纳托尔窗改进FFT动态谐波参数估计方法

Improved FFT Approach for the Dynamic State Estimation of Harmonic Parameters Based on the Nuttall Window

动态下的谐波参数估计是近年来的研究热门,快速傅里叶变换(fast Fourier transform,FFT)因其简单且易于嵌入式实现而得到广泛应用,但其参数估计准确度受频谱泄漏和栅栏效应的制约。分析纳托尔(Nuttall)窗的旁瓣特性,建立基于4项5阶Nuttall窗改进FFT的谐波参数估计算法,通过曲线拟合推导信号基波与各次谐波的频率、幅值和相位估计修正公式。仿真结果表明:4项5阶Nuttall窗抑制频谱泄漏效果好,改进FFT算法能对栅栏效应产生的影响进行有效修正,提出的方法能准确跟踪基波频率波动,有效抑制白噪声影响,提高谐波参数估计的准确度。

Research on the dynamic state estimation of harmonic parameters has become increasingly popular in recent years and the fast Fourier transformation （FFT） has been widely used for its simplicity and easy realization in the embedded system. However, the computation precision is still low, constrained by the spectral leakage and picket fence effect In this paper, the sidelobe characteristics of the Nuttall window are analyzed, and an improved FFT approach for harmonic parameters estimation based on four-term fifth derivative （FTFD） Nuttall window is proposed. The rectification formulae of frequency, amplitude and phase of the fundamental and harmonics are deduced by using the polynomial curve fit functions. Simulation results indicate that, spectral leakage can be effectively curbed by the FTFD Nuttall window, and the picket fence effect can be effectively modified by the improved FFT algorithm. The proposed method can precisely trace the fundamental frequency fluctuation, effectively curb the effect of white noise, and greatly improve precision in estimating harmonic parameters.