Detection and Classification of Induction Motor Faults Using DWPD and Artificial Neural Network: Case of Supply Voltage Unbalance and Broken Rotor Bars

EasyChair Preprint 7756

Abstract

It is obvious that time-frequency condition monitoring approaches have known better accuracy when combining it with AI techniques; however, this type of combined techniques is shyly applied, so it needs more encouragement. This article aims to detect and classify induction motor's faults, supply voltage unbalance and broken rotor bars, under several loads, with Artificial Neural Network (ANN), using as indicators: kurtosis and energies values. These values are calculated from Discrete Wavelet Packet Decomposition (DWPD) sub-bands of the stator current signals. The signals are obtained from the simulation of the squirrel cage induction motor (IM). The approach adopted here, is to treat the machine in terms of circuit. The occurred results are discussed.

Keyphrases: Artifitial Neural Network, DWPD, Energy, Induction motor, Kurtosis, Supply voltage unbalance, broken rotor bars

@booklet{EasyChair:7756,
  author    = {Meriem Behim and Leila Merabet and Salah Saad},
  title     = {Detection and Classification of Induction Motor Faults Using DWPD and Artificial Neural Network: Case of Supply Voltage Unbalance and Broken Rotor Bars},
  howpublished = {EasyChair Preprint 7756},
  year      = {EasyChair, 2022}}