Life Extension Analysis & Prognostics (LEAP) Laboratory-Directed Research and Development
A pervasive problem in both government and industry is the need to extend the useful life of systems. Economic pressures to maintain aging fleets of military and commercial equipment and vehicles are very real. Even with relatively new equipment, there is a tremendous cost-benefit of extending the time between overhauls, reducing the probability of a failure in the field, and increasing appropriate preventive repairs. A major predictor of the need for maintenance is the type of use and operating conditions that the product has experienced-such as environmental factors, duty factors, service history, etc. Keys to extending the useful life of each of these "systems" is (1) the capability to record "operational experience," and (2) the capability of integrating and analyzing the recorded data to produce reliable diagnostics and prognostics about the state of the system and its remaining useful life.
While advances in sensor technologies are making it feasible to install sensors on equipment for health monitoring, methods for analyzing system status and predicting system life expectancy need to be made more reliable and robust for data collected onboard systems in real-time. Prognostics is the process of predicting the future state of a system. Prognostics systems comprise sensors, a data acquisition system, and microprocessor-based software to perform sensor fusion, analysis, and reporting/interpreting of results with little or no human intervention in real-time or near real-time. Prognostics offers the promise of minimizing failures (especially failures "in the field"), extending the time between maintenance overhauls, and reducing life-cycle costs. Implementing prognostics is a challenging task on several levels: (1) hardware and sensor technologies, (2) analytically effective predictive methods, and (3) organization changes to capture the logistical benefits made possible by effective prognostic information. Still, the costs of this challenging task are dwarfed by the benefits of effective prognostics.
The goal of this PNNL internal research project was to define the architecture of a dynamic prognostic system and to develop robust, generalized statistical methods for predicting remaining life of mechanical systems.
Greitzer, F.L. May 2002. "Life Extension Analysis and Prognostics (LEAP) Architectures," Laboratory Directed Research and Development, Annual Report 2001, pp. 312-316. PNNL-13855. U.S. Department of Energy. Richland, Washington. (PDF 538KB)
Greitzer, F.L. April 2001. "Life Extension Analysis and Prognostics (LEAP) Architectures," Laboratory Directed Research and Development Annual Report, Fiscal Year 2000, pp. 447-450. PNNL-13501. U.S. Department of Energy. Richland, Washington. (PDF 454KB)
Greitzer, F. L. and TA Ferryman. (2001) Predicting Remaining Life of Mechanical Systems. ASNE Intelligent Ship Symposium IV, April 2-3, 2001, Philadelphia, Pennsylvania. (PDF 552KB)
Greitzer, F.L. April 2000. "Life Extension Analysis and Prognostics Architectures," Laboratory Directed Research and Development Annual Report, Fiscal Year 1999, pp. 85-88. PNNL-13203. U.S. Department of Energy. Richland, Washington. (PDF 51KB)
Greitzer, F. L., E. J. Stahlman, T. A. Ferryman, B. W. Wilson, L. J. Kangas, and D. R. Sisk. Development of a Framework for Predicting Life of Mechanical Systems: Life Extension Analysis and Prognostics (LEAP). SOLE '99 Symposium, August 31- September 2, 1999, Las Vegas, Nevada. (PDF 54KB)
Wilson, B. W., N. H. Hansen, C. L. Shepard, T. J. Peters, and F. L. Greitzer. Development of a Modular In-Situ Oil Analysis Prognostic System. SOLE '99 Symposium, August 31- September 2, 1999, Las Vegas, Nevada. (PDF 22KB)
Engine Removal Projection Tool
Project funded by the U.S. Navy
The U.S. Navy has over 3500 gas turbine engines used throughout the surface fleet for propulsion and the generation of electrical power. The U.S. Navy, Naval Sea Systems Command (NAVSEA) Marine Gas Turbine Information System (MGTIS) Program archives engines' historical operating and removal data for use in logistics and lifecycle planning. One analysis conducted by the MGTIS Program is to project the number of engine removals for the next ten years and determine engine down times between removals. Prior to this research and development effort, NAVSEA used a software tool created in the early 1970s to support this analysis.
A research project conducted by PNNL for the U.S. Navy, Naval Sea Systems Command, Naval Surface Warfare Center developed a new Engine Removal Projection software program that uses a prediction algorithm based on a defensible scientific methodology, implemented with efficient software coding, and that runs under Windows as a Web-based application. To conduct the research underlying the prediction methodology, PNNL tested over 60 techniques on almost 20 years of data collected from over 3100 gas turbine engine assemblies and 120 U.S. Navy ships. We investigated a number of techniques as the forecast basis including moving averages, empirical negative binomial, general linear models, Cox regression, and Kaplan Meier survival curves, most of which are documented in engineering, medical and scientific research literature. We applied those techniques to the data and used test set validation to quantify the accuracy of each method and choose the best algorithm. The selected method demonstrated significant improvement over the existing method and is based on good mathematical techniques, which are described in the paper.
The software uses the best algorithm in combination with user-friendly interfaces and intuitively understandable displays. The user can select a specific engine type, forecast time period, and op-tempo. Graphical displays and numerical tables present forecasts and uncertainty intervals. The technology developed for the project is applicable to other logistic forecasting challenges.
Ferryman, T.A., Matzke, B.D., Wilson, J.E., Sharp, J., Greitzer, F.L., & Hilferty, E. (2005) Engine Removal Projection Tool . Intelligent Ship Symposium VI. Villanova, PA. June 1-2, 2005 PNNL-SA-44900 (PDF 525KB)