用窗口傅里叶脊提取时变振荡信息

Time-varying oscillation information extraction by window Fourier ridge

Prony算法可以提取平稳信号的振荡频率和阻尼,但抗噪性能差,并且不能正确反映非平稳信号的振荡特性。提出用窗口傅里叶脊(WFR)分析非平稳振荡信号,其具有良好的抗噪性,WFR利用窗口傅里叶变换反映振荡信号的频域特性,以极大值点(脊点)反映信号振幅,通过窗口滑动反映振荡信号的时域特性,从而反映非平稳振荡信号的频率和阻尼时变性。算例说明,WFR不仅可以分析平稳信号的振荡信息,对于非平稳信号同样具有良好性能,更适合电力系统摇摆曲线的振荡信息分析。

Prony arithmetic extracts the oscillation frequency and damping of smooth signal but not those of non-smooth signal,and its anti-noise performance is quite poor.It is proposed to apply WFR（Window Fourier Ridge） to non-smooth oscillation signal.It has excellent anti-noise performance.WFR uses the window Fourier transform to display oscillation frequency character and the maximum point（ridge point） to show signal amplitude.The time-varying characters of frequency and damping are displayed by sliding window.The results of case study show that WFR has excellent performance in oscillation information analysis for both smooth and non-smooth signals,suitable for the that of power system swing curves.