Bayesian inference with overlapping data for systems with continuous life metrics

A Bayesian approach for generating inference from multiple overlapping higher level system data sets on component reliability parameters within systems with continuous life metrics (as distinct from on-demand systems) is presented in this paper. Overlapping data sets are those that are drawn simultaneously from the same process or system. The methodology proposed in this paper is exclusively based on time, but is easily transferrable to any other variable (such as distance, flow etc.). The approach is able to incorporate overlapping evidence from systems with continuous life metrics using a detailed understanding of the system logic represented using fault-trees, reliability block diagrams or equivalent representation. The reliability parameters of each component define the continuous reliability function associated with each sensor location. This paper offers a fully Bayesian method of analyzing multiple overlapping higher level data sets for complex systems with multiple instances of identical components. The scope of the paper is limited to binary-state systems and components that exist in either ‘failed’ or ‘successful’ states.