Pattern classification based intelligent numerical protection of turbogenerator

Abstract

This paper discusses the application of MFNN for the protection of turbogenerator against internal faults in any winding of the stator. The network has been used as pattern classifier for detection, identification and classification of the internal faults. The full cycle data of simulated fault currents in the phases and their parallel paths have been used for training and testing of proposed neural networks. The synchronous generator model based on direct phase quantities and modified winding function approach has been used to simulate different types of internal and external faults using electrical parameters of generators being used by utilities. All possible cases of internal faults in the stator winding have been considered and networks have been trained and tested accordingly. The test results of the selected architecture of the networks for different protection schemes indicate that the fault signal can be correctly identified and classified by the proposed networks.