摘要
风电功率波动分量的存在对风电功率的预测精度有重要影响。为提高风电功率的预测精度,本文从波动分量偏度入手,推导出预测精度与波动分量的偏度的相关性;并选取合格率作为精度指标,运用卡尔曼滤波模型和小波神经网络模型计算合格率,验证了合格率与偏度的相关性;最终得出波动分量的偏度与风电功率的预测精度呈显著负相关,该结果表明选取波动分量偏度较小的数据可提升风电功率预测精度。
The existence of the fluctuant component has restricted the prediction accuracy. For the purpose of improving the situation, we strated with the skewness of fluctuant to deduce the correlation between prediction accuracy and skewness. Skewness theory is put forward with the Kalman filtering and Wavelet neural network experiments. This paper selected the qualified rate as accuracy index and verified the correlation between the skewness and qualified rate. The result of the present work showed a significant negative correlation between the accuracy skewness of fluctuation component and wind power and also implied that the data with lower skewness can further enhance the prediction accuracy.
出处
《石河子大学学报（自然科学版）》
CAS
2016年第2期244-250,共7页
Journal of Shihezi University(Natural Science)
基金
石河子大学大学生创新训练计划项目(201510759037)
关键词
波动分量
偏度
预测精度
fluctuant component
skewness
prediction accuracy
