Fault location in distribution systems: A Method Considering the Parameter Estimation Using a RNA Online

This paper proposes a new methodology for single-phase short-circuit location in power distribution systems. The feeder behavior during single-phase faults is modeled in terms of the distance of the fault, the network parameters and the currents and voltages measured only at the substation. Thus, the proposed method does not require the installation of any additional measurement equipment in the network. The capacitances of the network were also considered in the model used, making it closer to the real system. The main innovation of the proposed method is the use of a neural network, which does not require prior training, to estimate the unknown parameters of nonlinear equations that model the feeder behavior during faults. The performance of the proposed method was evaluated in the test system IEEE 34 bars through the variation of fault resistance and load feeder. Furthermore, it was also evaluated the influence of the estimated short-circuit current in the determination of the fault location. The good results obtained, combined with simplicity and low cost of implementation, make the proposed method promising for application in a real feeder.