Research On Wind Speed Prediction Based On Hybrid Optimization Model

With the increasing depletion of fossil fuel and other traditional resources and the increasingly serious environmental pollution,countries all over the world are looking for clean energy,and wind energy as a renewable energy has attracted great attention in the energy sector.Ocean has a lot of wind energy resources.In recent years,ocean wind power generation has become mature.Wind speed is the key content of wind power generation evaluation.However,wind speed is easily affected by obstacles,terrain and other factors and becomes unstable.With the change of height,wind speed also changes accordingly.Therefore,accurate prediction of wind speed is challenging.Most scholars focus on improving the accuracy or stability of prediction,but ignore the importance of improving at the same time.Therefore,this paper proposes a hybrid model which can effectively improve the accuracy and stability of wind speed prediction,so as to reduce or avoid the adverse effects of wind farms on the power system.This paper proposes a new hybrid prediction model for wind speed prediction and research,including effective double decomposition data processing algorithm,combined prediction algorithm,gray wolf optimization algorithm and effective evaluation method.In this study,lmd-vmd decomposes the original wind speed data into a series of modes function(IMF).The partial autocorrelation function(PACF)is used to select the BP neural network of input variables,the limit learning machine(ELM)and Elman neural network are used to predict and integrate the training set and test set of these subsequences respectively,and the predicted results of the three prediction algorithms are used as the input of support vector regression(SVR)optimized by gray wolf optimization algorithm to predict.In order to verify the effectiveness of the proposed model,the wind speed data collected from Dandong,Liaoning Province andSheyang,Jiangsu Province are used as samples for one-step prediction and two-step prediction,and the comparative model and comprehensive evaluation mechanism are introduced.the results demonstrate that the hybrid system proposed in this paper has higher accuracy and stability than other common models.Therefore,the model lmd-vmd-bp-elm-Elman-gwo-svr can be effectively used for wind speed prediction and research.