基于改进K-means聚类算法的风电场动态等值  

RESEARCH ON DYNAMIC EQUIVALENT OF WIND FARM BASED ON IMPROVED K-MEANS CLUSTERING ALGORITHM

作　　者：吴红斌[1] 何叶 赵波 李鹏 季宇[3] 吴鸣[3] Wu Hongbin;He Ye;Zhao Bo;Li Peng;Ji Yu;Wu Ming(School of Electrical Engineering and Automation,Hefei University of Technology,Hefei230009,China;Zhejiang Electric Power Corporation Research Institute,Hangzhou310014,China;China Electric Power Research Institute,Beijing100192,China)

机构地区：[1]合肥工业大学电气与自动化工程学院,合肥230009 [2]浙江省电力公司电力科学研究院,杭州310014 [3]中国电力科学研究院,北京100192

出　　处：《太阳能学报》2018年第11期3232-3238,共7页Acta Energiae Solaris Sinica

基　　金：国家重点研发计划(2016YFB0900400;2016YFB0900404);国家电网公司科技项目(5211DS150015)

摘　　要：提出一种基于改进K-means聚类算法的动态等值建模方法,将风速、输出功率及转差率作为聚类指标对风电机组进行动态分群,然后将群内的风电机组分别进行等值,可实现风电场的动态等值多机表征。利用算例系统进行验证,仿真结果表明该文所提出的等值方法能更准确地反映双馈机组风电场并网点的动态特性。In the study of grid connection of wind farm,computing scale and calculating speed will be greatly increased by simulating on each wind power generator,so it is necessary to study the equivalence modeling of wind farm.A dynamic equivalent modeling method based on improved K-means clustering algorithm is proposed,the method is based on three state parameters of the wind speed,output power and the slip as the clustering index to cluster wind power generator dynamiclly,and then the wind power generators in the same group are equivalent as a wind generator,in order to realize the dynamic equivalent multi-machine representation of the wind farm.By texting the example system to verify,the simulation results show that the proposed method is more accurate to describes the dynamic characteristics of the wind farm of the doubly fed induction generator at the grid connection point.

关 键 词：风电机组 聚类算法 动态模型 等值研究 

分 类 号：TM743[电气工程—电力系统及自动化]