Wind Farm Power Prediction With Support Vector Machine Regression

Wind energy,as a renewable energy source of clean renewable energy,is an important part of the energy structure of the country and society.With the rapid rise of wind power generation industry,the realization of grid-connected large-scale wind farms and consequent power forecasting of wind turbines are also becoming more and more important.However,due to many factors that affect the power of wind turbines,and the influencing factors are random and intermittent,it is challenging and far-reaching to study the power of wind turbines.In this paper,by analyzing the existing wind turbine power forecast research,in order to improve the wind turbine power forecasting accuracy,a new improved firefly algorithm is proposed to realize the selection of the parameters of the support vector machine regression model;the power correlation degree with the wind turbine will be compared.Large factors are used as model inputs to improve prediction accuracy.The main research work is as follows:First of all,according to the analysis of existing wind turbine power forecast research,it is proved that SVM has advantages over traditional statistics theory.After analyzing the theory of minimizing empirical risk,minimizing structural risk,and VC dimension,the basic theory and principle of support vector machine regression are introduced.Power error is used as an evaluation criterion to compare with other algorithms,and it is proved that the support vector machine solves the problem of small samples.The machine learning problem has better generalization ability.By calculating the Pearson correlation coefficient between the SCADA monitoring project and the output active power,the four factors that have the greatest impact on the wind turbine power are obtained.Secondly,a power forecasting method based on the improved firefly algorithm using the improved firefly algorithm is proposed.The improved firefly algorithm is used to select the appropriate support vector machine regression model parameters and enhances the prediction accuracy.Based on the average absolute error and root mean square error,the results of the improved FA-SVR model are compared with the results of PSO-SVR and GS-SVR.It is proved that the improved FA-SVR model has better than GS-SVR model and PSO-SVR model.Higher accuracy and better results can be used to predict wind turbine power.Finally,the four factors that have a significant impact on the power of wind turbines constitute a four-dimensional set as the input to the support vector machine regression model.Through validity analysis and results analysis,the effectiveness of the improved algorithm is proved.