摘要
采用FFT算法进行电力系统谐波分析很难做到同步采样和整周期截断,由此造成的频谱泄漏将影响谐波分析的效果。通过对频谱泄漏机理的详细分析,导出了信号实际频谱和泄漏频谱之间的关系,在此基础上提出了一种利用相位差校正信号频率来恢复实际频谱的改进算法。该算法只需要较短的采样数据长度,就能达到较高的计算精度,具有延时小、响应速度快等特点,避免了常用的加窗插值算法通过延长数据采样长度来提高计算精度的缺点,在实时性方面有较大的优势。
There are difficulties in performing synchronized sampling and integral period truncation in the harmonic analysis of power system with fast Fourier transform. However, if not satisfied, the results will be influenced by spectral leakage. Based on the comprehensive analysis of spectral leakage, a formula is deduced between the leakage spectrum and actual spectrum. In this paper, an improved algorithm, which uses phase difference to correct signal frequency and then recoveries actual spectrum, is presented. The algorithm can calculate more accurately with short sample duration, and the response speed is faster. Compared with the widely used window and interpolation algorithm which has to prolong the sample duration to improve the calculation accuracy, the algorithm proposed is better on real-time.
出处
《电力系统及其自动化学报》
CSCD
北大核心
2007年第2期98-102,共5页
Proceedings of the CSU-EPSA
关键词
谐波
快速傅里叶变换
非同步采样
频谱泄漏
实时性
harmonic
fast Fourier transform（FFT）
asynchronous sampling
spectral leakage
real time
