Digital protective relaying using an adaptive neural network

The paper presents a new approach to the estimation of voltage and current phasors of a faulted power system by using an adaptive neural network represented by adalines. The neural estimator uses a nonlinear weight adjustment algorithm for the effective rejection of DC offset and noise. The fault apparent resistance, reactance and fault location of a typical transmission line are calculated using the estimated phasors. The results are quite comparable to the Kalman filter approach which requires more computation than the existing approach