Ultra-Short-Term Wind Power Forecasting Based On Time Series Analysis

In wind power generation,the accurate prediction of wind power can reduce the spinning reserve capacity,and can provide the relevant departments with a scientific grid scheduling scheme,which can greatly improve the utilization rate of clean energy.However,due to the volatility and uncontrollable nature of wind power,the wind power data is nonlinear and non-stationary.The ARIMA model in time series can be modeled and predicted according to the time series of wind power data,but the accuracy of the model is decreased with the increase of the predicted step size.In this paper,the empirical mode decomposition(Empirical Mode Decomposition,EMD)and the ARIMA model in time series are combined to improve the accuracy.In the process of application of EMD algorithm,the algorithm will appear in the decomposition process of modal aliasing phenomenon,in order to solve this problem,this paper presents an improved EMD algorithm,built integrated improved empirical mode decomposition methods(Modified Ensemble Empirical Mode Decomposition,MEEMD).In order to further improve the wind power prediction accuracy,this paper integrates improved empirical mode decomposition method,ARIMA model and sample entropy(Sample Entropy SE)fusion model advantages,established a mixed model of the MEEMD-SE-ARMA algorithm.In this paper based on the original data of the non steady wind power ultra short term forecasting,MEEMD algorithm is used to decompose a plurality of wind power sequence will be obtained after decomposition of recombination using sample entropy,divided into a series of complex differences in fresh air power sub obvious sequence,so that air power sub sequences obtained close to stationary data;the ARIMA model uses the time series modeling and forecasting of each sub sequence obtained by air power,time series should be fully considered in the process of modeling the generalized autoregressive conditional heteroscedasticity model and Lagrange multiplier test,and established the corresponding ARIMA model;the prediction of wind power sequence obtained by superposition reconstruction finally,the prediction results.This paper establishes the ARIMA-GARCH prediction model,EMD-ARIMA model,EEMD-ARIMA model and MEEMD-SE-ARIMA model,and compares the average absolute error of the prediction model,finally proved that MEEMD-SE-ARMA hybrid algorithm can improve the establishment of ultra short term wind power prediction accuracy effectively.