典型电力用户用电负荷特性分类技术的研究  

Technology research on typical electric power load characteristic classification

作　　者：高曦莹[1] 张冶[1] 扬爽[1] 李广翱[1] 蔡颖凯[1] 

机构地区：[1]辽宁省电力有限公司电力科学研究院,沈阳110000

出　　处：《沈阳工程学院学报（自然科学版）》2013年第4期323-325,共3页Journal of Shenyang Institute of Engineering:Natural Science

摘　　要：由于分布式发电的特性及影响,在电力系统控制当中能否对负荷特性做出准确的预测讲变的十分关键.通过对电力系统负荷特性常用的聚类的方法如K-means、凝聚型层次聚类等深入的研究,针对这几种算法,利用国际IRIS、曲线聚类审定通用指标、实际的特性曲线聚类结果 3个算法进行了验证,从而总结出含有分布式电源的系统电力负荷曲线高兴率聚类算法.Due to the nature and the influence of distributed generation, it is tremendously critical whether accurate predictions about to load characteristic can be made. We probed into the clustering methods which are frequently used to research load characteristic, such as K-means, condensing type hierarchy clustering, etc. Those algorithms are veri- fied by IRIS, clustering authorized general index and the virtual characteristic curve of the clustering results. And thus a load curve happy rate clustering algorithm with distribution power supply system has been generalized.

关 键 词：负荷特性 聚类算法 曲线聚类 

分 类 号：TM715[电气工程—电力系统及自动化]