Power forecasting approach of PV plant based on similar time periods and Elman neural network

A forecasting model based on similar time period model and Elman neural network forecasting model is presented. Similar time period model divides the forecasting day and historical days into 4 parts, according to the real-time weather forecast, the best historical data were selected to match each time period. K-fold cross volition was used to do the structure selection and parameter optimization for the Elman neural network forecasting model to get the minimum error model. Tests were conducted and results indicate that the model in the paper has a better performance in forecasting the changing law of output power in the single weather day, in addition, it also offers a better performance for the complicated weather day.