一种考虑广义负荷时变性的动态模型研究方法  

A Research Method for Dynamic Models Considering Generalized Load Time Variation Characters

作　　者：郑秋宏 韩蓓[1] 李国杰[1] 徐晨博 张利军 Zheng Qiuhong;Han Bei;Li Guojie;Xu Chenbo;Zhang Lijun(Key Laboratory of the Ministry of Education for Power Transmission and Power Conversion,Shanghai Jiao Tong University,Shanghai 200240,China;State Grid Zhejiang Electric Power Co.Economic Technology Research Institute,Hangzhou Zhejiang 310000,China)

机构地区：[1]上海交通大学电力传输与功率变换控制教育部重点实验室,上海200240 [2]国网浙江省电力公司经济技术研究院,杭州310000

出　　处：《电气自动化》2020年第2期33-36,共4页Electrical Automation

基　　金：国网浙江省电力有限公司科技项目(5211JY16000X)。

摘　　要：针对广义负荷随时间变化下难以进行动态建模的问题,提出一种结合聚类分析和总体测辨法的动态建模方法。对复杂的时变性场景采用k-means算法进行聚类,再进一步通过改进的粒子群算法辨识故障样本数据。最后,基于实际历史数据和DIgSILENT仿真平台的算例结果表明,能够在考虑负荷时变性的情况下有效地进行动态建模。In view of the difficulty with dynamic modeling due to generalized time-varying load characters,a dynamic modeling method combining clustering analysis and measurement-based approach was proposed.Complex time-varying scenarios were clustered in the K-means algorithm.Then,fault sample data was identified through improved particle swarm optimization.Finally,the results of calculation examples based on actual historical data the DIgSILENT simulation platform indicated that the proposed method could effectively perform dynamic modeling while load time variation characters were taken into account.

关 键 词：聚类 特征向量 广义综合负荷模型 负荷辨识 模型时变性 

分 类 号：TM714[电气工程—电力系统及自动化]