Fault Line Selection and Section Location in Distribution Network Based on Sparse Representation

Abstract

Distribution network lines are widely distributed with high failure probability. Accurate identification of fault lines and fault areas can shorten the troubleshooting time. Based on the time-domain information of three-phase current(TPC) and zero-sequence current(ZSC) before and after faults, a fault line selection(FLS) and fault section location(FSL) method based on sparse representation is proposed, which is applicable to all kinds of short-circuit faults in distribution network. The proposed method does not rely on Shannon Theory for fault current collection. The TPC and ZSC of each feeder of the bus and each measuring point of the fault line before and after the fault are collected, and spliced according to the sequence of A phase, B phase, C phase and zero sequence. The K-Singular Value Decomposition algorithm (K-SVD) is used for dictionary learning to construct an over complete dictionary that accurately matches the fault current characteristics; The Orthogonal Matching Pursuit algorithm (OMP) is used for adaptive dictionary sparse decomposition. The analysis shows that the maximum sparse coefficients (MSC) of fault line (FL) and non-fault line (NFL), fault section (FS) and non-fault section (NFS) fault current splicing signal have different characteristics. Therefore, the MSC of fault current splicing signals at each feeder of the bus bar and each measurement point of the fault line is used to construct the fault line selection and fault location criteria. MATLAB/Simulink simulation verifies the effectiveness of the proposed method.