基于统计聚类分析的短期风电功率预测

考虑到短期风电功率预测模型建立时，样本的选取对预测模型的精度有较大影响，提出了运用聚类方法对历史风速数据进行处理，实现了历史数据的自动分类。根据预测日的平均风速和最大风速等特征参数，按照相似度最大的原则，选择合适的类别作为预测建模用的训练样本。运用时间序列方法，建立风速预测模型，与不经过预处理的相比，所建立预测精度得到了提高，验证了运用聚类进行数据预处理的正确性。运用风力发电机的出力曲线，得到了未来日的风电功率的预测值，为含风电系统的电力系统运行计划的制定，提供了基础数据支持。

Short-term wind power prediction based on statistical clustering analysis

Considering that in the process of establishing short-term wind power prediction model,the sample selection would affect the predication accuracy of wind power model,it is necessary to process the history wind speed data prior to modeling.Data classification is automatically accomplished through the statistical clustering approach.With the criterion of maximal similarity,we select a group of data as our trained samples according to the average and maximum wind speed of prediction day.Then we establish the p...