Research On Power Harmonic Detection Technology Based On Variational Mode Decomposition

With the grid connected operation of new energy distributed power supply and the extensive use of various power electronic components in the power system,the security and stability of the power system has been widely concerned.These nonlinear loads and equipments aggravate the harmonic pollution of the power system.Accurate measurement of harmonic components and processing of these harmonic components are of great sign-ificance to the safe and reliable operation of power system.This paper starts from the following aspects:1.Analyzing and comparing several traditional harmonic detection methods,and introduce the signal processing method of variational mode decomposition and its application in various fields.Considering the problems of traditional signal processing methods,the Variational Mode Decomposition(VMD)algorithm is introduced into the field of harmonic detection.2.The basic theory and signal processing flow of VMD algorithm are introduced,and the same simulation signal is processed together with Empirical Mode Decomposition(EMD)algorithm.The simulation results show that VMD algorithm can effectively avoid the problems of mode aliasing and false components existing in EMD algorithm,which provides theoretical guidance for subsequent chapters.3.Aiming at the problem that the EMD harmonic denoising algorithm loses effective information and takes a long time in the process of noise reduction,the VMD-SVD harmonic denoising algorithm is proposed for noise reduction.By comparing the simulation results with EMD denoising algorithm,we can see that VMD-SVD denoising effect is better than EMD denoising algorithm,and the processing time is reduced.The simulation and experimental results verify the effectiveness of the VMD-SVD algorithm,which can ensure the accurate detection of noisy signals.4.In order to further improve the detection accuracy of harmonic parameters,VMD algorithm is used for frequency detection and HT algorithm is used for amplitude detection.The premise of frequency detection of VMD algorithm is to satisfy the optimal K value de-composition of VMD algorithm.The value of K selection method based on the change of instantaneous frequency mean is proposed.EMD selection method,correlation coefficient method,spectrum selection method and instantaneous frequency mean method are used to select the optimal decomposition value of K simulation signal VMD.The comparison of four methods shows that the instantaneous frequency mean method can select the optimal K value simply and quickly.Aiming at the problem of endpoint effect in HT amplitude detection,on the premise of ensuring the optimal K-value decomposetion,directly selecting the middle segment data for amplitude detection can avoid the impact of endpoint effect on detection.On the basis of the two improved methods,Improve VMD-HT(IVMD-HT)harmonic parameter detection algorithm is proposed,and SVD algorithm is used to pre denoise the noisy signal.Simulation and experimental data verify the effectiveness of SVD-IVMD-HT algorithm,and the detection accuracy is also improved.5.According to the actual distribution network signal,three kinds of harmonic signal models are constructed by using MATLAB platform,and two kinds of algorithms are used for processing.The processing results show that both algorithms can realize harmonic signal parameter detection,and the harmonic detection accuracy of SVD-IVMD-HT algorithm is higher than VMD-SVD-Prony algorithm,but the signal processing time of VMD-SVD-Prony algorithm is much lower than SVD-IVMD-HT algorithm.The result shows that the two algorithms can achieve complex harmonic signal parameter detection,and different harmonic detection algorithms can be selected for signal processing according to the actual needs.