论文部分内容阅读
针对含有频率相近谐波/间谐波成分的电力系统复杂信号，提出一种原子分解与加窗频移算法相结合的分析方法。首先对待分析信号进行加窗并应用快速傅里叶变换对加窗信号进行频谱分析，搜索幅值谱线的峰值，通过频域谱线的相位特性判断峰值谱线是否存在主瓣干扰。若存在主瓣干扰，使用原子分解算法在峰值谱线频率的一定邻域范围搜索最匹配的原子成分，提取受到主瓣干扰的谱峰谐波或间谐波参数。若不存在主瓣干扰，则使用加窗频移算法对峰值谱线进行频移，消除非同步采样造成的误差，获得谱峰谐波或间谐波参数。循环上述过程依次获得各谐波/间谐波的参数。算例表明本文方法可有效处理频率相近的谐波/间谐波信号。“,”In order to treat the complex electrical signals of harmonic/inter-harmonic components with closing frequencies, which prevailing in modern power systems, a new scheme is presented using the atomic decomposition and windowed frequency shifting algorithm. First of all, the analyzed signal is windowed and imposed with Fast Fourier Transform. Then, the Maximal Magnitude Component (MMC) needs to be determined by scanning the amplitude spectrum of the signal, and checking whether the main-lobe jamming exists in the MMC is followed according to the phase-frequency characteristics. If the main-lobe jamming exists, the atomic decomposition algorithm is used to search the best matching atom in neighborhood of the MMC frequency to obtain the exact parameters of predominant harmonics/inter-harmonics. If the main-lobe jamming does not exist, the windowed frequency shifting algorithm is employed to adjust the MMC frequency to remove the error produced by asynchronous sampling and to get the correct parameters of the predominant harmonics/inter-harmonics. The parameters are decomposed one by one from the complex signals by iterating the above process. The evaluation results show that the proposed approach can effectively dispose the complex signals of harmonics/inter-harmonics with closing frequencies.