| About this Abstract |
| Meeting |
2025 TMS Annual Meeting & Exhibition
|
| Symposium
|
Additive Manufacturing Modeling, Simulation and Machine Learning
|
| Presentation Title |
Machine Learning Enabled Process Optimization during 3D Printing of
Tablets |
| Author(s) |
Yizhou Lu, Kshitij Chitnis, Jaidev Chakka, Samrat Choudhury, Mo Maniruzzaman |
| On-Site Speaker (Planned) |
Yizhou Lu |
| Abstract Scope |
Fused deposition modeling (FDM), a 3D printing approach is capable of producing high quality
drug products. However, surface defects such as porosities are observed during FDM-based
printing of tablets. Optimizing the processing parameters for defect free tablets can be non-
trivial, time consuming, and costly process as the combinatorial processing condition search
space is simply too large. We used an adaptive design strategy, a machine learning algorithm to
determine the processing conditions needed to optimize the porosity on the tablet surface. As a
first step commercial Polylactic-acid filament was utilized to print formulation under different
processing conditions using batch and continuous printers. Image segmentation was performed to
determine porosity on the surface of tablets. Later, machine learning tools such as random
forest was used to identify key processing parameters that impact the porosity on the tablet
surface. Finally, machine learning predicted processing conditions needed for optimal porosity
were verified experimentally. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Additive Manufacturing, Machine Learning, |