Short-Term Wind Power Prediction Based on EMD-PSO-BiLSTM Combined Model
Wind power prediction is of great significance for the stable operation of wind power connected to the grid.In order to solve the problems of accuracy and model stability in wind power prediction,EMD-PSO-BiLSTM model is introduced.By empirical mode decomposition,the original wind power sequence is decomposed into a series of inherent mode functions to effectively capture multi-scale features in the data,and a prediction model is established for each mode sequence.In view of the excellent generalization ability of bi-directional long short-term memory neural network,a prediction model of each mode based on BiLSTM is studied.Further,particle swarm optimization algorithm was used to optimize BiLSTM parameters to solve nonlinear,high-dimensional and multi-modal problems of the model,and the optimal model of each modal component was obtained,and then the predicted value of wind power was obtained by summarizing the results of each modal component.Finally,the actual operation data of a wind farm in Hunan Province is taken as an example to verify that the EMD-PSO-BiLSTM model can effectively improve the short-term prediction accuracy of wind power.
NETL
NSTL
万方数据
