Title :
Predictive Bayes design of Scram systems: related mathematics and philosophical implications
Author :
Clarotti, Carlo A. ; Spizzichino, Fabio
Author_Institution :
ENEA, Rome, Italy
fDate :
9/1/1996 12:00:00 AM
Abstract :
The predictive Bayes approach to reliability provides a rational method for scoring different design solutions for systems which can fail. This statement is supported by solving the problem of optimal design of Scram systems (emergency systems) formed from constant-failure-rate components, and called into operation by plant-transients which occur according to a Poisson process. The lessons from this exercise are: in the predictive Bayes frame, any reliability problem is a sequential statistical decision problem; and design by mean-cost minimization (the use of Bayes approach, and in particular its predictive version) requires extensive numerical analysis
Keywords :
Bayes methods; decision theory; failure analysis; minimisation; redundancy; reliability theory; safety systems; stochastic processes; Poisson process; Scram systems; constant-failure-rate components; dangerous failure; decision problem; emergency systems; mathematics; mean-cost minimization; numerical analysis; optimal design; philosophical implications; plant-transients; predictive Bayes design; redundancy optimisation; reliability; renewal equation; safe failure; sequential statistical decision problem; Accidents; Calendars; Cost function; Equations; Mathematics; Numerical analysis; Production systems; Redundancy; State-space methods; Terminology;
Journal_Title :
Reliability, IEEE Transactions on
