Internal Fault Detection, Location, and Classification in Stator Winding of the Synchronous Generators Based on the Terminal Voltage Waveform.

In this paper, a novel method is presented for detection and classification of the faulty phase/region in the stator winding of synchronous generators on the basis of the resulting harmonic components that appear in the terminal voltage waveforms. Analytical results obtained through Decision Tree (DT) show that the internal faults are not only detectable but also they can be classified and the related region can be estimated. Therefore, this scheme can be used to protect the synchronous generators against the various internal faults. Fuji technical documents and data sheets for an actual salient pole synchronous generator (one unit of an Iran’s hydroelectric power plants) are used for the modeling. Simulations in Maxwell software environment are presented. All the related parameters, such as B-H curve, unsymmetrical air gap and pole saliency, slot-teeth effect, and other actual parameters, are considered to obtain a comprehensive model to generate acceptable terminal voltage waveforms without any simplification.