Fault-section identification for hybrid distribution lines based on principal component analysis and Euclidean distance

Abstract

In this paper, we present a single-ended principal-component-analysis-(PCA-) based fault-section identification method for a hybrid distribution line (non-effective grounding system), which is an overhead line combined with an underground cable when a single-line-to-ground (SLG) fault occurs. Ground-mode currents are extracted from the single-ended measuring points when faults occur in underground cables or overhead lines, and then the method of fault-section identification based on PCA is utilized to deal with the ground-mode or aerial-mode currents to realize the fault-section identification of the hybrid system. First, a loading plot of PCA is used to compare the projection of ground-mode currents on the PC1 coordinate axis to 0. Then, Euclidean distance is used to realize the fault-section identification to improve the reliability of criterion and avoid threshold setting in the PCA method. The performance of the method is tested in 35 kV hybrid distribution lines for a distribution system based on PSCAD/EMTDC for different fault locations, different fault inception angles (FIAs), and fault resistances with satisfactory results. The impact of distributed generation (DG) and lengths of cable and overhead lines on the accuracy of the proposed method is also investigated.