An Improved Wavelet Transform Power System Harmonic Detection Anti-aliasing Method

Harmonic detection is the foundation of harmonic control.Fourier transform is an effective method of harmonic detection,mainly used to detect stationary signals.Power system harmonics are mainly for the detection of non-stationary signals,and the Fourier transform seems powerless.The wavelet transform has the function of time-frequency localization,which can realize the detection of stationary signals and non-stationary signals at the same time,which meets the requirements of harmonic detection in power systems.However,the wavelet transform is prone to frequency aliasing during the harmonic detection process.The high-frequency information is mixed into the low-frequency information,the fundamental wave and the harmonics cannot be effectively separated,resulting in greatly reduced accuracy of harmonic detection and power system harmonics.It has great instability and complexity.Therefore,the effective methods to suppress or even eliminate wavelet aliasing has a very important application value for power system harmonic detection.1).In this paper,The wavelet aliasing phenomenon is analyzed from the mathematical and physical aspects respectively,and the traditional wavelet transform method is used to simulate the power system harmonics,The results show that spectrum aliasing is obvious.On this basis,the comparison of the existing three methods of eliminating aliasing,Due to the disadvantages of poor aliasing suppression and large computation,And propose a new algorithm for inserting odd extraction branch in the Mallat algorithm basic links in the existing and make further compensation,The results show that the new algorithm not only has good aliasing suppression effect,but also reduces the computation.2).In this paper,The characteristics and properties of several commonly used wavelet basis functions are analyzed,and the db20 wavelet with large vanishing moment order and good frequency band division is selected as the wavelet basis function of power system harmonic analysis.According to Nyquist sampling theorem and harmonic detection of power system,the appropriate sampling frequency is selected.3).The new algorithm is applied to the harmonic simulation experiment of power system with steady state harmonic components and simulation experiment with steady and transient harmonic components.Through analysis and example comparison,this method can not only separate the steady state harmonic signal,but also determine the starting and ending time of the transient harmonic.The simulation results are compared with the theoretical values,and the results show that the simulation values are very close to the theoretical values,which proves the effectiveness of the method.Through the analysis of simulation results,a new algorithm for inserting odd extraction branch in the Mallat algorithm basic links in the existing and make further compensation can effectively eliminate the spectrum aliasing phenomenon in the wavelet transform process,It can effectively separate the fundamental wave signal and the harmonic signal,to determine the occurrence and cut-off time of the non-stationary signal,and improve the accuracy of the harmonic detection of the power system.It has high application value in the harmonic detection.