11/08/2024
By Danielle Fretwell

The Francis College of Engineering, Department of Plastics Engineering, invites you to attend a Doctoral Dissertation defense by Ahmed Adisa on: Property Prediction in Fused Filament Fabrication.

Candidate Name: Ahmed Adisa
Degree: Doctoral
Defense Date: Friday, November 22, 2024
Time: 10 a.m. - Noon
Location: Ball Hall, Room 302

Committee:
Advisor: Amy Peterson, Ph.D., Professor, Plastics Engineering, University of Massachusetts Lowell

Committee Members*
1. David Kazmer, Ph.D., Professor, Plastics Engineering, University of Massachusetts Lowell
2. Amir Ameli, Ph.D., Associate Professor, Plastics Engineering, University of Massachusetts Lowell
3. Christopher Hansen, Ph.D., Professor, Mechanical Engineering, University of Massachusetts Lowell

Brief Abstract:
Additive manufacturing (AM) offers reduced lead time between design and manufacturing. Fused filament fabrication (FFF), the most common form of material extrusion additive manufacturing, enables the production of custom-made parts with complex geometry. Despite the numerous advantages of AM, reliability, reproducibility, and achievement of isotropic bulk properties in part remains challenging. FFF systems contain several input parameters which lead to varying part properties. These numerous input parameters lead to difficulty in controlling part properties. Maintaining consistent part properties is crucial to the wide-scale adoption of FFF. Furthermore, knowing the tensile property of a part is critical to its safe and optimum performance during use. The ability to predict the tensile properties of printed parts eliminates the need for time-consuming and expensive experiments.

This dissertation focuses on developing predictive models that have short execution time frames and can facilitate real-time property prediction. We used these models to determine the mechanical properties of printed parts with minimal information. In pursuit of this goal, we first investigated the interrelationships between process parameters, cross-sectional geometry, fracture behavior, and mechanical properties in FFF, to have an improved understanding of these complex relationships. We subsequently evaluated the effect of thermal contact resistance between print layers, and its effect on tensile strength predicted using a mathematical (physics-based) modeling approach. Thereafter, we used the experimental data from previously conducted experiments in training machine learning models for predicting mechanical and geometric properties. The study advances the use of simple predictive models in the quest to achieve real-time property prediction in FFF.