Detection of Bearing Faults by Using Stator Current in Inverter-Fed Induction Motor

In this study, inverter-fed induction motor of bearing fault detection is realized by stator current analysis and artificial neural network. Artificial failures are created by damaging various parts of the bearings used in the experiment. Current signals receieved from the motor of the faulty bearing are examined in time and frequency dimension The differences are investigated by comparing the obtained data with the current signal of the robust bearing. The dominant characteristics of each bearing are determined as statistical and spectral so that feature exraction are performed. Failure detection and classification are realized by articial neural network trained with these determined features. Failure detection are completed by classifying 95.3% accuracy rate