Data-driven foundations of microstructure and plastic localization relationships in additively manufactured alloys

Additively manufactured (AM) alloys develop complex, non-equilibrium microstructures that show cellular dislocation networks, low- and high-angle boundary (LAGB/HAGB) topologies, and chemical segregation. These profoundly reshape how plasticity develops and localizes, and in turn macroscopic properties. This dissertation establishes data-driven foundations linking AM microstructure to plastic localization and macroscopic properties, and works to translate those links into predictive tools. First, correlative high-resolution digital image correlation (HR-DIC) with EBSD/BSE/TEM demonstrates that, in AM 316L stainless steel, plastic localization characteristics are governed primarily by intragranular heterogeneities, cell structure and local misorientation, rather than classical descriptors used for wrought alloys such as grain size or Schmid factor. Second, a computer-vision pipeline generalizes rapid, statistical extraction of plasticity characteristics across alloys, and enabling orders-of-magnitude scale-up of statistical investigation of plasticity. It results in extending our fundamental understanding of AM microstructure effects on a large set of structural alloys. Third, leveraging an established quantitative relationship between plasticity characteristics and macroscopic properties, here fatigue strength, this work demonstrates fatigue strength prediction in AM alloys from the developed rapid plasticity characteristics evaluation. Finally, a macroscopic properties predictive model is developed that leverage both plasticity and microstructure: It consists of (i) supervised models that map microstructural descriptors to localization metrics, and (ii) deep generative encoders that transform raw diffraction patterns into spatially faithful latent maps capturing AM microstructure. Together these contributions yield an initial end-to-end framework, from microstructural characterization to plasticity to property prediction, that provides mechanistic insights and practical, rapid prediction of fatigue strength directly from AM microstructures, with implications for alloy qualification, process optimization, and closed-loop materials design.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Materials Science & Engr

Grantor

University of Illinois Urbana-Champaign

Year dc:date

2025

Author dc:creator

• Bean, Christopher Matthew

Contributors dc:contributor

• Stinville, Jean-Charles
• Bellon, Pascal
• Schleife, Andre
• Ertekin, Elif

• Additive manufacturing
• Metallurgy
• Digital Image Correlation
• Machine Learning
• Data-driven
• Plasticity
• Electron Backscatter Diffraction
• Encoding

Statement dc:rights

• 2025 Christopher Matthew Bean

Language dc:language

en

Handle dc:identifier

https://hdl.handle.net/2142/132494

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/132494