Research On Combined Forecasting Of Wind Power Based On BP Neural Network And Support Vector Machine

In recent years,due to the increasing energy consumption,environmental protection problems are becoming more and more serious The development of clean energy has become the strategic choice of all countries in the world.As a clean and pollution-free renewable energy,wind power has been paid attention to by all countries.In China,wind power has become the third largest energy,and the installed capacity of wind power continues to increase,ranking first in the world.However,wind power generation has the characteristics of randomness,volatility and uncertainty,which will affect the safe and stable operation of the whole grid after wind power integration.The accuracy of wind power prediction is related to the scheduling plan of wind power system.In order to improve the utilization rate of wind power and ensure the smooth grid connection and stable operation of wind power,it is necessary to improve the accuracy of wind power prediction.According to different standards,wind power prediction methods can be divided into numerical weather forecast data prediction and historical data prediction.In this paper.wind power prediction based on historical data is the research direction.Because the wind turbine stops working at low wind speed and the acquisition equipment fails or the scheduling is unfavorable,the recorded data is not accurate enough.The original data is preprocessed to obtain high-quality data sets.Secondly,the basic BP neural network and support vector machine(SVM)prediction model are established.The classical BP neural network has strong nonlinear fitting ability.nonlinear mapping ability and strong self-learning adaptive ability;The support vector machine model has strong approximation ability and generalization ability,and the nonlinear operation can be completed quickly by introducing kernel function.BP neural network is easy to fall into local minimum,poor stability and empirical problems of SVM model parameter design.Because differential evolution algorithm has faster convergence speed,more accurate results and better robustness than other optimization algorithms in solving the optimal solution problem,Therefore,the DE-BP prediction model of BP neural network optimized by differential evolution algorithm and the DE-SVM prediction model of SVM optimized by differential evolution algorithm are established.The simulation results show that DE-BP model and DE-SVM model are better than BP neural network model and SVM model.Finally,two fixed weight combination strategies are used to combine DE-BP model and DE-SVM model for wind power prediction,so as to improve the accuracy of wind power prediction-The combined forecasting model is simulated and verified with the processed high-quality data of a wind farm in Inner Mongolia.The first mock exam is compared with the four single models BP,SVM,DE-BP and DE-SVM,and the effectiveness of the combination method is verified.Mean absolute error(MAE),root mean square error(RMSE)and mean absolute percentage error(MAPE)are used to evaluate the above models.The results show that the combined prediction method is more accurate and effective than the single BP neural network or support vector machine prediction model This study provides a reference for the development of wind power forecasting in China.