REAL TIME PREDICTION OF COHERENT GENERATOR GROUPS

Abstract

This paper presents a scheme for predicting coherent generator groupings that may result following a disturbance that leads to transient instability. The proposed scheme uses rotor speed deviations of the individual generators in the power system as input data, and a multilayer perceptron neural network as decision making tool. The speed deviations are extracted 5 cycles after the tripping of a line or bus following a disturbance. The proposed scheme is able to predict coherent generator groups before they are formed. The prediction accuracy of the scheme for 114 fault cases was found to be 91.22%