Neuro-fuzzy-based scheme for stator winding protection of synchronous generator

Abstract

A neuro-fuzzy-based scheme for the protection and monitoring of a synchronous generator has been presented. The proposed scheme consists of two parts. In the first part, various operating states (normal operating, internal fault, and external fault states) of a synchronous generator have been simulated. Modeling of generator has been done in terms of phase quantities. In the second part (i.e., the fault-detection part), current samples from the line-side and neutral-end in addition to the samples from field current are used. An artificial neural network has been used for detecting the operating state, whereas fuzzy logic is used as a final trip decision-maker. The proposed technique has been tested extensively by using data that were generated using the PSCAD/EMTDC software package.