Back to search

University of North Texas

Physics-Informed and Data-Driven Frameworks for Thermal and Microstructure Prediction in Additive Manufacturing

Abstract

dc:description

Laser Powder Bed Fusion (LPBF) is an additive manufacturing process used to fabricate complex metallic components with high precision. However, the process involves rapid heating and cooling cycles that generate complex thermal behavior, which directly affects melt pool formation, microstructure evolution, defects, and final material properties. Traditional numerical simulation methods can accurately model these thermal processes, but they are computationally expensive and difficult to use for large-scale process optimization. This dissertation explores the use of machine learning and physics-informed machine learning methods for predicting thermal profiles and microstructural behavior in LPBF additive manufacturing. The research focuses on developing computationally efficient surrogate models that can reduce simulation time while maintaining physically meaningful predictions. Deep learning approaches, including Generative Adversarial Networks (GANs), are used to predict thermal field distributions and microstructural evolution from manufacturing process parameters. In addition, Physics-Informed Neural Networks (PINNs) are developed to incorporate governing heat transfer physics directly into the learning process. The proposed framework is designed to predict transient thermal behavior across varying process conditions using large-scale finite-element simulation data. The results demonstrate that the proposed machine learning models can significantly reduce computational cost while maintaining strong predictive performance for thermal profile prediction. This work highlights the potential of physics-informed machine learning for accelerating process optimization and enabling intelligent data-driven manufacturing systems in additive manufacturing applications.

Degree

thesis:*
Grantor dc:publisher
University of North Texas
Year dc:date
2026

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Manjunath, Aishwarya 1999-
Contributors dc:contributor
  • Pears, Russel
  • Dahotre, Narendra B.
  • Fu, Song
  • Guo, Xuan

Subjects

dc:subject × 3

Rights

dc:rights
Statement dc:rights
  • Public
  • Manjunath, Aishwarya 1999-
  • Copyright
  • Copyright is held by the author, unless otherwise noted. All rights Reserved.
Language dc:language
English

Identifiers

dc:identifier.*
Identifier
local-cont-no: submission_5059
ark: ark:/67531/metadc2652548

Chain of custody

source
Harvested from
University of North Texas
Base URL
digital.library.unt.edu/oai/
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
citation

Manjunath, Aishwarya 1999-. Physics-Informed and Data-Driven Frameworks for Thermal and Microstructure Prediction in Additive Manufacturing. University of North Texas, 2026. https://digital.library.unt.edu/ark:/67531/metadc2652548/