Research On Harmonic Detection Method Based On Sliding Window Iterative DFT And Waveletpacket Analysis

In recent years,with a variety of electronic equipment and nonlinear load connected to the grid,so that grid harmonic source surge, failure occurred more frequently,and grid suffered noise pollution tends to be serious. The noise and harmonics will bring a serious threat to the normal operation of power system,and the adverse effects on the power quality. For governance of the harmonic in the power grid,the prerequisite is detection and analysis to harmonic.While the harmonic component analysis,we must first be de-noising of power signal preprocessing to improve detection accuracy. Therefore,accurate harmonic detection can be used as a good basis for harmonic treatment,to improve power quality has important significance.This topic first briefly summarizes the background and concepts of power harmonic,the harm of harmonic and several current harmonic detection methods and trends; And it describes the definition,sources and characteristics of harmonics,and cited a typical waveform of harmonics,a brief description of the principle and standard of harmonic content analysis.Secondly,this paper studied the power harmonic signal de-noising pretreatment methods, explored the principles, advantages and disadvantages of empirical mode analysis (EMD) and de-noising wavelet analysis, combining the advantages of these two methods,it presents a power harmonic signal preprocessing method based on EMD-wavelet analysis.Studied the EMD decomposition principle,the principle of wavelet denoising and the selection of related parameters.Through Matlab simulation,respectively with three methods of EMD,wavelet and EMD-wavelet to do denoising pretreatment for power signal from Matlab library (leleccum signal) and several kinds of power harmonic signals with noise,and the signal-to-noise ratio and mean square error as an argument,compare the three methods,the denoising effect of the EMD-wavelet method in electric harmonic signals denoising pretreatment process effectiveness,prepare a premise for the following harmonic detection and analysis.Again,this paper discusses the electric power harmonic signal component analysis method,which is the focus of this article.The current main methods of harmonic analysis is divided into two kinds,one is to obtain fundamental,then get the other harmonics;the other is power harmonic signal is analyzed directly to obtain further fundamental and harmonic.In this paper,we study the principle and characteristics of Fourier analysis,sliding window iterative DFT,wavelet analysis and wavelet packet analysis methods,combined with their advantages and disadvantages,the sliding window DFT method and the wavelet packet analysis methods are combined to give a combination detection method.Through matlab simulation,respectively using Fourier analysis method,the sliding window iterative DFT method,iterative sliding window DFT-wavelet analysis method and sliding window iterative DFT-wavelet packet analysis method to detect and analyse the constructed steady and unsteady harmonic signal.The results show that the EMD-wavelet method in harmonic signal denoising preprocessing,signal-to-noise ratio and variance parameter values were relatively good,can be used as a reference method of pre-treatment.Sliding window iterative DFT-wavelet packet analysis method using the sliding window iteration DFT methods to separate the fundamental wave and harmonic firstly,then use wavelet packet analysis method to get every harmonic signal.This method in time-frequency domain has better signal characteristics,but also to the steady and non steady state harmonic signal detection and analysis, can be used as a reference method of harmonic analysis.