Activity Number:
|
160
- Quantifying Uncertainty
|
Type:
|
Topic Contributed
|
Date/Time:
|
Monday, July 30, 2018 : 10:30 AM to 12:20 PM
|
Sponsor:
|
Uncertainty Quantification for Complex Systems Interest Group
|
Abstract #328452
|
Presentation
|
Title:
|
A Geometric Approach for Calibrating Computer Models with Misaligned Functional Output
|
Author(s):
|
Lauren Hund* and James Derek Tucker and Justin Brown
|
Companies:
|
and Sandia National Laboratories and Sandia National Laboratories
|
Keywords:
|
functional data analysis;
model calibration;
uncertainty quantification;
computer models
|
Abstract:
|
Bayesian model calibration is a statistical framework for solving an inverse problem to estimate parameters input into a computational simulation model. Model outputs are coupled with experimental measurements to learn about the model inputs. When the outputs of the computer model are functions over time, functional misalignment must be incorporated into the calibration procedure to accurately estimate and quantify uncertainty in calibrated parameters. We propose a geometric approach to Bayesian model calibration of misaligned functions based on elastic shape analysis. Posterior inference on the calibration parameters incorporates both phase and amplitude divergence from the experimental data. We describe theoretical advantages of the elastic shape framework for functional calibration relative to alternative functional registration techniques and apply our approach to estimate and quantify uncertainty in dynamic material properties under extreme pressures.
|
Authors who are presenting talks have a * after their name.