Monitoring of Bearing Based on High-Order Spectra Analysis of Electric Power Supply Signals

Failures of bearing would cause sudden downtime and severe accidents that may require significant maintenance and economic consequences. Therefore, condition monitoring of bearing is extremely important for ensuring reliable and efficient operation. The analysis of electric power supply signals offers a non-intrusive and cost-effective method for detecting various incipient faults in rotary machines. However, using inverter or variable speed drive would influence the electric power supply signals, causing difficulties in detection and diagnosis results. To reduce these influences and develop effective detection techniques, this paper focuses on using electric power supply signals as an approach to detect incipient faults in the bearing of driving induction motor based on high-order bispectral analysis. It takes the advantage of providing abundant information from electric signatures of downstream driving motors to improve the performance of bearing’s health monitoring. Modulation signal bispectrum (MSB) and cyclic bispectrum (CCB) exhibit significance to detect normal and abnormal bearing conditions, whereas remarkable effort is dedicated from MSB result due to the inclusion of phase information with amplitude modulation in both lower and higher sidebands. It can conclude that effective measurements of electrical power supply signals utilize reliable fault diagnosis information in the mechanical components of the driving motor.