Research Of The Integrative Fault Line Selection Methods For Small Current Grounding Fault Based On Information Fusion

The neutral indirectly grounded way is mostly adopted in distribution systems of 6?66kV in China,but single-phase grounding fault is easy to happen in this mode,and the fault current is weak which is easier to be influenced by many factors such as system grounding mode,line type and structure,transformer characteristic and so on,so that fault line selection is difficult.Single fault line selection method can not guarantee the accuracy of fault line selection,and integrative methods based on information fusion has became a research hotspot in the fault line selection for small current grounding fault.This paper firstly introduces the background and significance of the subject,and explained the principle and application of the existing fault line selection method.The steady-state characteristics and transient characteristics of small current grounding fault are analyzed in detail,and the simulation model of small current grounding system is built by MATLAB software,that simulation analysis is carried out under different fault conditions including different voltage initial angles,different grounding resistances,different line types and different fault location.Than combined with the fault characteristics of the small-current ground fault,the amplitude of fundamental wave and the amplitude of fifth harmonic of steady-state zero sequence current are extracted by the FFT as the fault feature of the fundamental amplitude comparison method and the fifth harmonic method,and the transient energy value is extracted by the WPT as the fault feature of wavelet packet method.To overcome the shortage of single fault line selection,this paper adpots two kinds of information fusion methods to fuse the above three fault line selection methods.Support Vector Machine(SVM)is a kind of machine learning method with excellent generalization ability for small sample and nonlinear problems,which is very suitable for the problem of fault line selection for the small current grounding fault,but its classification performance is greatly affected by the parameter setting.To avoid blindness in parameter setting,this paper uses genetic algorithm(GA)to optimize the parameters of SVM,and proposes a fault line selection based on GA-SVM.Compared with the fault line selection method based on traditional SVM,the simulation results show that this method can achieve fault line selection better.Vague set theory is a more effective method for information fusion than traditional fuzzy set theory,which takes into account membership degree,non-membership degree and hesitation degree,and this paper applied it to fault line selection to propose a method of fault line selection based on fuzzy entropy weight of vague set for the small current grounding fault.In the process of information fusion based on Vague set theory,it is a key problem to determine the weight of each fault line selection method in the score function.In this paper,the weight of information entropy is determined according to the information entropy weight method,and the weight of fuzzy theory is determined according to the traditional fuzzy theory method,and the fuzzy entropy weight is synthesized by combining them.The simulation results show that the fault line selection method overcomes the shortage of single line selection method,and verifies the effectiveness of this method under different fault conditions.