This paper updates current efforts by the authors to develop fully-automated, online incipient fault detection and prognosis algorithms for drivetrain and engine bearings. The authors have developed and evolved ImpactEnergy/spl trade/, a feature extraction and analysis driven system that integrates high frequency vibration/acoustic emission data, collected using accelerometers and other sensors such as a laser interferometer to assess the health of bearings and gearboxes in turbine engines. ImpactEnergy combines advanced diagnostic features derived from waveform analysis, high-frequency enveloping, and more traditional time domain processing like root mean square (RMS) and kurtosis with classification techniques to provide bearing health information. The adaptable algorithm suite has been applied across numerous air vehicle relevant programs for the Air Force, Navy, Army, and DARPA. The techniques presented in this paper are tested and validated in a laboratory environment by monitoring multiple bearings on test rigs that replicate the operational loads of a turbomachinery environment. The capability of the software on full-scale test rigs at major OEMs (original equipment manufacturer) locations will be shown with specific data results. The team will review developments across these multiple programs and discuss specific implementation efforts to transition to the fleet in a variety of manned and unmanned platforms.
Recent case studies in bearing fault detection and prognosis
2006 IEEE Aerospace Conference ; 8 pp.
2006-01-01
768765 byte
Conference paper
Electronic Resource
English
A Probabilistic Fault Detection Approach: Application to Bearing Fault Detection
Tema Archive | 2011
|Manifold Learning Based Intelligent Fault Diagnosis and Prognosis
Springer Verlag | 2023
|