Purpose

The purpose of this paper is to improve the online monitoring level of low-frequency oscillation in the power system. A modal identification method of discrete Fourier transform (DFT) curve fitting based on ambient data is proposed in this study.

Design/methodology/approach

An autoregressive moving average mathematical model of ambient data was established, parameters of low-frequency oscillation were designed and parameters of low-frequency oscillation were estimated via DFT curve fitting. The variational modal decomposition method is used to filter direct current components in ambient data signals to improve the accuracy of identification. Simulation phasor measurement unit data and measured data of the power grid proved the correctness of this method.

Findings

Compared with the modified extended Yule-Walker method, the proposed approach demonstrates the advantages of fast calculation speed and high accuracy.

Originality/value

Modal identification method of low-frequency oscillation based on ambient data demonstrated high precision and short running time for small interference patterns. This study provides a new research idea for low-frequency oscillation analysis and early warning of power systems.

中文翻译：

---

基于环境数据的电力系统低频振荡模式识别方法

目的

本文的目的是提高电力系统低频振荡的在线监测水平。提出了一种基于环境数据的离散傅里叶变换（DFT）曲线拟合的模态识别方法。

设计/方法/方法

建立了环境数据的自回归移动平均数学模型，设计了低频振荡参数，并通过DFT曲线拟合估计了低频振荡参数。变分模态分解方法用于对环境数据信号中的直流分量进行滤波，以提高识别的准确性。仿真相量测量单位数据和电网测量数据证明了该方法的正确性。

发现

与改进的扩展Yule-Walker方法相比，该方法具有计算速度快，精度高的优点。

创意/价值

基于环境数据的低频振荡模态识别方法具有精度高，运行时间短，干扰小等优点。该研究为电力系统的低频振荡分析和预警提供了新的研究思路。