between numeric weather prediction(NWP)data and measured meteorological data to help for weather classification,so as to reduce the model reliance on the accuracy of forecasted meteorological indicators as well as improve the robustness of the model.In experiments based on the data of a photovoltaic power station in Jilin,the rationality of the weather classification method was demonstrated.The photovoltaic power probability prediction combined with the proposed weather classifier resulted in the prediction interval coverage probability closer to the preassigned confidence level,and a narrower mean prediction interval width.
photovoltaic power generationweather classificationphotovoltaic power probability predictiontime series K-means clusteringmulti-modal learninguncertainty