多指标FCM聚类算法在风电场聚类问题中的应用  

Application of Multi Index FCM Clustering Algorithm in Wind Farm Clustering Problem

作　　者：李观阳 刘洪正 李广磊 孙毅 刘志敏 LI Guanyang;LIU Hongzheng;LI Guanglei;SUN Yi;LIU Zhimin(Shandong University of Technology, Zibo 255000, China;State Grid Shandong Electric Power Research Institute,Jinan 250003, China;Shandong Zhongshi Yitong Group Co., Ltd. ,Jinan 250003, China)

机构地区：[1]山东理工大学,山东淄博255000 [2]国网山东省电力公司电力科学研究院,山东济南250003 [3]山东中实易通集团有限公司,山东济南250003

出　　处：《山东电力技术》2018年第2期1-5,17,共6页Shandong Electric Power

基　　金：国家科技支撑计划资助项目(2015BAA07B01)

摘　　要：本文首先介绍了FCM聚类算法相对于传统的K-means聚类算法、AP聚类的优势,并将FCM算法引入风电场多指标聚类中去。获取了东源风电场的详细模型数据及运行数据,借助Digsilent/Power factor软件搭建东源风电场模型并嵌入风速模型,根据风力发电机风速、风力机功率、故障结束0.4 s后的转子的转速作为三项聚类指标,对33台风机进行了聚类建模,并与详细模型进行了比较,验证了基于FCM聚类算法在处理风电场聚类问题上的优越性。The advantages of FCM（Fuzzy clustering-means） algorithm relative to the traditional K-means clustering algorithm and AP clustering are first explained,and the FCM algorithm is introduced into the multi index clustering of wind farm.The detailed model data and operation data of Dongyuan wind farm are obtained and the detailed model of Dongyuan wind farm is built with Digsilent/Power factor software and embedded into the wind speed model.Taking the wind speed of wind turbine,the power of wind turbine and the speed of rotor after 0.4 s as the three clustering indexes,cluster modeling of 33 fans is carried out,and then compared with the detailed model,which verified the superiority of FCM clustering algorithm in dealing with the clustering problem of wind farms.

关 键 词：FCM聚类算法 多指标聚类 Digsilent/Power factory软件 风电场模型 

分 类 号：TM614[电气工程—电力系统及自动化] TM71