基于分群辨识曲线的信号能量分析法

Signal Energy Method Based on Cluster Identification Curve

针对传统的信号能量分析法无法直接处理高阶模式下的非平稳振荡信号,结合振荡机组的分群辨识,提出用基于分群辨识曲线的信号能量分析法进行低频振荡参数的在线辨识,以拓展信号能量分析法的适用范围,并提高辨识效率。该方法借助信号能量分析法能处理上下包络线关于时间轴对称的信号这一特点,利用振荡机组分群辨识得到的两个发电机群的反向对称轨迹,采用信号能量分析法辨识主导振荡模式。仿真测试的结果表明基于分群曲线的低频振荡在线辨识方法结果准确、计算效率高,能够有效地识别系统的主导振荡模式。

The traditional signal energy method is unable to deal directly with the non-stationary vibration signals of higher-order mode, so, based on clustering identification curve, this paper proposed an online identification method for low frequency oscillation by using signal energy method in order to expand the scope of signal energy method and improve the efficiency of identification. Considering that signal energy method can deal with signal, whose upper and lower envelope are symmetrical about time axis, the proposed method takes advantage of symmetric trajectory produced by two power generation groups of oscillation units, and then identifies the dominant oscillation mode by using signal energy method. The simulation test results prove that the identification is accurate and of high computational efficiency.