Adaptive Harmonic Detection of Active Power Filter based on Improved VMD

Harmonic detection accuracy is the key factor affecting the performance of active power filter (APF). As an effective harmonic detection method, any complex signal can be adaptively decomposed into several intrinsic mode functions (IMFs) with a certain bandwidth by Variational Mode Decomposition (VMD). Aiming at the problems that VMD will be affected under the background of strong noise and the parameters are difficult to determine, an adaptive harmonic detection method based on twice VMD is proposed in this paper. Firstly, the proposed method uses VMD with a fixed decomposition layer A of 2 for denoising. Secondly, the VMD parameters are optimized from the perspective of IMF energy, and the second VMD decomposition is performed on the denoised signal. The simulation results show that, compared with EMD, EEMD and VMD, the proposed method can not only effectively reduce the fundamental extraction error, but also be suitable for aperiodic and non-stationary signals. It shows that the harmonic detection accuracy of APF can be effectively improved by this method.