Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Validation Metrics for Fixed Effects and Mixed-Effects Calibration

Journal Article · · Journal of Verification, Validation and Uncertainty Quantification
DOI:https://doi.org/10.1115/1.4049534· OSTI ID:1769907

The modern scientific process often involves the development of a predictive computational model. To improve its accuracy, a computational model can be calibrated to a set of experimental data. A variety of validation metrics can be used to quantify this process. Some of these metrics have direct physical interpretations and a history of use, while others, especially those for probabilistic data, are more difficult to interpret. In this work, a variety of validation metrics are used to quantify the accuracy of different calibration methods. Frequentist and Bayesian perspectives are used with both fixed effects and mixed-effects statistical models. Through a quantitative comparison of the resulting distributions, the most accurate calibration method can be selected. Two examples are included which compare the results of various validation metrics for different calibration methods. It is quantitatively shown that, in the presence of significant laboratory biases, a fixed effects calibration is significantly less accurate than a mixed-effects calibration. This is because the mixed-effects statistical model better characterizes the underlying parameter distributions than the fixed effects model. The results suggest that validation metrics can be used to select the most accurate calibration model for a particular empirical model with corresponding experimental data.

Research Organization:
Sandia National Laboratories (SNL-NM), Albuquerque, NM (United States)
Sponsoring Organization:
USDOE National Nuclear Security Administration (NNSA)
Grant/Contract Number:
AC04-94AL85000
OSTI ID:
1769907
Report Number(s):
SAND--2020-8540J; 690011
Journal Information:
Journal of Verification, Validation and Uncertainty Quantification, Journal Name: Journal of Verification, Validation and Uncertainty Quantification Journal Issue: 1 Vol. 6; ISSN 2377-2158
Publisher:
ASMECopyright Statement
Country of Publication:
United States
Language:
English

References (14)

On the Statistical Calibration of Physical Models: STATISTICAL CALIBRATION OF PHYSICAL MODELS journal February 2015
DRAM: Efficient adaptive MCMC journal December 2006
Relationship between pressure drop and heat transfer of developing and fully developed flow in smooth horizontal circular tubes in the laminar, transitional, quasi-turbulent and turbulent flow regimes journal February 2018
Pipe flow measurements over a wide range of Reynolds numbers using liquid helium and various gases journal June 2002
Some observations on skin friction and velocity profiles in fully developed pipe and channel flows journal August 1969
Mean-flow scaling of turbulent pipe flow journal October 1998
Friction factor and mean velocity profile for pipe flow at high Reynolds numbers journal September 2015
Monte Carlo sampling methods using Markov chains and their applications journal April 1970
A Simplex Method for Function Minimization journal January 1965
Bayesian calibration of computer models journal August 2001
On Information and Sufficiency journal March 1951
Inference from Iterative Simulation Using Multiple Sequences journal November 1992
An Adaptive Metropolis Algorithm journal April 2001
Calibrating the sqHIMMELI v1.0 wetland methane emission model with hierarchical modeling and adaptive MCMC journal January 2018