Reliability Modeling using Both System Test and Quality Assurance Data

Abstract

A system with several components may undergo full system pass/fail testing as well as quality assurance testing at the single component level. The component tests are informative about system reliability, although they measure different things than the full system tests. We present a Bayesian framework for integrating the two types of test data for better reliability estimates. Our formulation allows the reliability to depend on covariates such as age. One result of the inference is a better understanding of the relationship between component tests and system performance. We illustrate the ideas using a small subsystem of a larger (proprietary) system.