基于互高阶谱SVD算法的电力系统低频振荡频率估计  

Frequency Estimating of Low Oscillation in the Power System Based on Cross-high-spectrum SVD

作　　者：刘座铭[1] 王韵然[2] 高宇 王春瑛[4] 

机构地区：[1]东北电力大学,吉林吉林132012 [2]北京交通大学,北京100044 [3]黑龙江省电力勘察设计院,黑龙江150010 [4]吉林油田热电厂,吉林松原138000

出　　处：《东北电力大学学报》2007年第4期95-98,共4页Journal of Northeast Electric Power University

摘　　要：介绍了互高阶谱理论在互联电力系统低频振荡频率估计中的应用,详细讨论了其数学模型和处理过程。在相关高斯噪声与非相关噪声背景下,首先用互高阶谱 SVD 方法,将信号分解成信号矢量空间和噪声空间,通过求解噪声信号的特征方程则可知频率信号。该方法不需要任何关于有色噪声的先验信息就可以估计低频振荡频率。仿真分析的结果表明该方法具有良好的谱估计分辨率及谱估计稳定性,是低频振荡频率估计‘的有效工具。This paper presents high -order self-spectral theory for the analysis of low frequency oscillations in interconnected power systems and discusses mathematical model and procedure in detail. Under coherent Gaussian noise and incoherent noise, in this method the singular value decomposition （SVD） is applied, and the signal vector space and noise space are established; therefore, the amplitude, frequency, and phase of sinusoidal signal can be estimated without almost any prior information about colored noise. By dint of much analog simulation,proving the method has good resolution and stability of spectral estimation and is a powerful tool for oscillation research.

关 键 词：电力系统 低频振荡 频率估计 互高阶谱 奇异值分解 

分 类 号：TM744[电气工程—电力系统及自动化]