Research on Multi-Step Prediction of Short-Term Wind Power Based on Combination Model and Error Correction

The instability of wind power poses a great threat to the security system, and accurate prediction is beneficial large-scale entry into grid. To improve accuracy prediction, short-term multi-step model with error correction proposed, which includes complete ensemble empirical mode decomposition adaptive noise (CEEMDAN), sample entropy (SE), improved beetle antennae search (IBAS) kernel extreme learning machine (KELM). First, CEEMDAN decomposes original sequences set stationary sequence components. Then, new components reconstructed according SE value each component reduce workload subsequent prediction. are respectively sent IBAS-KELM for obtained, predicted values obtained by adding two. Finally, added obtain final value. results actual farm data show that has outstanding advantages in high-precision evaluation indexes combined constructed this paper at least 34.29% lower MAE, 34.53% RMSE, 36.36% MAPE compared other models. decreased 30.43%, RMSE 29.67%, 28.57%, error-corrected three-step 55.60%, 50.00%, 54.17% uncorrected method significantly accuracy.