Typical Fault Cause Recognition of Single-Phase-to-Ground Fault for Overhead Lines in Non-Solidly Earthed Distribution Networks

The Single-phase-to-ground fault (SPGF) affects reliability and safety of distribution system greatly. Accurate online recognition of fault causes can help find weak components in time and improve maintenance efficiency. A fuzzy inference system of typical fault cause recognition of SPGFs is proposed for overhead lines in non-solidly earthed distribution networks. Based on the survey of fault causes in a certain city in China, artificial grounding experiments are designed for six typical fault causes, including arrester breakdown, insulator flashover, line-to-crossbar discharge, line fallen on wet mud, line fallen on wet sand, and line fallen into pool. The waveform data for various causes are collected. Through multi-dimensional analysis on waveform data of various causes, five features are extracted: self-recoverability, zero current time, transition time, degree of distortion, and randomness. Also, their distribution ranges are determined. Based on above, a FIS for cause recognition for SPGFs is established. The experimental results and the comparison with BPNN model show that the proposed method has good performance and feasibility.