The application of ANN in fault diagnosis for generator rotor winding turn-to-turn faults

When turn-to-turn faults occurred to rotor winding of the generator, the terminal parameters of the generator will change. The condition of the rotor winding can be reflected by the terminal parameters, but itpsilas difficult to describe the relationship of fault information and terminal parameters by accurate mathematics expressions. Applying artificial neural network in rotor winding fault diagnosis can obtain a good result, and when there are faulty samples in the training samples of artificial neural network, the severity information of the generator faults can be obtained directly. But it is difficult to gain the faulty samples in practical applications. Through the analysis of magnetic motive force of the generator and application of artificial neural network for faulty samples, the fault diagnosis of turn-to-turn fault on generator rotor winding can be carried out.