Research On Power Quality Analysis Method Based On GA-VMD Harmonic Detection And High Oeder Singular Spectrum

The power quality disturbance signal is complex and diverse.The accurate identification of the disturbance signal can provide reference information for improving the power generation level of the power sector and improving the power quality,which plays an important role in the stable and safe operation of the power system.Based on the description of power quality characteristics,disturbance types and related characteristic parameters,this paper proposes a power quality disturbance recognition method based on high-order singular spectrum analysis and decision tree.First of all,due to the complicated working environment of the power grid,the obtained power quality signal is mixed with various noises.Therefore,this paper proposes signal denoising methods based on variational mode decomposition and permutation entropy and an improved wavelet threshold denoising method to preprocess the obtained power quality signal,this paper uses permutation entropy as the basis for evaluation,and provides reconstructed component screening criteria.Improving the threshold function and threshold selection,it can realize the adjustment of the threshold function by introducing a power function and adjustment factor to construct a new threshold function,and the hierarchical selection threshold is more targeted.It is verified by experiments on the noise-disturbed signal that both methods are good.The denoising effect can achieve effective separation of the disturbance signal and noise.VMD-PE is more suitable for complex disturbance signal denoising,while improved wavelet threshold denoising is more suitable for single disturbance signal denoising.Secondly,in the power quality disturbance signal,the harmonic signal is mixed with various frequency components,which is more complicated and more difficult to detect than other disturbance signals.Therefore,the paper proposes a genetic algorithm to optimize the variational mode decomposition detection method for harmonic disturbance signals.The genetic algorithm is used to realize the automatic selection of parameters.The experimental results show that the method has good detection results.Then,according to the characteristics of 9 kinds of power quality disturbance signals and 4 kinds of compound disturbance signals,high-order singular spectrum analysis is introduced,and the paper makes comparison and analysis for the selection of higher-order cumulant orders,and fourth-order singular spectrum analysis is selected as the method.The high-order singular spectral feature quantization analysis of various disturbance signals shows that the high-order singular spectrum can characterize the inherent characteristics of the disturbance signal.Two features of high-order singular spectral entropy and signal energy are taken as the classifier input,and also are evaluated and analyzed.The experimental results show that these two features have a good discrimination and the combination of these two features can effectively distinguish Kind of disturbance signal.Finally,the simulation data and measured data are taken as the research object,and the decision tree is selected as the classifier to identify the disturbance signal.The experimental results verify the effectiveness and superiority of the proposed method.