 |
Center for Nonlinear Studies |
To support effective decision making, engineers and scientists must comprehend and manage various uncertainties throughout analysis and design processes. Unfortunately, in today’s modern systems, uncertainty analysis can become cumbersome and computationally intractable for one individual or group to manage. This is particularly true for systems comprised of a large number of components, subsystems, or disciplines. In many cases, these components may be developed by different groups and even run on different computational platforms. In this talk we propose an approach for synthesizing uncertainty analysis tasks performed independently using the various information-sources used to model the components of a feed-forward system. Our proposed composition-based uncertainty analysis approach is shown to be provably convergent in distribution under certain conditions. The proposed method is illustrated on the quantification of uncertainty for a multidisciplinary gas turbine system and is compared to a traditional system-level Monte Carlo uncertainty analysis approach.
Host: Frank Alexander 5-4518