Enhancing Accuracy While Reducing Computation Complexity for Voltage-Sag-Based Distribution Fault Location

A fault-location method for radial distribution systems is proposed in this paper. The proposed method uses voltage and current phasors from feeder root and voltage sags measured at sparse nodes along the feeder, and pinpoints faults to the nearest node. Decision-tree (DT)-based fault segment identification is introduced before the process of node selection to reduce the computational complexity and improve fault-location accuracy. The method has been implemented on a practical distribution system and tested under a large number of fault scenarios. Test results are compared with those from the traditional voltage-sag-based fault-location algorithm using the same inputs, and the conclusion is that the proposed method can achieve more reliable results while maintaining computational simplicity. A quantitative method to suggest the optimal placement of measurement units based on the DT variable importance is proposed at the end.