Ma. Elsharkawi et al., LOCALIZATION OF WINDING SHORTS USING FUZZIFIED NEURAL NETWORKS, IEEE transactions on energy conversion, 10(1), 1995, pp. 140-146
Shorted turns in field winding of large turbogenerators are difficult
to detect and localize. We propose a technique whereby shorts are dete
cted and localized using an artificial neural network with a fuzzified
output. The method is based on injecting two simultaneous and identic
al waveform signals at both ends of the field winding. Selected featur
es of the received signals are used to train the neural network. Once
trained, the neural network can detect and localize short turns in the
field winding. The proposed method is verified by a field test on 60
MVA turbogenerator. The results show that the proposed method is quite
accurate and efficient.