Date of Award

2025-08-01

Degree Name

Doctor of Philosophy

Department

Mechanical Engineering

Advisor(s)

Yirong Lin

Second Advisor

Bill Tseng

Abstract

This dissertation presents an in-depth look at the mechanical behavior, size changes, and process improvement of ceramic and polymer materials made with additive manufacturing (AM). The work consists of four separate but related studies, each tackling important challenges in fabrication, characterization, and predictive modeling. Chapter 1 examines the nano-mechanical property changes of 3D-printed zirconia made using digital light processing (DLP). By employing nanoindentation and scanning electron microscopy (SEM), the effects of print orientation and sintering on hardness, elastic modulus, and microstructure are revealed. The results indicate that 0°-oriented sintered samples have up to 140% higher hardness compared to preconditioned samples, with mechanical properties comparable to those of conventionally processed ceramics. Chapter 2 introduces a Python-based image processing method to observe size changes during sintering. Optical images are analyzed through edge detection and spline fitting to monitor shrinkage and deformation. This non-contact technique supports mechanical testing by allowing real-time observation of shape changes. Chapter 3 details the creation of a high-load barium titanate (BaTiO3) slurry for DLP-based production of piezoelectric ceramics. The study assesses photoblocker concentrations, cure depth, and printability. A 90 wt% BaTiO3 mixture was successfully printed, reaching a target cure depth of 0.050 mm, showing that it is possible to produce dense lead-free ceramics using DLP. Chapter 4 investigates machine vision and machine learning to predict the mechanical behavior of 3D-printed silicone lattice structures. Cross-sectional images of the printed lattices are analyzed to identify geometric features. These features are then used to train predictive models, such as convolutional neural networks and genetic algorithms. This section emphasizes the potential of merging image processing with data-driven modeling for effective property prediction. These studies together improve our understanding of additive manufacturing processes and offer new tools for characterizing, optimizing, and predicting the performance of engineered materials.

Language

en

Provenance

Received from ProQuest

File Size

67 p.

File Format

application/pdf

Rights Holder

Diana Hazel Leyva Marquez

Download

Share

COinS