Application of Feed Forward Neural Network to Differential Protection of Turbogenerator

This paper discusses the application of MultiLayer Feed Forward Neural Network (MFNN) for the differential protection of the turbogenerator based on pattern classification. The cases of all the possible internal faults in the stator of the generator with lap winding have been simulated using Modified Winding function Approach. The simulated fault currents in the phases and their parallel paths at the terminal and the neutral end have been considered for training and testing of the proposed MFNN. Different networks has been accordingly trained and tested to detect, identify and classify the internal fault in the stator. From the test results it is clear that the proposed networks are capable of correctly identifying and classifying the fault signal. KeywordsSynchronous generator; Turbogenerator; Differential Protection; Pattern Classification; Artificial Neural Network; Feedforward Neural Network.