A thesis submitted in partial fulfillment of the requirements for the degree of Master of Engineering. School of Environment Resources and Development

Thesis (M.Eng.) - Asian Institute of Technology, 1999

The main objective of the digital relaying of transmission line is to determine the phasor repi"esentations of the voltage and current signals from their sample values and thereafter to calculate the apparent impedance of faulty line from the relay location to the fault point. Then determine whether the fault lies within the relay's protective zone or not. Since impedance of linear system is defined in terms of the fundamental frequency voltage and current sinusoidal waves, it is necessary to extract the fundamental frequency components of voltage and current signals from the complex post fault voltage and current signals. This work presents an adaptive neural network approach for the estimation of fundamental components from the complex post fault voltage and current signals. The neural estimator is based on the use of an adaptive perceptron consisting of Adaptive Linear Neuron network called ADALINE. The learning parameters in the proposed algorithm are adjusted to force the actual and desired outputs to satisfy a stable difference error equation, rather than to minimize an error function. Three numerical tests have been conducted for adaptive estimation of fault impedance. The first is the simulated signals, the second is signals from transient analysis program EMTP and the third is signals from EGAT system captured by fault recorder. The estimator tracks accurately the fundamental components of the signals data corrupted with harmonics and decaying de component during transient period. The performance of the proposed algorithm is found superior to the recursive DFT based algoritlm1s.