Modeling failure priors and persistence in model-based diagnosis

Probabilistic model based diagnosis com putes the posterior probabilities of failure of components from the prior probabilities of component failure and observations of sys tem behavior One problemwith this method is that such priors are almost never directly available One of the reasons is that the prior probability estimates include an implicit no tion of a time interval over which they are speci ed for example if the probability of failure of a component is is this over the period of a day or is this over a week A sec ond problem facing probabilistic model based diagnosis is the modeling of persistence Say we have an observation about a system at time t and then another observation at a later time t To compute posterior probabil ities that take into account both the observa tions we need some model of how the state of the system changes from time t to t In this paper we address these problems using tech niques from Reliability theory We show how to compute the failure prior of a component from an empirical measure of its reliability the Mean Time Between Failure MTBF We also develop a scheme to model persis tence when handlingmultiple time tagged ob servations