Development of Performance and Effectiveness Metrics For Mechanical Diagnostic Technologies
Abstract
In recent years, numerous anomaly detection and diagnostic technologies have been developed for various military and industrial applications to aid in the detection and classification of developing faults. In many cases, significant reductions in machinery total ownership costs have been achieved through the judicious application of these technologies. However, there is currently no consistent methodology available for assessing both the technical and economic benefits of these machinery diagnostic technologies. In response to this need, a virtual test bench is under development by the Navy for assessing the performance and effectiveness of machinery diagnostic systems. The test bench utilizes a plug n play' interface that can readily accept standardized diagnostic/prognostic tools and link them to real and model-based transitional data from appropriate condition based maintenance (CBM) platforms. The assessment process relies on a standard set of mathematical ground rules and a statistical framework to directly identify confidence bounds, robustness measures, and various diagnostic thresholds associated with specific mechanical diagnostic technologies. Specific performance and accuracy of the diagnostic algorithms at the component or subsystem level are evaluated with performance metrics, while system level capabilities in terms of achieving the overall operational goals of the diagnostic system will be evaluated with effectiveness measures. This qualification and validation methodology is utilized to compare a variety of diagnostic tools that are commonly used to analyze gearbox vibration.
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 05, 2001
- Accession Number
- ADP013509
Entities
People
- Christopher J. Savage
- Kathy Mcclintic
- Rolf F. Orsagh