Uncertainty Quantification in Atomistic Modeling of Metals and Its Effect on Mesoscale and Continuum Modeling: A Review
- PDF / 607,094 Bytes
- 15 Pages / 593.972 x 792 pts Page_size
- 9 Downloads / 204 Views
https://doi.org/10.1007/s11837-020-04436-6 Ó 2020 This is a U.S. government work and not under copyright protection in the U.S.; foreign copyright protection may apply
AUGMENTING PHYSICS-BASED MODELS IN ICME WITH MACHINE LEARNING AND UNCERTAINTY QUANTIFICATION
Uncertainty Quantification in Atomistic Modeling of Metals and Its Effect on Mesoscale and Continuum Modeling: A Review JOSHUA J. GABRIEL ,1,7 NOAH H. PAULSON,1 THIEN C. DUONG,2 FRANCESCA TAVAZZA,3 CHANDLER A. BECKER ,4 SANTANU CHAUDHURI,5,6 and MARIUS STAN 1 1.—Applied Materials Division, Argonne National Laboratory, Lemont, IL 60439, USA. 2.—Energy and Global Security, Argonne National Laboratory, Lemont, IL 60439, USA. 3.—Material Measurement Laboratory, National Institute of Standards and Technology, Gaithersburg, MD 20899, USA. 4.—Office of Data and Informatics, Material Measurement Laboratory, National Institute of Standards and Technology, Gaithersburg, MD 20899, USA. 5.—Manufacturing Science and Engineering, Energy and Global Security, Argonne National Laboratory, Lemont, IL 60439, USA. 6.—Civil, Materials, and Environmental Engineering, University of Illinois at Chicago, Chicago, IL 60607, USA. 7.—e-mail: [email protected]
The design of next-generation alloys through the integrated computational materials engineering (ICME) approach relies on multiscale computer simulations to provide thermodynamic properties when experiments are difficult to conduct. Atomistic methods such as density functional theory (DFT) and molecular dynamics (MD) have been successful in predicting properties of never before studied compounds or phases. However, uncertainty quantification (UQ) of DFT and MD results is rarely reported due to computational and UQ methodology challenges. Over the past decade, studies that mitigate this gap have emerged. These advances are reviewed in the context of thermodynamic modeling and information exchange with mesoscale methods such as the phase-field method (PFM) and calculation of phase diagrams (CALPHAD). The importance of UQ is illustrated using properties of metals, with aluminum as an example, and highlighting deterministic, frequentist, and Bayesian methodologies. Challenges facing routine uncertainty quantification and an outlook on addressing them are also presented.
INTRODUCTION Integrated computational materials engineering (ICME) describes the design of materials for target properties by the coupled use of experiments, computational simulations, and data-driven techniques. Atomistic simulation workflows that cross multiple time and length scales are becoming popular for the determination of physical properties critical to ICME. One often overlooked tenet of ICME, however, is the reliable quantification of uncertainties of material properties. This is especially important for the design of metals that are used in transportation, structural, health, and energy industries due to the mission-critical nature of the materials performance and the potential for loss of life should failures (Received July 27, 2020; accepted October 5,
Data Loading...
