摘要
针对基于评估模型、神经网络以及数据挖掘的三种配电网状态监测方法监测精度较低问题,提出一种基于回归分析的配电网负荷老化状态自动监测方法。选择变量参数进行灰度关联计算,筛选变量,利用筛选好的变量构建理论回归模型,估计未知参数,根据参数进行包括拟合优度检验、显著性检验以及残差分析检验的理论回归模型检验,并根据结果调整模型,得到最终的多元回归分析模型,实现配电网负荷老化状态自动监测。结果表明:与3种传统配电网状态监测方法,该方法的监测误差更小,监测精度更高。
Aiming at the problem of low monitoring accuracy of three distribution network condition monitoring methods based on Evaluation model,neural network and data mining,an automatic monitoring method of distribution network load aging condition based on regression analysis is proposed.Variable parameters are selected for gray correlation calculation,variables are screened,theoretical regression model is constructed by selected variables,unknown parameters are estimated,and theoretical regression model tests including goodness of fit test,saliency test and residual analysis test are carried out according to the parameters,and the final multivariate is obtained by adjusting the model according to the results.Regression analysis model is used to realize automatic monitoring of distribution network load aging.The results show that,compared with the three traditional methods of distribution network condition monitoring,the method has smaller monitoring error and higher monitoring accuracy.
作者
张雷
刘树枫
王浩
李聪革
郝磊
ZHANG Lei;LIU Shu-feng;WANG Hao;LI Cong-ge;HAO Lei(East Inner Mongolia Electric Power Co.,Ltd.Chifeng Power Supply Company,Chifeng Power Supply Co.,Ltd.,Chifengshi 024000,China;Sifang&Huaneng Power System Control Co Ltd,Beijing 100020,China)
出处
《电子设计工程》
2020年第5期130-133,138,共5页
Electronic Design Engineering
关键词
回归分析
配电网
电网负荷
老化状态
监测
regression analysis
distribution network
grid load
aging state
monitoring
