skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Annual Report Nucelar Energy Research and Development Program Nuclear Energy Research Initiative

Abstract

NERI Project No.2000-0109 began in August 2000 and has three tasks. The first project year addressed Task 1, namely development of nonlinear prognostication for critical equipment in nuclear power facilities. That work is described in the first year's annual report (ORNLTM-2001/195). The current (second) project year (FY02) addresses Task 2, while the third project year will address Tasks 2-3. This report describes the work for the second project year, spanning August 2001 through August 2002, including status of the tasks, issues and concerns, cost performance, and status summary of tasks. The objective of the second project year's work is a compelling demonstration of the nonlinear prognostication algorithm using much more data. The guidance from Dr. Madeline Feltus (DOE/NE-20) is that it would be preferable to show forewarning of failure for different kinds of nuclear-grade equipment, as opposed to many different failure modes from one piece of equipment. Long-term monitoring of operational utility equipment is possible in principle, but is not practically feasible for the following reason. Time and funding constraints for this project do not allow us to monitor the many machines (thousands) that will be necessary to obtain even a few failure sequences, due to low failure rates (<10{supmore » -3}/year) in the operational environment. Moreover, the ONLY way to guarantee a controlled failure sequence is to seed progressively larger faults in the equipment or to overload the equipment for accelerated tests. Both of these approaches are infeasible for operational utility machinery, but are straight-forward in a test environment. Our subcontractor has provided such test sequences. Thus, we have revised Tasks 2.1-2.4 to analyze archival test data from such tests. The second phase of our work involves validation of the nonlinear prognostication over the second and third years of the proposed work. Recognizing the inherent limitations outlined in the previous paragraph, Dr. Feltus urged Oak Ridge National Laboratory (ORNL) to contact other researchers for additional data from other test equipment. Consequently, we have revised the work plan for Tasks 2.1-2.2, with corresponding changes to the work plan as shown in the Status Summary of NERI Tasks. The revised tasks are as follows: Task 2.1--ORNL will obtain test data from a subcontractor and other researchers for various test equipment. This task includes development of a test plan or a description of the historical testing, as appropriate: test facility, equipment to be tested, choice of failure mode(s), testing protocol, data acquisition equipment, and resulting data from the test sequence. ORNL will analyze this data for quality, and subsequently via the nonlinear paradigm for prognostication. Task 2.2--ORNL will evaluate the prognostication capability of the nonlinear paradigm. The comparison metrics for reliability of the predictions will include the true positives, true negatives, and the forewarning times. Task 2.3--ORNL will improve the nonlinear paradigm as appropriate, in accord with the results of Tasks 2.1-2.2, to maximize the rate of true positive and true negative indications of failure. Maximal forewarning time is also highly desirable. Task 2.4--ORNL will develop advanced algorithms for the phase-space distribution function (PS-DF) pattern change recognition, based on the results of Task 2.3. This implementation will provide a capability for automated prognostication, as part of the maintenance decision-making. Appendix A provides a detailed description of the analysis methods, which include conventional statistics, traditional nonlinear measures, and ORNL's patented nonlinear PSDM. The body of this report focuses on results of this analysis.« less

Authors:
Publication Date:
Research Org.:
Oak Ridge National Lab. (ORNL), Oak Ridge, TN (United States)
Sponsoring Org.:
USDOE
OSTI Identifier:
885729
Report Number(s):
ORNL/TM-2002/183
TRN: US200617%%156
DOE Contract Number:  
DE-AC05-00OR22725
Resource Type:
Technical Report
Country of Publication:
United States
Language:
English
Subject:
29 ENERGY PLANNING, POLICY AND ECONOMY; 42 ENGINEERING; ALGORITHMS; CONTRACTORS; DATA ACQUISITION; DECISION MAKING; DISTRIBUTION FUNCTIONS; IMPLEMENTATION; MACHINERY; MAINTENANCE; MONITORING; MONITORS; NUCLEAR ENERGY; NUCLEAR POWER; PHASE SPACE; SEEDS; STATISTICS; TESTING; VALIDATION

