基于改进K-均值聚类的电动机群聚合  

Probe into Aggregation of Induction Motors Based on Improved K-means Clustering Algorithm

作　　者：左宇 顾琨 李博江 李玲 杨智元 曹晖 张钢 王雨晗[3] 

机构地区：[1]国网西安供电公司,陕西西安710032 [2]国家电网西北分部,陕西西安710049 [3]华北电力大学电气与电子工程学院,北京102206

出　　处：《陕西电力》2016年第10期32-37,共6页Shanxi Electric Power

基　　金：中央高校基本科研业务费专项资金资助(2015ZD03)

摘　　要：提出了一种新的电动机聚合方法。首先对电动机动态模型进行轨迹灵敏度分析,然后以灵敏度均值作为改进K-均值聚类算法的特征权值进行电动机分群计算,最后结合加权平均法和空转-堵转法计算等值电动机的参数。算例对Anderson 3机9节点系统进行了仿真计算,选取其中6个参数差异较大的电动机群作为研究对象,对其进行了动态等值,分析比较了本文方法与传统方法得到的等值电动机在不同故障切除时间下的系统稳定性。结果表明本文方法得到的等值电动机能够更好地反映原始电动机群的动态特性。This paper proposes a new method for aggregating induction motors. Firstly, trajectory sensitivity analysis is performed on the dynamic model of induction motors. Then an improved K-means clustering algorithm with feature weights obtained by averaging the trajectory sensitivity curve of each parameter is used to classify homogeneous motors into different groups. The parameters of the aggregated induction motor are determined based on two operating conditions, i.e. no-load and locked-rotor conditions. Simulations and analysis is carried out in Anderson 9-bus system with 6 different motors and the comparison between the system＇s stability obtained from the proposed method and the traditional methods under different fault clearing time is done. The result shows that the new method can improve the approximation of the dynamic characteristic of the original motor groups.

关 键 词：电动机聚合 K-均值聚类 特征权值 轨迹灵敏度 

分 类 号：TM346[电气工程—电机]