In order to improve the level of real-time monitor of low frequency oscillation, this paper proposes a decorrelation Least Mean Square (LMS) algorithm based on transversal filter model for online identification of low frequency oscillation modes. Based on the initial Least Mean Square (LMS) adaptive filtering algorithm, the improved algorithm relieves the relativity between input signals, and promotes algorithm's accuracy and convergence speed. Furthermore, this paper gives the results of the New-England 39-bus system's simulation data analysis and identification calculation of several mapping line of south China power grid, which proves the validity of this improved algorithm. And comparing with the identification results of basic LMS algorithm and traditional ARMA (auto-regressive moving-average) algorithm, it is argued that the new improved algorithm does not only have better accuracy and convergence speed, but also has practical project meaning. rnThis work is supported by National Program on Key Basic Research Project (973 Program) (No. 2009CB724505-1).%为提高电力系统低频振荡现象的实时监测水平,提出一种基于横向滤波器模型的解相关最小均方误差递推算法进行低频振荡模式辨识.该改进算法在原有最小均方自适应滤波算法的基础上,解除输入信号之间的相关性,提高了算法辨识的精度和收敛速度.通过对New-England 10机39节点系统的仿真数据分析以及南方某电网实测线路的辫识计算,其结果验证了该改进算法对低频振荡模式辨识的有效性.并通过与基本的LMS(最小均方)算法以及传统ARMA(自回归一滑动平均)算法辨识效果的比较,验证了该改进算法对低频振荡模式的辨识具有更好的精确性且提高了收敛速度,更具有实际的工程意义.
展开▼
