| With the rapid development of the economy and the massive consumption of fossil energy,environmental pollution and climate change have gradually become major challenges in China.Vigorously developing clean energy and gradually reducing the proportion of fossil energy consumption is the key to China’s efforts to improve the environment and achieve sustainable energy important step in development.As clean,efficient,and widely distributed renewable energy,wind energy plays an important role in the process of energy transformation and upgrading in China.However,in the process of wind power generation,the complexity and high volatility of wind speed affect the efficiency and safety of wind power.Accurate wind speed prediction is conducive to the dispatch planning and effective supervision of wind farms,which can significantly reduce the operating cost of wind power generation,and ensure the stability of wind power grid connection and the safety of power system dispatching.In order to achieve high precision and strong stability of wind speed prediction,this paper proposes a novel wind speed prediction strategy VMD-SGMD-DE-BP.Firstly,in view of the high volatility of wind speed series,this paper constructs a new secondary decomposition model VMDSGMD.First,the VMD model is used to extract the low-frequency components in the wind speed sequence and reduce its complexity and volatility,and then the SGMD model is used to further decompose the remaining high-frequency wind speed sequences to obtain independent and clean decomposed components.The proposed secondary decomposition model has a better decomposition effect than other single decomposition models(EMD,CEEMDAN,VMD)and published secondary decomposition models(WT-VMD,EMD-WPD,CEEMDAN-VMD,SSAVMD).Secondly,the parameters of the BP neural network are optimized by the optimization method DE,and the obtained optimization model DE-BP has the prediction performance of high precision and high stability,and can well fit the nonlinear fluctuations in the wind speed series.In order to verify the superiority of the overall model in prediction performance,this paper selects the actual wind speed data set of four seasons in Chengde Wind Farm for empirical research and analysis.This comparative experiment comprehensively evaluates the predictive performance,predictive stability,and generalization ability of different models from different focuses by using12 correlated comparative models,3 error metrics(MAE,RMSE,MAPE),and their percent improvement metrics,and DM statistical tests,respectively.The results show that the proposed wind speed prediction strategy has the highest prediction accuracy and superior generalization ability,and the proposed secondary decomposition model is suitable for nonlinear wind speed sequence analysis.Therefore,the proposed model can be applied to wind speed prediction research in the future,providing a certain reference for enhancing the stability of wind power grid connection and promoting the sustainable development of the wind power industry. |
