An Empirical Study On The Prediction Of Short-Term Wind Speed In China Based On Intelligent Hybrid Model

With the rapid development of the economy,air pollution in various regions of China has become more and more serious,and the development of clean energy has become a consensus.Wind power is one of the most popular directions for developing clean energy.Therefore,high precision and reliable wind speed prediction become critical.Due to the randomness and instability of wind speed data,it is difficult to accurately predict wind speed.In order to improve the accuracy of wind speed prediction,this paper proposes a hybrid wind speed prediction model SSA_VMD_LSTM_ELM based on time series data processing,Long-Short Term Memory(LSTM)neural network and Extreme Learning Machine(ELM)to predict the wind speed in coastal areas of China.The main components of the proposed model include:(1)using Singular Spectrum Analysis(SSA)to extract trend information and residual information of wind speed data;(2)The extracted trend information and residuals are respectively decomposed into a series of sub-sequences using a Variational Mode Decomposition(VMD)method;(3)The LSTM network is used to predict the low frequency subsequences obtained by SSA_VMD;(4)The ELM is used to predict the high frequency subsequences obtained by SSA_VMD.For the purpose of verifying the predictive performance of the proposed model SSA_VMD_LSTM_ELM,nine models including the common models and the simplified models of SSA_VMD_LSTM_ELM are selected as contrasts,including EEMD_GWO_SVR,CEEMD_GWO_SVR,VMD_GWO_SVR,EEMD_CS_SVR,CEEMD_CS_SVR,VMD_CS_SVR,SSA_ELM,VMD_ELM,VMD_LSTM_ELM.According to the results of four data experiments,the newly proposed SSA_VMD_LSTM_ELM model shows the best prediction effect in all compared models on the wind speed prediction of four coastal cities in China;on the three evaluation indicators of RMSE,MAE and MAPE,it is better than other models,and the newly popsed model has improved at least 30 percent accuracy than other models.