Research On Fault Line Selection Method For Ineffective Grounding System Based On VMD And SVM

In China,the existing 3-66kV medium voltage distribution network mostly adopts small current grounding system.If single-phase grounding faults,which account for more than 65%of the total fault types,cannot be identified and eliminated timely,the reliability of the system will be affected seriously.What’s more,fault characteristics of single-phase grounding fault can easily be affected by factors such as grounding resistance,on-site noise,and unstable arc,which brings a lot of difficulties for accurately determining the fault line.Therefore,in this thesis,the single-phase grounding fault line selection for small current grounding system is taken as the research object,and fault feature extraction,fault line determination and line selection accuracy are studied accordingly.The steady-state and transient fault current characteristics of the small current grounding system are analyzed,and the compositions of the fault transient zero-sequence current are deduced in detail.Then,it is concluded that compositions of the transient zero-sequence current under different fault closing angles are different.According to the conclusion above,a new transient fault line selection method is proposed based on Variational Mode Decomposition(VMD)and Support Vector Machines(SVM).First,VMD is introduced into the process of the zero-sequence current feature extraction,coupled with the above conclusion on the differences among zero-sequence current components and the influence of decomposition parameters on VMD performance,the VMD parameters are determined.Then,with the Hilbert envelope energy and the analysis of waveform correlation,the polarity and amplitude of the transient zero-sequence current is measured.Finally,the fault line is determined based on the SVM theory.By MATLAB simulation analysis,it is verified that the performance of VMD is superior to the empirical mode decomposition(EMD),and the method can achieve the fault line selection under different fault conditions and has the ability to select the fault line under the influence of the arc as well.In order to further improve the accuracy of the above method,it is improved as follows:First,the characteristics of the fifth order harmonic and the zero-sequence energy are applied in the process of fault line selection to avoid the single line selection criterion.Then,SVM is replaced by the Least Squares Support Vector Machine(LSSVM)to improve the classification accuracy.Besides,in order to improve the optimization capacity,Improved Fireworks Algorithm(IFWA)is used to optimize the parameters of LSSVM and to further improve the classification accuracy of classifier.Finally the multi-criterion fusion fault line selection method based on VMD-IFWA-LSSVM is established.The simulation shows that IFWA has better overall optimization capacity and the performance of it is better than particle swarm algorithm,genetic algorithm and fireworks algorithm.Meanwhile,the accuracy rate toward the test set of the multi-criterion fusion fault line selection method is up to 97.5%.It means this method has the value of application and practical.