Abstract
A hybrid technique for diagnosing broken rotor bar fault of induction motor using Multi-Wavelet Transform (MWT) and radial basis neural network is presented. The stator currents of induction motor are preprocessed using multi-wavelet transform and the decomposed components are obtained. Then, these features are given as input to the neural network to identify fault. This paper compares the proposed hybrid technique with MWT-Feed Forward Neural Network (FFNN) and Discrete Wavelet Transform-FFNN techniques. These techniques are compared using the concept of classifier performance. From the simulation results, it is evident that the proposed method is superior to other methods with regard to objective proposed.
Highlights
Squirrel cage induction motors are considered to be workhorse in number of industrial applications (Aderiano et al, 2008)
The multiwavelet transform is applied to the stator current signal and features are extracted by decomposing the signal
The analysis showed that the proposed hybrid technique (MWT-Radial Basis Neural Network (RBNN)) is better than that of MWTFFNN and DWT-Feed Forward Neural Network (FFNN)
Summary
Squirrel cage induction motors are considered to be workhorse in number of industrial applications (Aderiano et al, 2008). The non-stationary nature of the stator current of induction motor leads to the use of wavelet transform for condition monitoring. Broken rotor bars are one of the easiest induction motor faults to be detected using steady-state stator current condition monitoring.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
