基于模糊聚类的电力负荷特性的分类与综合

THE CHARACTERISTICS CLASSIFICATION AND SYNTHESIS OF POWER LOAD BASED ON FUZZY CLUSTERING

在阐述负荷特性分类与综合内涵及意义的基础上,以变电站综合负荷构成成分比例为负荷特性分类和综合的基本特征,基于模糊聚类原理,提出了模糊等价关系和模糊C均值算法的2种分类方法。基于模糊C均值法可以通过优化理论获得聚类中心矩阵,同时完成负荷特性分类与综合。对某省48个变电站采用加权平均的方法进行聚类分析,得出了基于模糊等价关系的聚类综合特性并与模糊C均值算法的聚类中心矩阵进行了比较分析。结果表明,两者都具有良好的聚类综合能力;基于模糊C均值法的聚类能力明显优于基于等价关系的聚类法,而且聚类结果更为合理有效。两种方法都成功地解决了负荷建模中变电站特性分类处理的复杂性与主观性。

The article points out that the composition proportion of aggregate load is the essential load characteristics The authors clarified the signification of classification and synthesis for power load and proposed two classification approaches based-on fuzzy equivalent relation clustering and the fuzzy C means clustering. The latter can obtain the optimal clustering center and aggregate characteristics of load synchronously. It arranged 48 power substations under categories using this two methods and gained the class synthesis characteristic, Analyzing the clustering central matrix of this two methods draws the conclusion that this two methods have excellent clustering ability and the fuzzy C means clustering is surpassed the equivalent relation clustering, The results of fuzzy C means clustering are more reasonable and effective than the fuzzy equivalent relation clustering. This two clustering methods have overcame the difficulties resulting from complexity and subjectivity of load characteristic classification in load modeling. They have clear conception and convenient calculation and their exactness and validity are proved by the engineering instances.