Ruoqi (Rosie) Gao

Advisor: Hamid Garmestani

Will defend a doctoral dissertation entitled,

An Integrated Process–Texture–Property–Residual Stress Modeling Framework for Laser Powder Bed Fusion of Ti-6Al-4V, with Experimental Microstructural Characterization

On

Wednesday, July 8th, 2026 at 9:00 AM, EST

At 

Price Gilbert Memorial Library, Room 4222 (Dissertation Defense Room)

and virtually via Zoom:

Zoom Meeting ID: 825 911 3860

Passcode: 662031

Abstract

Laser powder bed fusion (LPBF) enables near-net-shape manufacture of geometrically complex Ti-6Al-4V components, but its steep thermal gradients and rapid, cyclic solidification generate non-equilibrium microstructures, strong crystallographic texture, and large residual stresses that limit dimensional accuracy and structural reliability. Residual-stress models that resolve crystallographic anisotropy are computationally costly, while the fast analytical models suited to surveying process space typically assume isotropic elasticity and isotropic (von Mises) yield — neglecting the elastic and plastic anisotropy that process-induced texture imparts. This dissertation addresses this gap through complementary modeling and experimental studies of LPBF Ti-6Al-4V, unified by the process–structure–property–residual stress chain.

The modeling work assembles a computationally efficient, physics-based forward chain. A semi-analytical thermal model represents the laser as a moving single-ellipsoidal volumetric source, linearizes the temperature-dependent heat equation through the Kirchhoff transformation, and incorporates latent heat to predict the steady-state melt-pool geometry, reproducing width and depth across ten experimental single-track cases to within 3.3% and 2.8% on average. The predicted thermal field drives a two-phase texture model that generates the prior-β solidification texture by competitive columnar growth, applies the inter-layer scan rotation, and transforms to α′ through the twelve Burgers variants, reproducing the characteristic ⟨100⟩∥BD prior-β fiber and the scattered α′ texture against two datasets. The reconstructed textures are homogenized by elastic and visco-plastic self-consistent schemes into effective orthotropic elastic constants and Hill yield-stress ratios, which show the build direction to be both the stiffest (E₃/E₁ ≈ 1.1) and the strongest. Incorporating these texture-derived properties into an analytical residual-stress model predicts that texture does not simply scale residual stress but redistributes it, lowering the transverse and raising the along-scan component.

The experimental work characterizes as-built LPBF Ti-6Al-4V across a deliberate scan-speed and hatch-spacing process window, combining electron backscatter diffraction, porosity analysis, microhardness, and depth-resolved residual-stress measurement on common material. Increasing scan speed monotonically refines the prior-β grain structure and weakens texture, while lath morphology and retained β fraction vary non-monotonically and lack-of-fusion porosity emerges at the highest speed. The dataset documents how these features co-evolve across the window and frames the open questions a definitive microstructure–residual-stress linkage must resolve.

Committee

Prof. Hamid Garmestani - School of Materials Science and Engineering

Prof. Preet Singh - School of Materials Science and Engineering

Prof. Aaron Stebner - School of Materials Science and Engineering

Prof. Steven Y. Liang - School of Mechanical Engineering

Prof. Saïd Ahzi - Faculty of Physics and Engineering, Université de Strasbourg