Short-Term Wind Power Prediction Technology Based On Intelligent Combination Mode

Wind energy is a kind of renewable energy with rapid development,wide application range and inexhaustible resources.With the increasing demand for wind power generation,the number of wind farms has gradually increased.However,the randomness and intermittency of wind energy have brought great challenges to the reliability of wind power grid connection and the safe and stable operation of the power system.Therefore,it is of great significance to improve the accuracy of power prediction.This paper proposes an improved combined wind power prediction model based on the research of existing prediction methods.The specific research content is as follows:1.Aiming at the strong randomness and instability of the wind power sequence,a complete ensemble empirical mode decomposition of adaptive noise is carried out.However,the high-frequency subsequence generated after decomposition still has strong randomness.The sample entropy is used to quantify the complexity of each subsequence,and the subsequence with larger entropy value is subjected to variational modal decomposition.The echo state network(ESN)model based on signal decomposition is established to predict each component separately,and the prediction results are superimposed into the final wind power prediction value.The effectiveness of the model is verified by simulation.Compared with the single ESN model,the mean absolute percentage error of the model is reduced by about 41.881%.2.In order to further reduce the prediction error,the grey wolf optimizer(GWO)algorithm is used to optimize the parameters of the ESN model.Aiming at the shortcomings of the standard GWO algorithm of low optimization accuracy and easy to fall into local extremum,the nonlinear convergence factor and the improvement of the combination with the beetle antennae search algorithm are carried out.The standard GWO algorithm,the standard particle swarm optimization algorithm and the improved GWO algorithm are used to optimize the four classical test functions respectively,which verifies that the improved GWO algorithm has higher optimization accuracy and is not easy to fall into the local optimum.A short-term wind power prediction model based on the improved GWO algorithm to optimize ESN is established to predict the wind power sequence after signal decomposition.The simulation results show that the mean absolute percentage error of the model is reduced by4.719% compared with the signal decomposition model.3.The algorithm proposed in this paper is verified by the prediction system platform.First,a set of wind power prediction system during the internship period is introduced.The model proposed in this paper is used to predict the short-term wind power of Jiangsu Yancheng Dafeng Phase I wind power plant.The degree of fitting of the prediction curve and the accuracy of error analysis verify the effectiveness of the model proposed in this paper in the prediction system platform.4.Finally,an economic dispatch model of the power system including wind farms is established.Based on the previous wind power prediction,the improved firefly algorithm is used to solve the model.The simulation obtains the optimal output combination of the unit,which saves the total cost of power generation and verifies the effectiveness of the proposed model and algorithm.