Bayesian Inference for Multilevel Fault Tree Models

Abstract

This chapter describes how information and data may be available at various levels in a fault tree model, and how these may be used in a Bayesian analysis framework to perform probabilistic inference on the model. For example, we might have information on the overall system performance, but we might also have subsystem and component level information. We demonstrate the analysis approach using a simple fault tree model containing a single top event (a “super-component”) and two sub-events (i.e., piece-parts). Also, we show how OpenBUGS can be used for the example models to estimate the probability of meeting a reliability goal at any level in the fault tree model.