Detection of shorted turns in the field winding of turbogenerators using the neural network MLP

The work presents a digital prototype for simulation of shorted turns detection system in the field winding of the turbogenerators. In the development of the system, the neural network MLP, trained by backpropagation algorithm, was used to approach the problem of recognition of patterns that characterized shorted turns, in the field winding of the turbogenerators. The patterns of defects were obtained by simulation of turbogenerators. The work also presents a range of values limit of powers for turbogenerators operation, considering the short circuits conditions in field winding