Real time wireless condition monitoring of induction motor

Abstract

Induction motor (IM) plays a pivotal role in majority of the industrial applications. This makes it all the more important for condition monitoring. Stator current monitoring is preferred, as the stator current possesses principle discriminative features, which are the key for effective detection and classification of faults in the IM. This paper portrays a methodology to wirelessly monitor the stator currents of the IM to ascertain the health of the IM and the nature of the supply. The IM stator current is wirelessly acquired using a ZigBee based Wireless Sensor Network (WSN) with a sampling frequency of 1.8 kHz. Wavelet transform is applied to the stator current of the IM for the extraction of the feature of the broken rotor bar fault and voltage sag/ swell, if any, present in the supply. Daubechies wavelet −9 (Db9) is used for the same. The fifth level approximate wavelet coefficients are given as inputs to the Artificial Neural Network (ANN) for the classification. The ANN records a classification accuracy of 91.6%.