基于加新型自卷积窗的频谱插值谐波分析方法

摘要

为了解决快速傅里叶变换在电力谐波分析方法中存在采样的不同与非整数周期截取而造成的栅栏效应和频谱泄露的问题，提出一种基于 P 阶三角自卷积窗改进 FFT 频谱插值的电力系统谐波分析方法。首先使用 P 阶自卷积窗截取信号，之后选取幅值最大的频率附近的离散的频谱3条谱线进行加权平均计算来确定谐波谱线的准确位置，进而可以得到谐波的幅值、相位和频率，最后通过多项式拟合的方法，提出谐波修正的公式。通过仿真分析，所提出的自卷积窗函数能降低频谱泄露和栅栏效应带来的影响。采用改进的频谱插值算法可以提高电力谐波的检测精度，有助于对谐波的应用分析。%In power system harmonic analysis method,fast Fourier transform (FFT)has the problems of barrier effect and spectral leakage caused by the difference in sampling and the interception of non-integer cycle.To tackle them,we proposed in this paper a power system harmonious analysis method which is based on FFT spectrum interpolation improved by P order triangle self-convolution window.The first step is to use P order self-convolution window to intercept signals,after that three spectrum lines of discrete spectra close to the frequency with largest amplitude are selected for weighted average calculation to determine precise positions of the harmonic spectra,and then the amplitudes,phases and frequencies of the harmonics can be got.Finally,by using the method of polynomial fitting,we proposed harmonic correction formula.Simulation results show that the proposed self-convolution window function can reduce the influence brought about by spectrum leakage and barrier effect.To use the improved spectrum interpolation algorithm can raise the precision of electric power harmonic detection and contribute to the applied analysis of harmonic.