Quantifying Reliability Uncertainty from Catastrophic and Margin Defects: A Proof of Concept

• Activity Number: 334
• Type: Topic Contributed
• Date/Time: Tuesday, August 2, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Section on Physical and Engineering Sciences
• Author(s): John Lorio*+
• Companies: Sandia National Laboratories
• Keywords: Method of Moments ; Bayesian Analysis ; Bootstrap Analysis ; System Reliability ; Catastrophic Failure Modes ; Margin Failure Modes

Abstract

We aim to analyze the use of component level reliability data, including both catastrophic failures and margin failures, to estimate system level reliability and uncertainty. In this paper, a catastrophic failure is the failure of a component to produce any output and a margin failure is the failure of a component's output to meet a functional requirement. While much work has been done to analyze margins and uncertainties at the component level, a gap exists in relating this component level analysis to the system level. We apply methodologies for aggregating uncertainty from component level data to quantify overall system uncertainty. We explore three approaches towards this goal, the Classical Method of Moments, Bayesian, and Bootstrap methods. These three approaches are used to quantify the uncertainty in reliability for a system of mixed series and parallel components for which both discrete (pass/fail) and continuous margin data are available. This paper provides proof of concept that uncertainty quantification methods can be constructed and applied to system reliability problems. We also show that the three fundamentally different approaches give comparable results.