About this Abstract |
| Meeting |
2025 Annual International Solid Freeform Fabrication Symposium (SFF Symp 2025)
|
| Symposium
|
2025 Annual International Solid Freeform Fabrication Symposium (SFF Symp 2025)
|
| Presentation Title |
Machine Learning-Based Prediction of Toughness in Fused Filament Fabrication: Leveraging In-Process Annealing With Enhanced Printheads |
| Author(s) |
Rakin Ahmed, Tanvir Ahmed Shanto, Md Mahmudur Rahman, Robert M Taylor, Ankur Jain |
| On-Site Speaker (Planned) |
Rakin Ahmed |
| Abstract Scope |
Fused Filament Fabrication (FFF) is widely adopted for its affordability and design flexibility, but printed Acrylonitrile Butadiene Styrene (ABS) parts often exhibit poor Z-direction strength due to weak interlayer adhesion and thermal stresses. To address this, an enhanced printhead enabling in situ annealing was employed to promote improved bonding through controlled thermal management. This study leverages machine learning (ML) to predict mechanical toughness from thermal features captured during printing and cooling. Forty-four thermal descriptors were analyzed using correlation and Principal Component Analysis, retaining 99.8% variance. Multiple classifiers, including k-Nearest Neighbors, Decision Trees, Naïve Bayes, Support Vector Machines, Artificial Neural Networks, XGBoost, and CatBoost, were evaluated. CatBoost achieved the highest balanced accuracy (83.33%) and Area Under the Curve (76.19%) for toughness classification. Regression models offered limited continuous predictions. The integration of thermal monitoring with ML shows strong potential for non-destructive, real-time quality assurance in enhancing thermoplastic additive manufacturing reliability. |
| Proceedings Inclusion? |
Planned: Post-meeting proceedings |