Research On Forecasting Method Of Wind Speed And Wind Power Based On Combinatorial Model

With the increasingly prominent energy problems,the development and utilization of wind energy and other clean energy has been paid more and more attention all over the world.However,wind power itself has strong volatility and instability.Large-scale unstable wind power will bring a severe test to the safe and stable operation of the power system.The short-term wind power prediction is not only beneficial to wind farm to adjust the wind power generation plan,but also to the power dispatching department to arrange the scheduling plan reasonably,and to reduce the influence of wind power grid connection on the power system.Therefore,the research on wind power prediction has become the focus of research and has very important practical significance.Therefore,on the basis of the existing wind speed prediction models,this paper proposes a wind speed combination prediction model based on wavelet decomposition and fusion neural network,which reduces the influence of volatility and instability on wind speed prediction.Then considering the problems existing in parameter selection of support vector regression machine,whale algorithm is used to find the optimal parameters of support vector regression machine,and a wind power prediction model is established on the basis of this.The main research contents are as follows:(1)The processing of wind speed historical data by wavelet decomposition.In view of the instability of wind speed time series,randomness will directly affect the prediction accuracy.Wavelet decomposition is used to decompose the historical data of wind speed time series,and a series of wind speed components with different frequencies are obtained.It provides the basis for the research of wind speed prediction.(2)Different forecasting models are selected according to the frequency characteristics of the decomposed wind speed components.In order to reduce the influence of the highest frequency component on the wind speed prediction,a fusion neural network prediction model is established and the wind speed prediction is carried out.The remaining wind speed components are divided into low frequency components group and RBF neural network is selected to predict the wind speed.The final prediction of wind speed is obtained by equal weight superposition of each component of the forecast wind speed.The wind speed historical data of wind farm are taken as experimental samples,and the feasibility and effectiveness of the method are verified by comparative experiments.(3)According to the relationship between wind speed and actual wind power,a wind power forecasting model based on support vector regression is established.In order to reduce the influence of the model penalty parameter and kernel function parameter selection on the prediction results,the whale optimization algorithm is used to find the optimal parameters of the support vector regression machine,so as to improve the performance of the prediction model.The predicted wind speed is inputted into the training model,and the wind power forecast is obtained.The actual wind speed and wind power historical data are taken as experimental samples and compared with those before optimization.The results show that the prediction accuracy of support vector regression model after optimization of whale algorithm is higher,which proves the reliability and effectiveness of the proposed method.(4)In order to apply wind speed and wind power forecasting method to practice better,a wind speed-wind power forecasting system is designed and implemented on the basis of the above-mentioned wind speed and wind power forecasting methods.The system realizes user management,wind speed and wind power prediction.The system is simple and friendly in operation,convenient in management,and has certain practical application value.