基于K-means聚类算法的风电场机群划分方法  

Partitioning Method of Wind Turbine Grouping Based on K-means Clustering Algorithm

作　　者：徐立亮 胡仁祥 张毅 常喜强 闫亚东 

机构地区：[1]国网吐鲁番供电公司,新疆吐鲁番838000 [2]国网新疆电力调度通信中心,新疆乌鲁木齐830002

出　　处：《四川电力技术》2015年第6期72-75,84,共5页Sichuan Electric Power Technology

摘　　要：针对风电场动态等值建模的难题,采用K-means聚类算法,探讨了风电场机群划分问题,致力于达到风电场并网运行点特性的一致性。为减小等值算法带来的误差,在风力机功率转换特性不变的前提下,采取只对同一机型的风电场进行等值划分的方式来完成。并以某实际风电场为例进行算例仿真,结果表明,等值前后风电场并网运行点特性保持一致,等值算法能够准确反映风电场机组的动态响应特性。For the problem of dynamic equivalent modeling for wind farms, K - means clustering algorithm is adopted, and the partitioning issues of wind turbine grouping are discussed, which is devoted to achieve the consistency for the characteristics of parallel operation point of wind farms. For reducing the error brought by the equivalent algorithm, and on the premise that the power conversion characteristics of wind turbine are unchanged, it completes equivalence partitioning only on the same type of wind farms. And the example simulation is carried out taking an actual wind farm for example. The results show that the characteristics of parallel operation point consist with each other before and after the equivalence of wind farms, so the equivalent algorithm can accurately reflect the dynamic response characteristics of wind turbines in wind farms.

关 键 词：聚类算法 风电场 等值参数 等值模型 

分 类 号：TM614[电气工程—电力系统及自动化]