Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
1270296 | International Journal of Hydrogen Energy | 2016 | 7 Pages |
Focusing on the ultra-short term power prediction of wind farm, a novel method considering operational condition of wind turbines was proposed in this research. Pearson correlation coefficient between the output power and the operational condition was analyzed firstly through the SCADA data of wind turbine, which can illustrate the specific relevance of the SCADA monitoring items on the output wind power of the wind turbine. Then a Support Vector Regression (SVR) model was established to predict the wind power of single wind turbine with the meteorological and SCADA monitoring information. The predicted results of the model considering the operation condition was better than which of the model only considering meteorological information. Finally, considering that the wind turbines at different space positions have different contributions to output power of wind farm, the regression model was established with the input of prediction power of each wind turbine and the output of prediction power of wind farm. The method in this paper was denoted efficient by the result that the prediction error of wind farm regression model was less than that of the model summed all the prediction power of wind turbines.