System reliability assessment with an approximate reasoning model
The projected service life of weapons in the US nuclear stockpile will exceed the original design life of their critical components. Interim metrics are needed to describe weapon states for use in simulation models of the nuclear weapons complex. The authors present an approach to this problem based upon the theory of approximate reasoning (AR) that allows meaningful assessments to be made in an environment where reliability models are incomplete. AR models are designed to emulate the inference process used by subject matter experts. The emulation is based upon a formal logic structure that relates evidence about components. This evidence is translated using natural language expressions into linguistic variables that describe membership in fuzzy sets. The authors introduce a metric that measures the acceptability of a weapon to nuclear deterrence planners. Implication rule bases are used to draw a series of forward chaining inferences about the acceptability of components, subsystems and individual weapons. They describe each component in the AR model in some detail and illustrate its behavior with a small example. The integration of the acceptability metric into a prototype model to simulate the weapons complex is also described.
- Research Organization:
- Los Alamos National Lab., Technology and Safety Assessment Div., NM (United States)
- Sponsoring Organization:
- USDOE, Washington, DC (United States)
- DOE Contract Number:
- W-7405-ENG-36
- OSTI ID:
- 303958
- Report Number(s):
- LA-UR-98-2453; CONF-980965-; ON: DE99001236; TRN: AHC29904%%31
- Resource Relation:
- Conference: International conference and workshop on reliability and risk management, San Antonio, TX (United States), 15-18 Sep 1998; Other Information: PBD: [1998]
- Country of Publication:
- United States
- Language:
- English
Similar Records
An approximate reasoning-based method for screening high-level-waste tanks for flammable gas
Development of a Severe Accident Analysis Engine Using Approximate Reasoning