A Short-Term Photovoltaic Power Output Prediction for Virtual Plant Peak Regulation Based on K-means Clustering and Improved BP Neural Network

摘要

In order to formulate a reasonable scheduling plan of virtual power plant (VPP), a prediction method of photovoltaic (PV) output based on K-means and improved BP neural network is proposed. Firstly, the structure of virtual plant for peak regulation is introduced. Then, the historical data of PV is clustered by K-means to distinguish different weather conditions. To improve the prediction accuracy, genetic algorithm (GA) is used to improve the BP neural network. Finally, a short-term prediction model based on improved BP neural network is established in Matlab. The simulation results show that using clustered photovoltaic data and improved BP neural network to predict the output of PV on similar days has a higher prediction accuracy.