A fuzzy-logic-based fault recognition method using phase angles between current symmetrical components in automatic DFR record analysis

This paper presents a new fault classification technique combining accurate fault identifiers with a new fuzzy tool which deals with uncertainty and possible inconsistency. Only the three phase currents of each feeder are required to detect the faulty line and the fault type such as LG (line-to- ground fault), LLG (two-line-to-ground fault) or LL (line-to-line fault). No knowledge of the pre-fault conditions and no training from examples is required. First, we describe the fault identifiers based on phase angles between current symmetrical components. Their major advantage is the very low variation around their theoretical values regarding to important variations of the network and the fault parameters. Then, we introduce the fuzzy-logic tool comparing the standard with the new proposed tools, which gives an idea of the consistency of the found class. At last, we give some numerical examples and test results using EMTP simulations.