摘要
针对电网运营评价指标多而繁杂,难以对指标类别与指标间的相关性定性定量分析等问题,本文提出了一种基于模糊聚类与互信息的电网运营指标筛选方法。首先,利用系统运行效率、可靠性、经济性基本原则对国内外现有对电网运营评价的相关指标进行初步筛选,然后建立中心指标利用模糊聚类分析法计算其他指标对中心指标的模糊隶属度,通过指标隶属度的判别完成对指标类别进行定性化分类,其次,选择指标特征并利用最大相关性筛选最优特征子集,利用互信息理论计算最优特征子集的最大群相关性熵,同时采用最小冗余系数做约束保持最优特征子互信息冗余在最小值对,完成相关次级指标类间最优关联度筛选。最后对某典型区域电网进行运营评价对比验证,结果表明:本文所提的指标筛选方法能够简单有效的评价智能电网的运营水平。
In view of the numerous and complicated grid operation evaluation indicators,it is difficult to qualitatively and quantitatively analyze the correlation between indicator categories and indicators.This paper proposes a grid operation indicator screening method based on fuzzy clustering and mutual information.First,Use the basic principles of system operation efficiency,reliability,and economy to conduct preliminary screening of relevant indicators for evaluating grid operation at home and abroad.Then establish the central index and use the fuzzy cluster analysis method to calculate the fuzzy membership degree of other indexes to the central index,and complete the qualitative classification of the index category through the discrimination of the index membership degree.Secondly,select index features and use the maximum correlation to filter the optimal feature subset,use mutual information theory to calculate the maximum group correlation entropy of the optimal feature subset,and use the minimum redundancy coefficient as a constraint to maintain the optimal feature sub-mutual information redundancy.Remaining in the minimum pair,complete the optimal relevance screening between the relevant secondary index classes.The results show that the proposed index screening method can simply and effectively evaluate the operation level of the smart grid.
作者
李嘉周
尹远
刘俊勇
王电钢
黄林
唐杰
LI Jiazhou;YIN Yuan;LIU Junyong;WANG Diangang;HUANG Lin;TANG Jie(State Grid Sichuan Electric Power Company Information&Telecommunication Company,Chengdu 610000 Sichuan,China;College of Electrical Engineering Sichuan University,Chengdu 610065 Sichuan,China)
出处
《电力大数据》
2020年第9期42-49,共8页
Power Systems and Big Data
关键词
运营指标
模糊聚类
互信息
指标筛选
最大相关性
operational indicators
fuzzy clustering
mutual information
Indicators of screening
maximum correlation
