A fuzzy-based fault section identification method using dynamic partial tree in distribution systems

In a distribution management system (DMS), fault section identification (FSI) is an essential function for the selfhealing capability of smart distribution networks. Conventional FSI methods based on the flag of fault indicators (FIs) cannot be properly utilized for distribution networks with distributed energy resources (DERs) since false FIs, unintentionally activated by fault currents flowing from the downstream of a fault location due to the presence of wye-delta transformers for the DER interconnection, are expected to be generated. For more accurate FSI, this paper proposes a fuzzy-based FSI method using the fault current information measured on feeder remote terminal units (FRTUs) installed at automatic switches in distribution lines (DLs). The proposed method takes advantage of the fuzzy evaluation to provide an intuitive index for decision-making by quantifying the possibility of fault for each candidate section. Furthermore, to achieve faster FSI by reducing waiting time for FSI, the partial trees using only validated FIs with fault current information for the faulty DL are dynamically generated, thereby enabling the fault section can be identified regardless of whether the fault current information from every FRTU is available or not. Case studies on IEEE 33-bus distribution system considering various fault conditions are conducted and simulation results show that the proposed method can accurately identify correct fault sections under false FI and communication failure situations with the improved time required for the FSI.

Automated summary

This paper proposes a fuzzy-based fault section identification (FSI) method that utilizes fault current information measured on feeder remote terminal units (FRTUs) installed at automatic switches. The method provides an intuitive index for decision-making by quantifying the possibility of fault for each candidate section. Additionally, the method improves the FSI time by dynamically generating partial trees using validated fault current information.