Research On Short-term Power Prediction Of Wind Power Based On Extreme-Point Symmetric Mode Decomposition

As an economic,practical and safe renewable energy,wind energy has been widely concerned by human beings.The development of renewable energy has become an inevitable trend in the world.Because of the intermittence and randomness of wind speed,wind power has great fluctuation.Therefore,prediction technology is particularly important.Effective prediction can not only advance reasonable dispatch of integrated wind power grid,but also reduce the cost of wind farm operation and establish a supporting system.The development of facilities and related industries will enhance the market and social benefits of wind power.In the paper,a hybrid prediction model based on ESMD is proposed to predict the short-term active power.Firstly,the original data of wind power is preprocessed,and the meteorological data indicators at different altitudes are screened by correlation analysis.Secondly,the ADF test and the BDS test are used to test wind power and meteorological data indicators,and the ESMD is used to decompose wind power step by step into several stationary intrinsic mode components(IMF)and trend(R).Thirdly,the sample entropy method is used to reconstruct the new components according to the principle of similarity of features.Lastly,Grey Wolf algorithm(GWO)is used to optimize the connection weights and thresholds of Extreme Learning Machine(ELM).The GWO-ELM model is constructed to predict the new components separately and superimpose the predicted values to get the final predicted values.The validity and practicability of the model proposed in this paper are verified by empirical research.In this paper,the power data and meteorological data of Baihe wind farm from April 1 to April 30 are used as empirical data samples.The results show that the data series of wind power does not obey the normal distribution and has the characteristics of non-linearity and non-stationary.Wind speed,wind direction,air pressure,temperature and humidity in 10 meters above sea level have the highest correlation with wind power.Statistical analysis of several components from frequency and amplitude shows that the curve component with weak fluctuation of frequency index has a larger amplitude proportion in the original sequence,and the IMF7 and R should be effectively predicted.The prediction performance of ESMD-GWO-ELM hybrid model proposed is better than that of single-form model,because ESMD reduces the dimension of modeling and prediction,and improves the prediction accuracy.In addition,GWO-ELM model has faster convergence and stronger generalization than single ELM,so the proposed model is suitable for short-term wind power prediction.