Charging Transient for Fault Location: Analytical Method versus Artificial Neural Network

In this paper, two single-ended algorithms for fault location are presented with reference to the unearthed operated sub-transmission Italian grid (60 kV). Both algorithms deals with the correlation between the charging frequency of sound phases to ground capacitances with the fault position. In the first one, the frequency response of a lumped parameter circuit in the Laplace domain is linked to the fault distance. In the second one, the frequency spectra of the transient current waveforms are used as a database for the training of an Artificial Neural Network, which identifies the fault distance by analysing the input data. The fault location accuracy of the two proposed methods are compared in order to reveal strengths and weakness of both algorithms. The developed procedures have been applied to an overhead line modelled in EMTP-rv environment, whereas the fault location algorithm has been implemented in Matlab environment.