Optimization of gmr sensor location on induction motor for online rotor fault prediction using wavelet transform and machine learning

Abstract

Induction motor (IM) is used in various industrial applications newlinebecause of reliability and robustness. In Induction Motor (IM), broken rotor newlinebar fault leads to excess amount of current flow in the stator and causes newlineunbalanced magnetic field distribution between stator and rotor. The newlineunbalanced magnetic field damages the stator winding and increases the newlinevibrations in IM. Condition monitoring based fault detection of motor at early newlinestage is essential for smooth run of motor. IM fault detection techniques are newlineclassified as three categories such as Signal-based techniques, Model-based newlinetechniques, Knowledge-based techniques. In signal-based techniques, newlineparameters such as current, voltage and leakage flux are used for continuous newlinemonitoring of Induction motor. Motor Current Signature Analysis (MCSA) is newlinewidely used for rotor bar fault detection. Many numbers of MCSA based newlinemethods are developed for rotor bar fault detection through spectral analysis newlinetechniques such as autoregressive-based spectrum methods, Wavelet newlineTransforms, Taylor Kalman approach, low-frequency, load torque newlineoscillations and high-resolution parameter estimation. newline