Research On Method Of Power Quality Disturbance Analysis

With the development of power industry and the widely use of electronic devices and impact loads, the power quality problems become more and more prominent. It also becomes the focus of power supply department and the electricity users. This paper takes power quality disturbance signal as research object, using Discrete Fourier Transform(DFT), Time-time Transform(TT), Local Mean Decomposition(LMD) etc signal processing methods to research the detection and classification method of power quality disturbance.This paper introduces the research status of power quality disturbance analysis,it also analysis and compare several different feature extraction method and classification method of power quality disturbance. It expounds the harm and control measures of harmonic, flicker and unbalanced three-phase voltage power quality disturbance, also discuss the standards of power quality disturbance. This paper introduces the basic principle, characteristics and realization method of TT-transform and LMD-transform, which are applied to the power quality disturbance analysis.An approach based on TT-transform was proposed for detecting and classifying power quality disturbance. According to the frequency characteristic curve of the DFT of the diagonal elements corresponding to the TT-transform matrix and the diagonal elements corresponding to the TT-transform modular matrix,three characteristic values are extracted to do preliminary judgment about whether the disturbances contain harmonic. For disturbances without harmonic or with harmonic,the start-stop time and the amplitude of power quality disturbance can be detected and calculated by different disturbance characteristic curves according to TT-transform and two-TT-transform,then the fourth characteristic can be extract. At last, the classification of power quality disturbance is done through the rule tree.A identification and classification method of power quality disturbance based on LMD and Support Vector Machine(SVM) is proposed. The Product function(PF)components and allowance are obtained by LMD-transform, the instantaneous amplitude of PF component is its corresponding envelope signal, while the instantaneous frequency of PF component can be obtained by pure frequency modulation function. Then through detection of the disturbance’s mutation momentand calculation of amplitude of power quality disturbance, three characteristic values are extracted. Finally, using SVM method realizes the classification of power quality disturbance.By the two mentioned methods, the characteristic values of voltage sag, voltage swell, voltage interruption, voltage flicker, transient oscillatory, transient pulse,harmonic, voltage sag with harmonic, and voltage swell with harmonic are extracted,thus the classification of power quality disturbances is done by the combination of rule tree or SVM. Simulation results verify that the two proposed method can both effectively extract characteristics of power quality disturbances, detect and classify disturbances accurately, and they also have a certain ability to resist noise.