| 引用本文: | 徐志钮,律方成,李燕青,李和明.基于ICA的含噪电力系统信号的频率测量[J].电力自动化设备,2008,(8): |
| .ICA-based frequency measurement of power system signal with noise[J].Electric Power Automation Equipment,2008,(8): |
|
|
| 摘要: |
| 为提高在噪声环境下频率测量的准确性,提出一种结合独立分量分析(ICA)和加汉宁窗插值算法的含噪电力系统信号的频率测量方法.该方法使用ICA对混有噪声的信号进行分离获得电力系统信号后,使用加汉宁窗插值的傅里叶算法获得电力系统频率.由于ICA对噪声和有用信号进行了很好的分离,因此频率测量的精确度得到了显著提高.仿真分析了含有白噪声和脉冲噪声的情况.即使噪声的幅值大于信号幅值数百倍时,使用所提出算法后频率误差绝对值的最大值从2.4 Hz左右分别减少到了0.0028 Hz和0.0005 Hz,表明所提出算法在极低信噪比时仍具有较高的精确度. |
| 关键词: 电力系统 频率测量 白噪声 脉冲噪声 |
| DOI： |
| 分类号:TM935.1 |
| 基金项目: |
|
| ICA-based frequency measurement of power system signal with noise |
|
XU Zhiniu LU Fangcheng LI Yanqing LI Heming
|
| Abstract: |
| A frequency measuring algorithm for power system signals with noise is proposed to improve measurement accuracy,which combines ICA(Independent Component Analysis) and Hanning -windowing interpolation algorithm. ICA is used to separate noise from power system signals and Hanning-windowing interpolation algorithm is used to calculate signal frequency,therefore,the accuracy of frequency measurement is improved distinctly. Signals with white noise and impulse noise are analyzed by simulation. Even if the amplitude of white or impulse noise is hundreds times larger than that of signal,the maximal absolute value of frequency error calculated by the proposed algorithm is reduced from about 2.4 Hz to 0.002 8 Hz or 0.000 5 Hz respectively,which indicates that the proposed algorithm has higher accuracy even when SNR(Signal - to - Noise - Ratio) is low. |
| Key words: ICA |
