Evaluation of two knowledge-based fault locators for power distribution systems

This paper presents the implementation of two knowledge-based methods to locate faults on power distribution systems with distributed generation, to improve the power quality as demanded by the regulated electrical markets. Fault location methods could use additional measurements as new information to improve the results. ID3 decision tree and k nearest neighbor methods are implemented using measurements at substation and at the distributed generation nodes. The methods are validated using a modification of the IEEE 34-node test feeder, with two distributed generators. Several operating scenarios are considered by varying the model parameters of the power system. Additionally, a sensitivity analysis is performed to find the most critical parameters that affect the fault location methods. The results show the good performance of the fault locators and help to determine the most critical parameters of the power distribution system.