基于递推随机子空间的电力系统低频振荡辨识  

Low-frequency oscillation identification based on recursive stochastic subspace for power system

作　　者：马燕峰[1] 刘伟东[2] 赵书强[1] 范振亚 

机构地区：[1]华北电力大学电力工程系,河北保定071003 [2]国网天津市电力公司,天津300072 [3]国网内蒙古东部电力有限公司赤峰供电公司,内蒙古赤峰024000

出　　处：《电力自动化设备》2016年第12期43-49,共7页Electric Power Automation Equipment

基　　金：中央高校基本科研业务费专项资金资助项目(2016MS86);国家电网公司大电网重大专项资助项目(SGCC-MPLG019-2012)~~

摘　　要：为实时提取低频振荡模式信息,采用基于随机子空间的低频振荡递推辨识方法。引入基于双边迭代的子空间递推方法实现随机子空间递推辨识,以提高辨识快速性和灵活性。利用递推误差并结合低频振荡数据的特点,提出一种能够保证快速平稳递推的遗忘因子和加权因子选择策略。对理想数据、仿真数据和WAMS数据分别采用所提方法进行分析,验证了该方法的可行性。A method of RSSI（ Recursive Stochastic Subspace Identification） is proposed to extract the information of low-frequency oscillation mode in real time. The sub-space recursive method based on biiteration is introduced to realize the RSSI with enhanced identification speed and flexibility. According to the characteristics of low-frequency oscillation data,a strategy based on the recursive error is proposed for selecting the forgetting factor and weighting factor,which ensures the fast and stable recursion. The applications of the proposed method to the ideal data,simulative data and WAMS data are analyzed,verifying its feasibility.

关 键 词：电力系统 随机子空间 低频振荡辨识 子空间追踪 

分 类 号：TM712[电气工程—电力系统及自动化]