Stator faults diagnosis of the converter-fed induction motor using symmetrical components and neural networks

The paper deals with the diagnostic problems of the converter-fed induction motors in the case of stator faults. For diagnostic purposes the neural detector was developed trained and tested using measurement data of negative components of the stator current and impedance, for different levels of stator failures, load torque and converter frequencies. Feedforward neural network with very simple internal structure used for the detection of fault level gave satisfactory results, what is very important in practical realization.