Georgia Institute of Technology
Microstructure evolution and the influence on material properties in additive manufacturing
Abstract
dc:description.abstractDespite the transformative potential of additive manufacturing (AM) as a sustainable, inverse-philosophy alternative to traditional subtractive methods, crucial for global decarbonization, its widespread adoption faces fundamental challenges. Metal AM processes are inherently complex, governed by multi-physical phenomena (thermal, fluid, phase transformations) and exhibiting material-specific behaviors that vary significantly across alloy systems. This complexity hinders the reliable prediction and control of microstructural evolution (grain size, texture, defects), final mechanical properties (elastic modulus, Poisson's ratio, yield strength, residual stress), and defect formation, limiting its ability to replace conventional manufacturing. This research addresses these core challenges by establishing a quantitative understanding of the entire metal AM process chain. The primary aim is to elucidate the evolution of critical microstructural features and their deterministic impact on macroscopic mechanical performance, thereby bridging the persistent gap between microstructural evolution and macro-scale properties. To achieve this, we developed a comprehensive suite of physics-based closed-form analytical models, rigorously validated against experimental data, encompassing: (1) precise thermal distribution modeling accounting for moving heat sources and boundary conditions; (2) microstructure evolution models predicting grain growth and crystallographic texture formation; (3) mechanical property prediction models for elastic modulus, Poisson's ratio, yield strength, and residual stress; and (4) defect formation models. Targeting the critical industry challenge of processing parameter optimization, where traditional trial-and-error is a prohibitively expensive and high-fidelity numerical simulation is often computationally intractable, this work pioneers a transformative hybrid methodology. We strategically integrate these foundational physics-based analytical models with a genetic algorithm (GA) optimization engine. This framework synergistically combines the physical models with GA-driven adaptive sampling to achieve orders-of-magnitude reduction in computational cost, significant minimization of data acquisition requirements, and accelerated convergence efficiency. Importantly, the framework provides powerful bidirectional prediction capabilities: enabling highly efficient forward prediction of microstructure and properties from given process parameters, and intelligent inverse optimization of manufacturing conditions to meet target performance specifications. This research crystallizes a scientific framework for causality-aware manufacturing. By deriving closed-form models from first principles, it deciphers the ontological basis of metal AM process-structure-property (PSP) relationships, delivering not only real-time predictive capability but also a mechanistic epistemology for process cognition. The integration of evolutionary optimization elevates this into a self-actualizing scientific paradigm: physical laws provide causal grounding while genetic algorithms enable autonomous hypothesis refinement. This physics-encoded computational paradigm transcends empirical approaches, establishing provably generalizable principles for PSP control in cross-material applications, a new archetype of scientific inquiry merging reductionism with complex systems optimization, and a replicable conjectural-refutation template for closed-loop knowledge generation in industrial science.
Degree
thesis:*- Level thesis:degree_level
- Doctoral
- Department dc:contributor.department
- Mechanical Engineering
- Grantor dc:publisher
- Georgia Institute of Technology
- Year dc:date.issued
- 2025
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Huang, Wei
- Advisor dc:contributor.advisor
-
- Liang, Steven Y.
- Committee members dc:contributor.committeemember
-
- Garmestani, Hamid
- Kurfess, Thomas R.
- Melkote, Shreyes N.
- Shi, Jianjun
Subjects
dc:subject × 4Rights
- Language dc:language.iso
- en_US
Identifiers
dc:identifier.*- Handle dc:identifier.uri
- https://hdl.handle.net/1853/82065