基于经验模式分解的有源滤波器谐波检测

Harmonic Detection of Active Power Filter Based on Empirical Mode Decomposition

本文把经验模式分解(EMD)应用于有源滤波器的谐波检测。对谐波电流进行EMD分解,得到一系列固有模态函数(IMF),由于EMD分解的完备性,不同的IMF代表了不同的频率分量,即得到基波和各次谐波IMF分量,从而完成谐波的检测。在分解中,采用三次样条插值求包络曲线,并通过对信号两端添加极值点的方式来减轻边界效应的影响。仿真结果表明,基于EMD的谐波检测方法,不仅能有效地将电力系统谐波自适应地分离出来,而且对于有源滤波器的谐波检测具有较高的实时性。

Empirical mode decomposition （EMD） is adopted to detect harmonics of APF in this paper. The power current signals are decomposed by EMD method, and then a series of intrinsic mode function （IMFs） are obtained. Owing to different IMFs correspond to different frequencies and the integrity of EMD, harmonic detection can be achieved. During the procession of EMD, the cubic spline interpolating polynomial is employed. Furthermore, in order to alleviate edge effects the method of adding extremes is used. Simulation indicates that harmonic detection of APF based on EMD can not only separate the harmonic efficiently and self-adaptively, but also have high real-time.