摘要
实现准确的中压配变负荷预测有助于降低配变损耗和提升无功补偿控制效率。考虑配变负荷数据易缺失和高噪声的特征,应用Hermite插值和高斯滤波技术提升负荷数据质量,提出了基于高斯过程回归的配变负荷预测方法,具有鲁棒性好和准确性高的特征。经实际典型配变负荷数据验证表明,所提方法能够提高配变负荷预测的准确性。最后,比较了不同历史数据规模和区间时预测准确性的差异。
The accurate load prediction of medium-voltage distribution transformer helps reduce distribution transformer loss and improve the control efficiency of reactive power compensation. Considering easy loss of distribution transformer data and the characteristics of high noise,this paper uses the Hermite interpolation and Gaussian filtering techniques to improve the quality of load data,proposinga load forecasting method of distribution transformer based on Gaussian process regression,which method has the characteristics of good robustness and high accuracy. The verification of actual typical distribution transformer datashows that the proposed method can improve the accuracy of load forecasting. Finally,the paper makes a comparison on the differences of prediction accuracy in different historical data size and intervals.
作者
沙倚天
SHA Yitian(Nanjing Power Supply Branch,State Grid Jiangsu Electric Power Co.,Ltd.,Nanjing Jiangsu 210019,China)
出处
《湖北电力》
2019年第6期56-60,77,共6页
Hubei Electric Power
关键词
中压配变
负荷预测
HERMITE插值
高斯滤波器
高斯过程回归
medium-voltage distribution transformer
load forecasting
Hermite interpolation
Gaussian filter
Gaussian process regression
