A new acoustic emission sensor based bearing fault diagnostic technique

Bearing fault diagnosis by quantifying acoustic emission (AE) data has been an area of interest for recent years due to the numerous advantages over vibration based techniques. However, most AE based methodologies to date are data-driven technologies. This research takes a new approach combining a heterodyne based frequency reduction technique, time synchronous resampling (TSR), and spectral averaging to process AE signals and extract condition indicators (CIs) for bearing fault diagnosis. The heterodyne technique allows the AE signal frequency to be shifted from several MHz to less than 50 kHz, which is comparable to that of vibration based techniques. Then, the digitized signal is band pass filtered to retain the information associated with the bearing defects. Finally, the tachometer signal is used to time synchronously resample the AE data to allow the computation of a spectral average and the extraction of CIs for bearing fault diagnosis. The presented technique is validated using the AE signals of seeded fault steel bearings on a bearing test rig. The result is an effective physics based approach validated to diagnose all four fault types: inner race, outer race, ball, and cage.