Augmenting system tests with component tests for reliability assurance

Abstract

To successfully determine if a system meets some reliability standard, typically assurance tests are conducted on the full system. This proven approach is extremely effective. However, it can be cost prohibitive when full system tests are expensive. Previous work has incorporated component and subsystem data to supplement the component prior distributions before a full system assurance test. Although this can be helpful, we propose an additional advance of allowing component and subsystem tests as part of the assurance test. We show that by augmenting full system tests with component and subsystem tests, in some cases, we can reduce the cost of the assurance test. For a system that can be decomposed into independent series and parallel parts, the theory works perfectly. In the case where we cannot accurately classify parts using the series or parallel structure, we add a generalization that works for systems that are “almost-series” or “almost-parallel.”