Power Qualityc Lassifiction Method Based On CEEMD And Improved S Transform

The problem of power quality disturbance is complex and diverse,and the detection and recognition of various disturbance types is the precondition of improving the power quality.Aiming at the problem of voltage disturbance parameter detection and feature extraction,this paper focuses on improving EEMD and optimizing S transform(S-Transform,ST)method,and uses multi-feature extraction and multi-classifier comparison to achieve the classification of power quality disturbance signals.First of all,aiming at the EEMD(Ensemble Empirical Mode Decomposition,EEMD)are modal aliasing phenomenon,running slow and other issues,we propose CEEMD(Complete Ensemble Empirical Mode Decomposition,CEEMD),and introduce the decomposition method of principle and simulation analysis in detail.Finally,we prove that the decomposition method has good effect and advantages of decomposition in the calculation of the cost.Secondly,aiming at the difficulty of parameter optimization of generalized S transform(Generalized S-Transform,GST),an optimized generalized S transform(Optimized Generalized S-Transform)is proposed and applied to power quality disturbance parameter detection.The parameters of the fundamental frequency points are set independently to highlight the characteristics of time domain disturbance,which is convenient for parameter optimization of other frequency points,focusing on the frequency domain perturbation,so that OGST can simultaneously represent the high-accuracy time-domain disturbance and frequency-domain disturbance information.The optimization parameters are put forward,which provides adaptability and theoretical basis for calculating parameters.The detection of disturbance parameters such as disturbance time,disturbance amplitude,harmonic component and so on is realized.The analysis of simulated and measured data shows that OGST has strong anti-jamming ability and high detection precision..Aiming at the characteristics of OGST,a set of feature vectors is proposed,which is the standard of classification,and the rationality of the classification standard is explained.Finally,we introduce the principles of several classifiers,especially introduce the classification principle and classification process of ELM classifier,and use ELM classifier to classify 12 kinds of power quality disturbances.The analysis results of simulation and engineering data show that the proposed method has better noise resistance and higher classification accuracy.Finally,the basic classification performance of the above classifiers is compared,and time-cost advantages and accuracy of classification in this paper are highlighted,and the applicability of ELM classifier to power quality disturbance signals is highlighted.