基于自适应神经模糊滤波的低频振荡Prony分析  

Prony Analysis of Low Frequency Oscillations Based on Adaptive Neural-Fuzzy Filtering

作　　者：侯王宾[1] 刘天琪[1] 李兴源[1] 

机构地区：[1]四川大学电气信息学院,四川省成都市610065

出　　处：《电网技术》2010年第6期53-58,共6页Power System Technology

基　　金："十一五"国家科技攻关计划项目(JS2008113506594)~~

摘　　要：传统Prony法在分析低频振荡时对输入信号要求较高,存在着对噪声敏感的弱点。文中提出一种自适应神经模糊滤波和改进Prony法相结合的低频振荡分析方法。该方法先用自适应神经模糊滤波对低频振荡信号进行滤波,再用改进Prony法对滤波后的信号进行分析。其中改进Prony法有效阶数用归一化奇异值法确定。将该方法用于分析IEEE4机2区系统表明,在有色噪声影响下,该方法仍能相对准确地辨识出低频振荡主导模式,验证了其有效性。In view of the sensitivity of traditional Prony method to noise and its higher requirements to input signal while low-frequency oscillation is analyzed, a new low-frequency oscillation analysis method which integrates adaptive neural-fuzzy filtering with improved Prony method is proposed. Firstly, the fuzzy filtering of low-frequency signal is performed by adaptive neural-fuzzy inference system, then the filtered signal is analyzed by improved Prony method, and the effective orders of improved Prony method are determined by normalized singular values. Applying the proposed method to IEEE 4-machine 2-area system, simulation results show that the proposed method can accurately recognize the dominant mode of low-frequency oscillation under the affect of colored noise, and the effectiveness of the proposed method is oroved.

关 键 词：低频振荡:有色噪声 自适应神经模糊滤波 改进 PRONY法 归一化奇异值法 

分 类 号：TM76[电气工程—电力系统及自动化]