PhD Oral Exam - Allyson Fontes, Mechanical Engineering

When studying for a doctoral degree (PhD), candidates submit a thesis that provides a critical review of the current state of knowledge of the thesis subject as well as the student’s own contributions to the subject. The distinguishing criterion of doctoral graduate research is a significant and original contribution to knowledge.

Once accepted, the candidate presents the thesis orally. This oral exam is open to the public.

Abstract

Data-driven Thermal Modeling and Interlaminar Bond Improvement for in-situ Automated Fiber Placement (AFP) of Thermoplastic Composites Current manufacturing processes for thermoplastic fiber-reinforced polymer composites are slow and rely on manual interventions and post-consolidation in an autoclave. In-situ manufacturing of thermoplastic composites using Automated Fiber Placement (AFP) offers the potential to eliminate the secondary thermal process, thereby reducing manufacturing time and costs. However, the short processing times in in-situ AFP result in poorer part quality than with conventional methods (e.g., autoclave, compression molding). For high-performance industries to adopt in-situ AFP, the interlaminar bond strength, void content, and crystallinity of AFP-produced parts must improve significantly. Despite extensive research within the AFP community, achieving autoclave-level bond quality remains challenging due to difficulties in understanding interactions among process parameters, especially in multilayered samples. This thesis focuses on enhancing in-situ AFP of thermoplastics by addressing two key areas: thermal history prediction and bond-strength improvement. A unique 3D dataset of experimental temperatures across various heat input parameters has been collected, providing insights into parameter interactions. Two data-driven thermal models have been developed to demonstrate the application of machine learning for process control. The first model was a data-driven neural network, whereas the second was a theory-guided neural network that encoded domain knowledge to improve generalization performance. Analyzing this temperature data offered a deeper understanding of critical parameters and informed a new reconsolidation treatment process. To achieve autoclave-level bonding, a novel in-situ treatment process was proposed, and a new device, the in-situ consolidator device (ICD), was devised. The novel in-situ treatment process was tested and validated in both static and dynamic modes. In both modes, bond strength was measured through Single Lap Shear (SLS) tests. Results from the in-situ treatment were compared with those from autoclave-consolidated, in-situ AFP-consolidated, and in-situ repassed samples. The proposed in-situ treatment achieved near-autoclave-level bond strength in both operational modes. Overall, this novel process provides a promising pathway toward high-throughput, fully automated, repeatable in-situ AFP manufacturing.