TRANSMISSION LINE PROTECTION USING WAVELET TRANSFORM

Abstract

The dissertation aims at detecting, classifying and estimating the location of the different shunt faults that arise in double end fed two-area connected transmission line system. To deal with the problem of detection and classification of faults, three different methods have been proposed and for estimating the location of faults, one novel method has been presented. The first fault detection and classification method uses the reconstructed approximation coefficients obtained at fourth level after the decomposition of current signals using Discrete Wavelet Transform (DWT) with 'db4' as mother wavelet. In the second method, the reconstructed detail and approximate coefficients of current signals up-to third level which are obtained after the decomposition of them using DWT with 'dbl' as mother wavelet are used as features for detecting and classifying the faults using Artificial Neural Networks and Support Vector Machines independently. Similarly, the third method uses only the reconstructed first level detail coefficients of all the three phase voltage and current signals obtained by applying DWT with 'db4' as mother wavelet are used as features for detecting and identif,'ing the type of fault using ANN and SVM separately. In the method presented for estimating the location of the fault, the reconstructed approximate coefficients obtained by fourth level decomposition of all the post-fault three phase voltages and currents using 'db4' as mother wavelet are normalized with respect to the corresponding reconstructed approximate coefficients of all the pre-fault three phase voltages and currents. Then, these normalized coefficients are used as features for ANN for estimating the location of the fault. The different fault situations simulated using PSCAD/EMTDC software and the performances of the proposed schemes are compared with some of the already existing schemes. The results obtained have turned out that the proposed methods offer better accuracy in detecting, classifying and estimating the location of faults for wide variations in different system and line parameters when compared to the existing approaches.