Short-Term Wind Power Prediction for Extreme Weather Scenarios Based on Tensor Autoregression Completion Algorithm and ARMA Error Correction

Le résumé fourni par la source

Wind power is a new energy source vigorously built in China. Improving the accuracy of wind power prediction is of great significance to improve the stability of power grid. In view of the current situation of small wind power prediction data and strong randomness in extreme weather scenarios, a short-term wind power prediction model for extreme weather scenarios with tensor auto regression completion algorithm and ARMA error correction is proposed, and studied through the scenario of extreme high temperature. First, conduct extreme high temperature scenario discrimination and data screening to obtain basic research data. Then, using the tensor kernel norm to capture the long-term trend of the time series, and the tensor autoregressive norm to capture the short-term trend of the time series, to complete the missing values in the wind power data. Afterwards, the long short-term memory (LSTM) neural network wind power prediction model is established, and the input features were screened by Pearson correlation analysis, and the prediction results are corrected by auto regressive moving average (ARMA) error correction method. Finally, based on the actual data of the power grid operation in Gansu Province, China, the simulation results verify the effectiveness of this method.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.