Induction Motor Bearing Fault Diagnosis Using Taylor-Fourier Filters on Stray Flux Signals

Bearing ball fault detection has been a topic of great interest since it is the most common type of fault in induction motors. Bearing faults are usually detected by vibration analysis, given the mechanical nature of the fault. However, vibration signals may suffer from noise and even the effect of vibration from other surrounding machines. In this regard, flux analysis has emerged as a reliable source of information on the machine status for fault detection. In this work, two outer race bearing fault damage levels are detected by comparing the amplitude estimations given by the Taylor-Fourier transform in a specific frequency selected with kurtosis-base criteria. Results suggested that for all the scenarios, the fault was successfully detected.