Intelligence Scheme for Fault Location in a Combined Overhead Transmission Line & Underground Cable

Abstract

Abstract This paper focuses on comparison of fault position schemes for a long transmission line joint with underground cable. To carry out fault location, two hybrid schemes were implemented. One is the impedance method based on modal transformation (MT) and the other one is the intelligent technique based on artificial neural network (ANN). In this paper one cycle of post fault current and voltage signals were collected initially from the transmission line ends for fault analysis purpose. The first method to analyze the fault position comprises of MT technique in which initially pre-processing of the data is done by Clarke’s Transformation (CT) and Discrete Fourier Transform (DFT). CT decouples the signal and DFT extracts the phasors. Thereafter the fault location is calculated by Power System distributed line modal and MT concept. The second method focuses on estimation of fault distance by ANN with Wavelet Transform (WT) in which WT is used to extract six statistical features in order to pre-process the raw faulted data. These features are then given to ANN for finding fault position. The enactment of the suggested scheme is verified by testing it under different type of fault position, such as variation of fault resistance, inception angle and type. After extensive simulation, it was found that intelligent scheme i.e ANN-WT schemes are capable of locating the fault more accurately and less sensitive to parameter variation than the impedance method i.e MT-CT-DFT.