低频振荡模式在线辨识的最小平均M估计自适应滤波算法  

A Least Mean M-Estimation Adaptive Filtering Algorithm for On-Line Identification of Low-Frequency Oscillation Modes

作　　者：刘贵富[1] 卢继平[1] 徐玉韬[1] 

机构地区：[1]输配电装备及系统安全与新技术国家重点实验室(重庆大学),重庆市沙坪坝区400044

出　　处：《电网技术》2012年第10期178-183,共6页Power System Technology

基　　金：国家重点基础研究发展计划项目(973项目)(2009CB724505-1);国家"111"计划资助项目(B08036)~~

摘　　要：研究了常规最小均方二乘自适应滤波算法应用于低频振荡的基本原理,以及最小平均M估计(least meanM-estimate,LMM)算法的基本原理。首次将LMM算法应用于低频振荡模式的在线辨识,采用新英格兰-39节点系统时域仿真数据和某电网的同步相量装置(phasormeasurement unit,PMU)实测数据对所提方法进行了测试,验证了LMM算法在低频振荡模式在线辨识的鲁棒性和有效性,对稳态类噪声数据和动态数据均可辨识,并能消除异常数据对辨识结果的影响。Conventional least mean squares（LMS） adaptive filter algorithm is only applied to estimate ambient data but not to estimate the ring down data.The least mean M-estimation（LMM） algorithm is evolved from traditional LMS algorithm and its objective function is M-estimate function.In this paper an LMM adaptive filtering algorithm is proposed and it is the first time to apply LMM algorithm to online identification of low-frequency oscillation modes.The proposed method is tested by New England 39-bus system and phasor measurement unit measured data of a certain power grid to verify the robustness and effectiveness of applying LMM algorithm in online identification of low-frequency oscillation modes.Test results show that both steady-state noise data and dynamic data can be identified and the affects on abnormal data on identification results can be eliminated.

关 键 词：最小均方二乘算法 最小平均M估计算法 低频振荡 在线辨识 动态数据 

分 类 号：TM714[电气工程—电力系统及自动化]