基于优化原子分解的含频谱干涉谐波/间谐波检测算法

Harmonics/inter-harmonics with spectrum interference rapid detection based on optimization of atomic decomposition

针对现代电网中含主瓣干扰和负频率干扰的谐波/间谐波信号,提出了一种基于优化原子分解的快速检测方法。对待分析信号加Hanning窗并进行FFT谱估计,根据频域谱线的相位特性和谱线幅值的比值校正理论,构建一种高可信度综合频谱干涉判据,判断频域内发生主瓣干扰或负频率干扰的谱线。对受干扰谱线在其邻域的一定频率间隔范围内进行粒子群优化(PSO)算法粗搜索,将得到的寻优参数重新离散化,在更小的范围内采用原子分解算法,快速准确获得谐波/间谐波参数。本文方法大大减小了匹配追踪算法的寻优范围和计算量,并能保证搜索的全局最优。仿真结果表明本文算法可快速有效分析含频谱干涉的复杂谐波/间谐波信号。

In this paper, an optimized atomic decomposition algorithm is proposed to rapidly detect and analyze the harmonic / inter-harmonic signal that contains main-lobe interference and negative frequency interference in the modern power grid. Firstly, the FFT spectrum estimation is utilized to the analyzed signal with Hanning window functions, and based on the phase characteristic and the amplitude ratio correction theory of the frequency domain spectral, a dynamic spectrum interference criterion is proposed and proved to be highly credible. When the main lobe interference and negative frequency interference are detected from the harmonic/ inter-harmonic spectrum lines, the PSO algorithm is used to search the approximate range of parameters in adjacent of spectrum lines, then the rough parameters are discretized and the atomic decomposition algorithm is applied to search the best matching atom in a smaller range to obtain the precise parameters of harmonic/ inter-harmonic signals. According to analysis, the searching iterations and computation are reduced, and the searching result is also guaranteed to be the global optimal solution. The simulation results demonstrate that the proposed method can rapidly and effectively detect the harmonic/ inter-harmonic signal that contains spectrum interference.