一种改进FFT多谱线插值谐波分析方法

针对谐波分析中加窗FFT计算存在运算量大的问题,对常用窗函数进行比较,利用莱夫-文森特(RifeVincent,RV)窗优越的频谱特性,提出一种基于4项RV(I)窗多谱线插值FFT改进算法。通过分析加窗信号傅里叶变换的频域表达式,利用窗函数主瓣内相邻谱线间的相位特性,以及谐波频点附近的最大值谱线、次大值谱线和较大值谱线确定频率谱线的准确位置,改进了修正谐波幅值、频率偏差的计算方法,满足谐波分析准确度要求的同时,大幅降低运算量,提高谐波分析的实时性。仿真结果表明,提出的谐波分析方法能有效克服频率波动的影响,提高谐波测量的准确度,且能有效抑制白噪声的影响。

An Improved Harmonic Analysis Algorithm of Multi-spectrum-line Interpolation FFT

Aiming at the shortcomings of large computing capacity about the commonly used interpolation FFT algorithm that combined with complex window function, by comparing different windows and analyzing a good side lobe characteristic of Rife-Vincent window, an improved algorithm of multi-spectrum-line interpolation based on 4-term Rife-Vincent (I) window is proposed. After analyzing the signal model in frequency domain, and using the phase characteristic of adjacent spectrum lines in main lobe of window, the amplitudes of three neighboring spectrum lines near the harmonic frequencies are used to locate the accurate position of the harmonic spectrum lines, eventually, the correction formulas of amplitude and frequency deviation can be obtained, which can greatly reduce the computing capacity and improve the real-time performance, as well as satisfy the requirement of accuracy of harmonic analysis. The simulation results show that the proposed method in this paper can effectively overcome the influence of frequency fluctuations, enhance the accuracy of harmonic measurement, and can effectively restrain the influence of white noise.