ANN-based techniques for estimating fault location on transmission lines using Prony method

A system is proposed to locate faults on transmission lines using the traveling wave phenomenon. Prony method is used to analyze the voltage or current signal at the local bus and extract its modal information. This information represents the traveling wave generated by the fault and can be used to estimate its location. Artificial neural networks (ANNs) are used to estimate the fault location based on the modal information. Two schemes are presented: the first with a voltage input signal; and the second with a current input signal. Training and testing data are obtained by ATP-EMTP simulations. Tests were done to study the effect of fault resistance variation and fault incidence angle on both schemes. The system has the advantages of utilizing single-end measurements, using either a voltage or current signal; no pre-fault data is required.