The detection of rotor faults using artificial neural network

The detection of broken rotor bars in tree-phase squirrel cage induction motors by means of current signature analysis is presented. In order to diagnose faults, a Neural Network approach is used. At first the data of different rotor faults are achieved. The effects of different rotor faults on current spectrum, in comparison with other fault conditions, are investigated via calculating Power Spectrum Density (PSD). Training the Neural Network discern between "healthy" and "faulty" motor conditions by using experimental data in case of healthy and faulted motor. The test results clearly illustrate that the stator current signature can be used to diagnose faults of squirrel cage rotor.