一种修正的FFT高精度谐波分析方法

A Novel Way of Improving Harmonic Analysis Precision Based on Amending FFT Algorithm

信号的谐波分析一直是信号处理的基本方法之一。谐波分析要求对周期信号整周期采样,否则将导致频谱泄漏。本文通过对国内外一些提高谐波分析精度的有效方法的研究,在深入分析了采样周期对谐波分析精度的影响下,从傅立叶算法的基本原理出发,提出了一种新型的谐波分析方法。该方法利用非整周期采样数据及泰勒展开原理,估计理想的整周期采样序列,从而在FFT计算结果产生一个修正项,可以在不增加硬件设备及采样数据的前提下提高谐波分析精度,并通过计算机仿真证实了FFT修正方法的有效性。

Harmonic analysis is a basic way of signal analysis. It demands to synchronous-sample for a period signal, otherwise it must bring spectra leakage. This paper proposesd a new harmonic analysis way through investigating some domestic and foreign effective methods of improving harmonic analysis and the effect of sampling period for spectra analysis. The method can improve harmonic analysis precision under the condition of not adding another hardware and sampling data by amending FFT consequence to use asynchronous-sampling data and Taylor expansion theory to estimate corre- sponding ideal synchronous-sampling sequence. The validity of the method is tested by computer simulation.