电力系统谐波频率检测的高阶累积量近似联合对角化算法

Simultaneous Joint Diagonalization of High-order Cumulant Algorithm for Harmonic Frequency Detection of Power System

提出了1种用于电力系统谐波频率恢复的高阶累积量联合对角化算法。该算法以电力系统谐波采样数据的四阶累积量矩阵为基础,对1组四阶累积量矩阵进行联合对角化,然后根据root-music算法的原理,利用对角化所得的对角矩阵和联合对角化器,检测电力系统中的谐波频率,为谐波的分析和治理提供可靠依据。试验结果表明本算法优点如下:一是仅用1组采样数据就构造了多个四阶累积量矩阵来共同辨识信号频率,提高了采样数据的利用率,提高了算法精度;二是低信噪比时,仍可以自动有效地抑制高斯白噪声和高斯色噪声。

A simultaneous joint diagonalization of high-order cumulant algorithm for power system harmonic frequency detection is introduced in this paper.In the algorithm,the simultaneous joint diagonalization is carried on to a group of fourth-order cumulant matrix.Then according to the principle of root-music algorithm,the unitary matrix and the diagonal matrix which are obtained can be used to retrieve the harmonic frequencies.The result will provide a reliable basis for the harmonic analysis and refrain.The experiment suggests that this approach has some virtues：with only a series of measurements,several fourth-order cumulant matrix to identify the signals are constructed with one group of sampling data.This can improve the utilization rate of the measurements and the precision of the algorithm;it can effectively and automatically refrain not only the gaussian white noise but also the color noise even when the SNR is low.