Grounding Faulted Feeder Detection Based on Fuzzy Clustering Algorithms

Abstract

Many power distribution systems have been operated with neutral ineffectively earthed, and earth fault current is no more than a few tens of amperes. Traditional earth fault detection methods based on zero sequence current has poor precision and sensitivity in this case. For improvement, a novel principle for feeder grounding fault protection based on fuzzy clustering algorithms is presented in this paper. First, the historical data are divided into two groups by fuzzy clustering algorithms. The space relative distance among detected pattern and two cluster centers is then calculated to discriminate the faulted feeder. It can detect high impedance grounding faults (HIGF). The scheme has been verified by EMTP simulation, and results show that the proposed scheme always trips all kinds of grounding faults with high sensitivity and robustness in neutral ineffectively earthed power systems (NIEPS).