A new approach to quantification of margins and uncertainties for physical simulation data.

This paper proposes the use of statistical tolerance interval methodology as an approach to quantification of margins and uncertainties (QMU) for physical simulation data. We review the standard -factor methodologies and discuss potential limitations. The tolerance interval methodology is introduced and demonstrated with several examples. A new figure-of-merit is proposed and its properties are explored. These methodologies are intended for a performance characteristic that has shown the potential for low margin or margin that is changing with age. Hence, we require a well-understood dataset that has been through a comprehensive engineering analysis. This paper provides recommendations for an engineering analysis that will result in a dataset that is eligible for a rigorous analysis using these proposed methodologies. Finally, we present an overview of the probability of frequency approach commonly used in computational simulation QMU applications to highlight the similarities with this proposed methodology for physical simulation QMU applications.