The Method of Photovoltaic Power Forecast based on Seasonal Classification and Limit Learning Machine

Abstract The seasonal distribution characteristics of photovoltaic power plant output fluctuation are analyzed, and a short-term power forecasting method based on seasonal classification is proposed. Firstly, the seasonal distribution characteristics of photovoltaic output and its fluctuation are analyzed. Secondly, the forecasting model of photovoltaic output in different seasons is established by the Limit Learning Machine neural network. Finally, an empirical analysis is carried out by using photovoltaic output data. The results show that the seasonal classification method of short-term PV power forecast is better than the unclassified model.