From FMECA to Decision: A Fully Bayesian Reliability Process

Reliability testing of complex military systems is often hindered by the difficulty or cost of the testing. Therefore, conventional methodologies that rely on repeated measurements under differing conditions cannot be used. However, these systems are often designed using subcomponents that do have testing data available. In this manuscript, we demonstrate how Bayesian methodologies can be used throughout the testing process to combine subcomponent reliability score, under various levels of certainty, in order to inform, and maximize the impact of, a small number of full system tests. We show how this can be used for the U.S. Army’s long-range precision fires lines of effort and conclude that Bayesian inference should be incorporated into future experimental design for these complex systems.