Citation Formats

Hively, LM. Annual Report Nucelar Energy Research and Development Program Nuclear Energy Research Initiative. United States: N. p., 2003. Web. doi:10.2172/885729.
Hively, LM. Annual Report Nucelar Energy Research and Development Program Nuclear Energy Research Initiative. United States. https://doi.org/10.2172/885729
Hively, LM. 2003. "Annual Report Nucelar Energy Research and Development Program Nuclear Energy Research Initiative". United States. https://doi.org/10.2172/885729. https://www.osti.gov/servlets/purl/885729.
@article{osti_885729,
title = {Annual Report Nucelar Energy Research and Development Program Nuclear Energy Research Initiative},
author = {Hively, LM},
abstractNote = {NERI Project No.2000-0109 began in August 2000 and has three tasks. The first project year addressed Task 1, namely development of nonlinear prognostication for critical equipment in nuclear power facilities. That work is described in the first year's annual report (ORNLTM-2001/195). The current (second) project year (FY02) addresses Task 2, while the third project year will address Tasks 2-3. This report describes the work for the second project year, spanning August 2001 through August 2002, including status of the tasks, issues and concerns, cost performance, and status summary of tasks. The objective of the second project year's work is a compelling demonstration of the nonlinear prognostication algorithm using much more data. The guidance from Dr. Madeline Feltus (DOE/NE-20) is that it would be preferable to show forewarning of failure for different kinds of nuclear-grade equipment, as opposed to many different failure modes from one piece of equipment. Long-term monitoring of operational utility equipment is possible in principle, but is not practically feasible for the following reason. Time and funding constraints for this project do not allow us to monitor the many machines (thousands) that will be necessary to obtain even a few failure sequences, due to low failure rates (<10{sup -3}/year) in the operational environment. Moreover, the ONLY way to guarantee a controlled failure sequence is to seed progressively larger faults in the equipment or to overload the equipment for accelerated tests. Both of these approaches are infeasible for operational utility machinery, but are straight-forward in a test environment. Our subcontractor has provided such test sequences. Thus, we have revised Tasks 2.1-2.4 to analyze archival test data from such tests. The second phase of our work involves validation of the nonlinear prognostication over the second and third years of the proposed work. Recognizing the inherent limitations outlined in the previous paragraph, Dr. Feltus urged Oak Ridge National Laboratory (ORNL) to contact other researchers for additional data from other test equipment. Consequently, we have revised the work plan for Tasks 2.1-2.2, with corresponding changes to the work plan as shown in the Status Summary of NERI Tasks. The revised tasks are as follows: Task 2.1--ORNL will obtain test data from a subcontractor and other researchers for various test equipment. This task includes development of a test plan or a description of the historical testing, as appropriate: test facility, equipment to be tested, choice of failure mode(s), testing protocol, data acquisition equipment, and resulting data from the test sequence. ORNL will analyze this data for quality, and subsequently via the nonlinear paradigm for prognostication. Task 2.2--ORNL will evaluate the prognostication capability of the nonlinear paradigm. The comparison metrics for reliability of the predictions will include the true positives, true negatives, and the forewarning times. Task 2.3--ORNL will improve the nonlinear paradigm as appropriate, in accord with the results of Tasks 2.1-2.2, to maximize the rate of true positive and true negative indications of failure. Maximal forewarning time is also highly desirable. Task 2.4--ORNL will develop advanced algorithms for the phase-space distribution function (PS-DF) pattern change recognition, based on the results of Task 2.3. This implementation will provide a capability for automated prognostication, as part of the maintenance decision-making. Appendix A provides a detailed description of the analysis methods, which include conventional statistics, traditional nonlinear measures, and ORNL's patented nonlinear PSDM. The body of this report focuses on results of this analysis.},
doi = {10.2172/885729},
url = {https://www.osti.gov/biblio/885729}, journal = {},
number = ,
volume = ,
place = {United States},
year = {Thu Feb 13 00:00:00 EST 2003},
month = {Thu Feb 13 00:00:00 EST 2003}
}