Abstract
Based on the characteristics of the power curves of wind turbine generator systems and the principles of the support vector machine (SVM), a piecewise support vector machine (PSVM) model is proposed in this article to improve the precision of short-term wind-power prediction systems. The operation data from a wind farm in north China are used to verify the proposed model, and the average mean error and root mean squared error of the PSVM model are 4.76% and 68.83 kW less than that of an SVM model respectively. Results of parameter optimization confirm the robustness of the PSVM model.
ACKNOWLEDGMENT
The authors gratefully acknowledge the support from Ministry of Science and Technology of China for the “863” Program of China: Wind Power Prediction Method Study and System Development (2007AA05Z428).The authors are also grateful for the valuable advice from the reviewers of both the International Journal of Green Energy and IGEC-IV.