A fully subjective approach to modelling inspection maintenance

The fully subjective, or fully Bayesian approach discussed in this paper provides straightforward means of presenting uncertainty related to future events to decision makers. This approach is described in the context of an inspection maintenance decision problem, and is contrasted with the classical probabilistic approach that assumes the existence of true probabilities and probability distributions which have to be estimated. Key features of the fully subjective approach are discussed, including integration of engineering judgements, uncertainty treatment and type of performance measures to be used. An example related to the problem of identifying "optimal" inspection intervals for an extrusion press is used to illustrate the principles described. (C) 2002 Elsevier Science B.V. All rights reserved.