|
Activity Number:
|
363
|
|
Type:
|
Invited
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Physical and Engineering Sciences
|
| Abstract - #305346 |
|
Title:
|
Uncertainty Quantification for Combining Experimental Data and Computer Simulations from Multiple Data Sources
|
|
Author(s):
|
Brian J. Williams*+ and Dave Higdon and Jim Gattiker
|
|
Companies:
|
Los Alamos National Laboratory and Los Alamos National Laboratory and Los Alamos National Laboratory
|
|
Address:
|
P.O. Box 1663, Los Alamos, NM, 87545-0001,
|
|
Keywords:
|
calibration ; computer experiments ; Gaussian process ; functional data analysis ; uncertainty quantification ; predictive science
|
|
Abstract:
|
This work focuses on combining observations from field experiments with detailed computer simulations of a physical process to carry out inference. This typically involves calibration of parameters in the simulator and accounting for inadequate physics. We consider physical applications for which the field data and simulator output are multivariate. Multivariate data lead to computational challenges for implementing the framework. We consider adaptive basis methods to achieve significant dimension reduction. This methodology is extended to incorporate multiple sources of field data and simulator output into a joint calibration and prediction analysis. Different sources of data inform on specific calibration parameter subsets, which are not required to be disjoint. We illustrate the proposed methodology with experimental data and simulations that inform on gas equations of state.
|
