Information Fusion for Improved Prognostics and Health Management of Military Vehicles
Abstract
The development of Condition Based Maintenance (CBM) for wheeled land vehicles has been an interest to both the Army and Marine Corps for over ten years. Previous initiatives have shown that, while steady state sensor information can assist with diagnostics, prognostics will require sensors that more directly measure the physical phenomenon associated with the root cause of degradation and failure of mechanical components. The aviation industry has a successful history of CBM that includes vibration analysis and fluid (oil) analysis, both of which attempt to measure the effects of these root cause failure mechanisms. Vibration signature analysis and fluid analysis can provide complementary information regarding the health of key components in wheeled land vehicle drive trains. Mechanical defects and faults in rotating equipment produce characteristic vibration signatures at specific frequencies related to the design and operation of the overall system; these vibration signatures can be evaluated and trended through joint time--?frequency analysis to first detect and diagnose faults and then to prognose the remaining useful life of the system. Fluid analysis enables detection of wear through monitoring particulates and debris in the oil, often giving an earlier indication of impending faults and early degradation than vibration analysis alone. However, fluid analysis is not sensitive to all the mechanical faults that can be detected through vibration analysis. By combining the two approaches into an integrated health monitoring system, faults of various types should be detectable, diagnosable, and prognosable. A proof--?of--?principle demonstration of the integration of vibration and fluid analysis for monitoring drive train components under seeded faults and accelerated aging is proposed. The ultimate goal of this research is a demonstration that the results of vibration and fluid analysis can be combined through a hybrid analysis, along with state variable indicators, to provide early warning of incipient faults and accurate prognosis of failures throughout the component life. In this proof--?of--?principle demonstration, a testbed will be developed and constructed to run a single wheeled land vehicle component to failure with and without seeded faults. An automotive differential, which has an fluid reservoir that is required for fluid analysis and various bearings and gears for seeded fault introduction, is one acceptable subsystem for testing. The automotive differential includes several internal components for testing seeded faults and accelerated degradation: [1] pinion bearings (tapered roller), [2] carrier bearings (tapered roller), [3] spider gear bearings (sleeve type), [4] ring and pinion gear set (helical cut gears), [5] spider gear set (straight cut gears), and [6] axle bearings (spherical roller). An objective of this testing is to experimentally assess the correlation of measureable parameters to degradation mechanisms by running some of the internal differential components (bearings and gears) to failure in an accelerated manner. In order to maximize the value of vibration and fluid analyses for health management and maintenance planning, both need to be performed online in near--?real time as the vehicle operates. The proposed proof of principle experiments will demonstrate the applicability of these analyses for health management and will develop a framework and models for combining information from these analyses for accurate, high--?confidence prognostics. Development of smart vibration sensors and online fluid analysis sensors is beyond the scope of this work but will ultimately be driven by the results of these tests to meet required accuracy, uncertainty, and processing speed requirements.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jul 07, 2017
- Source ID
- N000141712441
Entities
People
- Jamie Coble
Organizations
- Office of Naval Research
- United States Navy
- University of Tennessee