Short-term Power Prediction Of Wind Farm Based On Wind Speed Decomposition

With the increasing influence of wind power on the power grid,accurate prediction of wind power is of great significance to the economic and stable operation of power systems.This paper proposes a short-term wind power prediction method based on wind speed decomposition.The main research contents and results are as follows:(1)An improved CEEMD decomposition method is proposed to decompose wind speed.Before the wind speed is decomposed,the correlation degree matching and mirror continuation processing are performed on the data to avoid the terminal flying phenomenon that occurs when the CEEMD decomposition is performed directly.Using actual data for simulation,the results show that the improved CEEMD decomposition can avoid terminal flying phenomenon without adding too much white noise,and there are fewer false components in the decomposition result.(2)An improved wind speed prediction model of CS-ELM is proposed to predict wind speed.The initial parameters of the ELM model can be optimized by the improved CS algorithm,thereby improving the prediction accuracy of the ELM model.The simulation results using actual data show that the improved CS-ELM model has higher prediction accuracy than the standard ELM model,and combined with the improved CEEMD decomposition algorithm can obtain better prediction results.(3)An improved wind speed and power mapping model of FA-LSSVM is proposed.By improving the FA algorithm,the internal parameters of the LSSVM model are optimized to improve the mapping accuracy of the LSSVM model.K-means clustering algorithm is used to classify training data and prediction data,and then establish mappingmodels for different types of data to predict power data.The simulation results using actual data show that both the improved FA algorithm and K-means clustering can improve the prediction accuracy of the mapping model.Combined with the improved CS-ELM model and the improved FA-LSSVM model,the final indirect power prediction model is obtained.Through simulation verification,the indirect power prediction model has higher prediction accuracy.