A Bayesian calibration approach to the thermal problem

✍ Scribed by Dave Higdon; Charles Nakhleh; James Gattiker; Brian Williams

• Publisher: Elsevier Science
• Year: 2008
• Tongue: English
• Weight: 246 KB
• Volume: 197
• Category: Article
• ISSN: 0045-7825
• DOI: 10.1016/j.cma.2007.05.031

✦ Synopsis

Many of the problems we work with at Los Alamos National Laboratory are similar to the thermal problem described in the tasking document. In this paper, we describe the tools and methods we have developed that utilize experimental data and detailed physics simulations for uncertainty quantification, and apply them to the thermal challenge problem. We then go on to address the regulatory question posed in the problem description. This statistical framework used here is largely based on the approach of Kennedy and O'Hagan [Kennedy, M., O'Hagan, A., Bayesian calibration of computer models (with discussion), J. Royal Stat. Soc. B 68 (2001) 425-464], but has been extended to deal with functional output of the simulation model.