This paper investigates induction motor fault detection and diagnosis using Artificial Neural Networks (ANN). The ANN techniques include feedforward backpropagation networks (FFBPN) and self organizing maps (SOM), used individually and in combination. Common induction motor faults such as bearing faults, stator winding fault, unbalanced rotor and broken rotor bars are considered. The ANNs were trained and tested using dynamic measurements of stator currents and mechanical vibration signals. The effects of different network structures and the training set sizes on the performance of the ANNs are discussed. This study shows that, while the feedforward ANNs give satisfactory results and the SOMs can classify the type of motor fault during steady state working conditions, using a combination of SOM and FFBPN techniques yields superior fault detection and diagnostic accuracy. In addition, incipient motor fault detection has been investigated. The above results show that improved induction motor maintenance strategies may be possible through the use of comprehensive on-line induction motor condition monitoring and fault diagnosis systems.
