BEARING INCIPIENT FAULT DETECTION: TECHNICAL APPROACH, EXPERIENCE, AND ISSUES

摘要

This paper describes a technical approach for detection and prediction of incipient bearing faults in complex machinery applications. The authors have developed and adapted the ImpactEnergy™ suite of vibration signal processing algorithms that are sensitive to bearing fatigue and wear events but minimize sensitivity to machinery noise generated by gearing. The ImpactEnergy™ algorithm suite develops an array of time and frequency domain features generated through an adaptive enveloping and demodulation scheme to provide quality diagnostic outputs for advanced machinery health monitoring. The technique reviewed here provides early detection and severity assessment of bearing faults by utilizing multiple regions of the vibro-acoustic energy spectrum. The multiple bandwidth strategies and feature extraction algorithms begin with a foundational understanding of mechanical fault propagation in drive train bearings. Experience with these algorithms and correlation of the indicators within seeded and transitional tests is offered in this paper. The authors discuss specific issues of observability and application of these techniques to the Expeditionary Fighting Vehicle.