Research on Photovoltaic Power Prediction Based on BP Neural Network Algorithm

Abstract

The output power of photovoltaic (PV) power plant is influenced by the fluctuation of meteorological conditions, which, coupled with the intermittent nature of the power output, has raised concerns about the stability of power grid. According to the variations in meteorological conditions, PV power prediction technology can predict the PV output power in advance to decrease the influence of PV power generation on the stable operation of the power grid. In this paper, the PV power prediction model is built on Matlab simulation software with a BP neural network algorithm. After that, the power prediction calculation of a PV power station is carried out under the operation conditions of sunny and cloudy days. The results show that, in the periods of prediction, the power curve of the predicted value is close to the power curve of the real value, which demonstrates the great accuracy of the BP prediction model established in this paper.