中文摘要：

风电功率短期预测对电力系统的调度运行有着重要意义.为提高风电功率短期预测的精度,构建基于改进的经验模态分解法（EMD）和支持向量机（SVM）相结合的预测模型,进行风电功率的短期预测.首先,采用镜像延拓算法对预处理后的功率序列进行处理,从而抑制经验模态分解法分解过程中的端点效应;同时,采用分段三次埃尔米特插值代替三次样条插值,由此得到的包络线可以有效改进EMD的欠冲或过冲问题;然后用改进后的EMD方法将风电功率序列分解成不同的分量,再针对各分量分别构建各自的SVM模型进行预测,最后将各预测分量进行叠加,由此得到总的风电功率预测值.实验结果表明,相比与其他的短期功率预测模型,改进的EMD-SVM预测模型具有更高的预测精度,具有一定的应用价值.

英文摘要：

Wind power short - term forecast is important to the dispatch and operation of power system. In order to improve predictive accu- racy of short -term wind power, a predictive model based on improved empirical mode decomposition （EMD） and support vector machine （SVM） is constructed for wind power short -term forecast. Firstly, the mirror extending method is employed to deal with the preproeessed power series, thereby reducing the boundary effect of EMD. Then, instead of traditional cubic spline interpolation, the piecewise cubic Her- mite interpolation is used. Envelope obtained can effectively improve the overshoots/undershoots. Next, wind power series can be decomposed into different series by the improved EMD. Last, SVM based on each component is used to forecast power by each component. The total wind power prediction result is obtained through reconstructing eventually. Experiment results show that the new predictive model proposed by this paper has higher prediction accuracy by comparing with other models and has certain application value.