Plant monitoring using possibilistic-probabilistic methodologies
In the process of monitoring the performance of nuclear power plant systems with the objective of fault identification and diagnosis, uncertainty is introduced. This uncertainty must be accounted for in rule-based systems, which are used for this purpose. In this context, uncertainty consists of two parts: randomness, meaning the uncertainty of occurrence of an object, and fuzziness, expressing the imprecision of the meaning of the object. For instance, for an engineering device, let us consider the proposition: E will increase {approximately}20% with a probability of 90%. The variable E could be any performance parameter of interest. The uncertainty in this statement has two components: one related to randomness, which is expressed by a probability of 90%, and another component due to fuzziness, which is expressed by {approximately}20%. From fuzzy set theory, the meaning of {approximately}20% can be precisely defined by a possibility distribution {sub {pi}X}, which may be induced by physical constraints, or it may be epistemic in nature, in which case it is induced by a collection of propositions. From probability theory, the construction of a probability density function based on frequency distributions from experimental and sensor data and observations can precisely define the meaning of the words a probability of 90%. Discussed is a methodology for inferencing and decision making under uncertainty for use in rule-based systems, where the knowledge about the modeled system is both probabilistic and possibilistic, and apply it to the monitoring and diagnosis of the performance of a nuclear reactor's residual heat removal (RHR) system.
- OSTI ID:
- 5228657
- Report Number(s):
- CONF-871101-; CODEN: TANSA; TRN: 90-002769
- Journal Information:
- Transactions of the American Nuclear Society; (USA), Vol. 55:3; Conference: Joint meeting of the American Nuclear Society and the Atomic Industrial Forum, Los Angeles, CA (USA), 15-19 Nov 1987; ISSN 0003-018X
- Country of Publication:
- United States
- Language:
- English
Similar Records
Anticipatory systems using a probabilistic-possibilistic formalism
Estimating Terrorist Risk with Possibility Theory
Related Subjects
NUCLEAR POWER PLANTS
REACTOR MONITORING SYSTEMS
RHR SYSTEMS
FAULT TREE ANALYSIS
HIGH PRESSURE COOLANT INJECTION
LEAKS
LOW PRESSURE COOLANT INJECTION
PERFORMANCE
PROBABILITY
PUMPS
COOLING SYSTEMS
ECCS
ENERGY SYSTEMS
ENGINEERED SAFETY SYSTEMS
NUCLEAR FACILITIES
POWER PLANTS
REACTOR COMPONENTS
REACTOR COOLING SYSTEMS
REACTOR PROTECTION SYSTEMS
SYSTEM FAILURE ANALYSIS
SYSTEMS ANALYSIS
THERMAL POWER PLANTS
220200* - Nuclear Reactor Technology- Components & Accessories