A novel decomposition-ensemble prediction model for ultra-short-term wind speed

Accurate ultra-short-term wind speed prediction is of great significance to the power generation efficiency of wind farms, and also has a good application prospect in the field of meteorology. In this study, a novel prediction model based on empirical mode decomposition and improved sparrow search algorithm optimized reinforced long short-term memory neural network is proposed. The implementation process of the proposed prediction model is: (a) the ultra-short-term wind speed is decomposed by empirical mode decomposition algorithm. (b) the decomposed components are predicted by reinforced long short-term memory neural network. (c) the hyperparameters of long short-term memory neural network are optimized by designed improved sparrow search algorithm. (d) the predicted values of each reinforced long short-term memory model are superimposed to get the final predicted value. The actual ultra-short-term wind speed data is taken as the research object. R2, RMSE, RRMSE, MRMSE, MAE, MAPE, TIC, IA, Pearson's test, prediction error distribution box-plot, and Taylor diagram are used to judge the performance indexes of the designed prediction model. In case studies, the performance of the designed prediction model is compared with standard LSTM, EMD and standard LSTM, EMD and standard SSA optimized reinforced LSTM, and other state-of-the-art models. The results show that the proposed ultrashort-term wind speed prediction model achieves the best prediction performance and prediction accuracy.

Automated summary

A novel wind speed prediction model based on empirical mode decomposition and improved sparrow search algorithm optimized reinforced LSTM was proposed. Through optimization, prediction, and superimposition steps, the model achieved the best prediction performance and accuracy.