Data-driven distributed photovoltaic power prediction

Abstract The accurate prediction of photovoltaic power is conducive to the accurate dispatch of the power system, and can provide support for the decision of peak and frequency modulation of the power grid. Based on the above considerations, this paper proposes a data-driven distributed photovoltaic power prediction method, which uses artificial neural network to clarify the mapping relationship between meteorological and temporal data and power data, thus improving the accuracy of power prediction through data classification. In convolutional network data-driven methods, power time series and meteorological time series data are organized in a 2D form.