Mehdi Samiei Sarkhanloo - Department of Electrical Engineering, Ardabil Branch, Islamic Azad University, Ardabil, Iran
Davar Ghalledar
Mohammad Reza Azizian - Sahand University of Technology, Tabriz, Iran

In this paper, first space vector model of symmetrical induction motor is extended to asymmetrical structure due to stator turn to turn fault. Then a novel scheme in order to diagnosis ofthis fault is proposed where a neural network with space vector pattern has used. In this scheme, new indexes using space vector properties of stator voltage and current, is defined that oscillation of this indexes is due to unbalancing of stator voltage and current.This scheme using suitable learning of neural network is able to recognize stator winding turn-fault regardless admissible unbalancing of the inputvoltage, balanced variation of three phase voltage, and admissible variation of load torque. Simulations were performed in Matlab/Simulink environment to validate the proposed method

Induction Motor; Stator Turn to Turn Fault; Space Vector; Fault Diagnosis; Neural Network