Testing Two Methods for Earth Fault Location in Closed-Ring Operated Distribution Networks: Lessons Learned from Field Tests

To test two newly developed methods for location of earth faults in closed-ring distribution networks, field tests have been carried out in a ring-operated 22 kV resonant grounded network. The results demonstrate poorer accuracy than expected, so this paper presents analyses of some likely sources of error. First, the impact of inaccurate line data, measurement inaccuracies and phasor estimation issues are described and studied analytically. Examples from the field tests demonstrate that all of these are likely contributors to the results, although some more than others. For one of the methods based on negative sequence currents, the inaccurate measurement of the three phase currents appears to be the most likely single reason for the low accuracy observed. For the method based on zero-sequence quantities, inaccurate line data is assumed to be another key factor. From the analyses, it is concluded that dedicated sensors are needed for accurate negative sequence current acquisition, and this is currently not available in the test network's substations. For line parameter acquisition it is argued that manufacturer data and analytically obtained data will not be accurate, and measurement of the actual lines and cables in the network is the preferred approach.