基于频率约束的经验模态分解的谐波检测方法

Harmonic Detection Method Based on Frequency Constraining Empirical Mode Decomposition

在使用经验模式分解进行谐波检测时,经常会产生模态混叠。针对这一问题,提出了一种基于频率约束的EMD算法,该方法通过对固有模态函数中零点的分布加以频率约束,重新设计其筛分准则,减小了产生模态混叠的概率。分别采用基于频率约束的EMD、传统EMD以及瞬时无功功率法(p-q法)进行了谐波检测仿真实验。结果表明,该方法能够比较好地抑制传统EMD中存在的模态混叠问题,有效检测出基波电流与特定次谐波,且在检测精度以及基波电流相位的一致性方面都优于瞬时无功功率法。该算法可以比较好地分解出电网谐波,具有良好的应用前景。

Mode mixing often ocurrs while using empirical mode decomposition （EMD） in harmonic detection. To solve the problem, a new EMD algrothim was proposed based on frequency constrain. By frequency constrainting the distribution of zero points in intrinsic mode function （IMF） and redesigning the screening criteria, the occurrence probability of mode mixing was reduced. Simulation researches of harmonic detection had been done seperatly, which were based on frequency constraint EMD method, the traditional EMD method and instantaneous reactive power method （p-q method）. The result shows that the improved algrothim could overcome the mode mixing problem in traditional EMD, detect the fundamental current and the spetific harmonic effectively and get better detection accuracy and the consistency of the fundamental current phase than p-q method. The algrothim work better in harmonic decomposition to be of good prospects in application.