Prognostic Analysis of the Tactical Quiet Generator

The U.S. Army needs prognostic analysis of mission-critical equipment to enable condition-based maintenance before failure. ORNL has developed and patented prognostic technology that quantifies condition change from noisy, multi-channel, time-serial data. This report describes an initial application of ORNL's prognostic technology to the Army's Tactical Quiet Generator (TQG), which is designed to operate continuously at 10 kW. Less-than-full power operation causes unburned fuel to accumulate on internal components, thereby degrading operation and eventually leading to failure. The first objective of this work was identification of easily-acquired, process-indicative data. Two types of appropriate data were identified, namely output-electrical current and voltage, plus tri-axial acceleration (vibration). The second objective of this work was data quality analysis to avoid the garbage-in-garbage-out syndrome. Quality analysis identified more than 10% of the current data as having consecutive values that are constant, or that saturate at an extreme value. Consequently, the electrical data were not analyzed further. The third objective was condition-change analysis to indicate operational stress under non-ideal operation and machine degradation in proportion to the operational stress. Application of ORNL's novel phase-space dissimilarity measures to the vibration power quantified the rising operational stress in direct proportion to the less-than-full-load power. We conclude that ORNL'smore » technology is an excellent candidate to meet the U.S. Army's need for equipment prognostication.« less

[1]  Lee M. Hively,et al.  Epileptic Event Forewarning From Scalp EEG , 2001, Journal of clinical neurophysiology : official publication of the American Electroencephalographic Society.

[2]  Jb Andriulli Development of Proof-of-Concept Units for the Advanced Medium-Sized Mobile Power Sources (AMMPS) Program , 2002 .

[3]  V. Protopopescu,et al.  Detecting dynamical change in nonlinear time series , 1999 .

[4]  Lm Hively,et al.  Epilepsy Forewarning Using a Hand- Held Device , 2005 .

[5]  Irene A. Stegun,et al.  Handbook of Mathematical Functions. , 1966 .

[6]  C. E. SHANNON,et al.  A mathematical theory of communication , 1948, MOCO.

[7]  Henry D. I. Abarbanel,et al.  Analysis of Observed Chaotic Data , 1995 .

[8]  Leon M. Tolbert,et al.  Advanced power generation systems for the 21st Century: Market survey and recommendations for a design philosophy , 1999 .

[9]  Xiao Li,et al.  Study and performance evaluation of some nonlinear diagnostic methods for large rotating machinery , 1993 .

[10]  V. Protopopescu,et al.  Timely detection of dynamical change in scalp EEG signals. , 2000, Chaos.

[11]  Lee M. Hively Epileptic Seizure Forewarning by Nonlinear Techniques , 2001 .

[12]  Lee M. Hively,et al.  Channel-consistent forewarning of epileptic events from scalp EEG , 2003, IEEE Transactions on Biomedical Engineering.

[13]  Lee M. Hively,et al.  Data-driven nonlinear technique for condition monitoring , 1997 .

[14]  L M Hively,et al.  Machine failure forewarning via phase-space dissimilarity measures. , 2004, Chaos.

[15]  Lee M. Hively,et al.  Sensitive Measures of Condition Change in EEG Data , 1999 .

[16]  Lm Hively,et al.  Annual Report for NERI Proposal No.2000-0109 on Forewarning of Failure in Critical Equipment at Next-Generation Nuclear Power Plants , 2001 .

[17]  Lee M. Hively,et al.  Nonlinear crack growth monitoring , 2001 .

[18]  Lee M. Hively,et al.  Advanced Physiological Monitoring of FCS Soldiers , 2004 .