摘要
针对传统Prony方法对噪声敏感和辨识精度不高的局限性,提出一种新的低频振荡模式辨识方法,实现了在有噪声干扰情况下低频振荡模式的准确辨识。该方法基于数学形态学设计出一种多结构元素的并行复合形态滤波器,可有效滤除多种噪声,保留更多的有用信息。对消噪后的信号采用基于总体最小二乘法-旋转不变技术的信号参数估计(TLS-ESPRIT)算法进行辨识,从而获取低频振荡各个模式参数。通过算例仿真,说明所提出的方法是可行和有效的。
Aiming at how to avoid the limitations of Prony method that is sensitive to noise and not so accurate, a new method for identifying low frequency oscillation modes is proposed. It realizes accurate identification of oscillation modes with mass of noise. The method is based on mathematical morphology to design a class of generalized multi-structuring-elements parallel complex morphology filter. The morphological filter is applied to effectively suppress the noises and retain the main characteristic components of signal. TLS-ESPRIT algorithm is then used to detect the de-noised signal and obtain the low frequency oscillation modes. The simulative results have shown that the method is feasible and effective.
出处
《电力系统保护与控制》
EI
CSCD
北大核心
2012年第3期114-118,123,共6页
Power System Protection and Control
基金
国家自然科学基金项目(51077103)
武汉大学自主科研项目(5082009)~~
关键词
低频振荡
数学形态学
TLS-ESPRIT算法
振荡模式辨识
low frequency oscillation
mathematical morphology
TLS-ESPRIT algorithm
oscillation mode identification
