Research On Short-term Wind Speed And Wind Power Prediction Of Wind Farms

With the improvement of social level,the energy crisis is becoming more and more serious.Wind energy as the most potential renewable energy has been widely concerned.However,due to the uncertainty and nonstationarity of wind power,the grid connected operation of wind power will seriously affect the dispatching and stability of power system.Improving the prediction accuracy of wind speed and wind power is an effective way to reduce the impact of wind power access to the grid and improve the stable operation of power system.Therefore,based on the historical data of wind farm,this paper studies the wind speed and wind power prediction,and the main contents are as follows:(1)In view of the traditional wind speed prediction method with human oriented,this paper deeply explores the dynamic characteristics of wind speed time series.According to the correlation dimension and the largest Lyapunov exponent of the chaotic time series,the chaotic characteristics of the wind speed time series are identified,and then the multi-scale recursive analysis of the wind speed time series is carried out by combining the complete ensemble empirical mode decomposition with adaptive noise(CEEMDAN)and recursive theory.In order to reduce the amount of calculation,the sample entropy(SE)is used to calculate the entropy of components,and the components are recombined according to the similarity and proximity of entropy.The phase diagram,recursion diagram and recursion quantitative analysis of the wind speed time series decomposed by CEEMDAN still has chaotic characteristics,and the certainty and stability of the decomposed components are improved compared with the original wind speed time series,which lays a theoretical foundation for the subsequent construction of wind speed prediction model of wind farm combined with chaos correlation theory.(2)In order to further reduce the wind speed prediction error caused by the dimension of the input data,a combined short-term wind speed prediction model is proposed,which combines the complete ensemble empirical mode decomposition with adaptive noise,phase space reconstruction and gated recurrent neural network.Firstly,the complete ensemble empirical mode decomposition with adaptive noise(CEEMDAN)method is used to decompose the wind speed data into relatively stable intrinsic mode function components;Secondly,the SE method is used to calculate the entropy value of each component and recombine the components according to the similarity and proximity of entropy value to reduce the amount of calculation;Then,the input dimension and delay time of each new component are calculated by using the PSR method to optimize the optimal input set of prediction;Finally,the wind speed value of each new component after reconstruction is predicted based on grated recurrent unit(GRU)method,and the prediction results of all new components is superimposed to get the final prediction result.The experimental results show that the combined forecasting model proposed in this paper can effectively improve the prediction accuracy.(3)When finding the relationship between wind speed and wind power according to the historical wind speed and wind power data of wind farm,power curve and neural network are generally used for indirect short-term wind power prediction,and the prediction effect of neural network is better.Therefore,a wind power prediction model based on Elman-Ada Boost neural network is proposed.Firstly,in order to truly reflect the performance of the wind turbine,the change point grouping quartile method is used to clean the abnormal data in the historical wind speed and wind power data of the wind farm;Then,the Elman-Adaboost neural network model is used to train the data of wind speed and wind power after clearning;Finally,the wind speed predicted by CEEMDAN-PSR-GRU model is imported into the trained Elman-Adaboost neural network model to realize wind power prediction.Compared with BP neural network and Elman neural network,the experimental results show that Elman-Ada Boost neural network model has the highest prediction accuracy of wind power.