Bayesian approach to alloy simulation, A

Show simplified metadata
dc.contributor.advisorToberer, Eric
dc.contributor.advisorStevanovic, Vladan
dc.contributor.authorNovick, Andrew
dc.contributor.committeememberMaughan, Annalise
dc.contributor.committeememberGómez-Gualdrón, Diego A.
dc.contributor.committeememberPak, Alexander J.
dc.date.accessioned2025-04-25T20:14:10Z
dc.date.available2025-04-25T20:14:10Z
dc.date.issued2024
dc.date.updated2025-04-14T16:27:49Z
dc.descriptionIncludes bibliographical references.
dc.description2024 Fall.
dc.description.abstractAlloy simulation is—seemingly—rife with intractable mathematical problems: the combinatorial explosion of atomic decorations, the irreducible global nature of convex hulls, and the curse of dimensionality in configuration space. Due to the importance of alloys across materials science, considerable attention has been given to surmounting these computational obstacles. The typical angle is to tackle daunting mathematical problems with increasingly complex model-Hamiltonians, ranging from lattice models to machine-learned interatomic potentials. Even in the fortunate cases where these models provide accurate predictions, extracting insight from large-scale simulation is challenging, limiting our theoretical understanding. Herein, I take a Bayesian approach to alloy simulation. All predictions provided from calculations are thus represented as probabilistic distributions over possible outcomes rather than isolated, singular values. The classic materials science questions are used to ground this thesis: i) can a given chemical composition be made as a single-phase alloy; and ii) what will be its local and long-range atomic structure? Equipped with Bayesian modeling, I show that relatively few calculations are necessary to provide sufficient estimations for the stated questions. As such, advanced alloy simulation is returned to the domain of first-principles calculations, enabling accuracy and functionality. While I focus solely on first-principles simulation, model-Hamiltonian research also stands to benefit from the developed Bayesian approach, which could accelerate exploration across time scales, length scales, and composition spaces. If there is an overarching thesis to this amalgamation of work, it is the following. Bloated mathematical constructs conceal the underlying simplicity of alloys—through the success of relatively simple simulation, I highlight the elegance within these disordered materials.
dc.format.mediumborn digital
dc.format.mediumdoctoral dissertations
dc.identifierNovick_mines_0052E_12994.pdf
dc.identifier.otherT 9891
dc.identifier.urihttps://hdl.handle.net/11124/180392
dc.languageEnglish
dc.language.isoeng
dc.publisherColorado School of Mines. Arthur Lakes Library
dc.relation.ispartof2024 - Mines Theses & Dissertations
dc.rightsCopyright of the original work is retained by the author.
dc.subjectalloys
dc.subjectBayesian
dc.subjecthigh-entropy
dc.subjectstatistics
dc.titleBayesian approach to alloy simulation, A
dc.typeText
dspace.entity.typePublication
thesis.degree.disciplinePhysics
thesis.degree.grantorColorado School of Mines
thesis.degree.levelDoctoral
thesis.degree.nameDoctor of Philosophy (Ph.D.)
Files
Original bundle
Now showing 1 - 2 of 2
Thumbnail Image
Name:
Novick_mines_0052E_12994.pdf
Size:
20.39 MB
Format:
Adobe Portable Document Format
Download
Thumbnail Image
Name:
Novick_mines_0052E_316/thesis permissions.pdf
Size:
1.71 MB
Format:
Adobe Portable Document Format
Download
License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
1.33 KB
Format:
Item-specific license agreed upon to submission
Description:
Download
Collections
2024 - Mines Theses & Dissertations