About this Abstract |
| Meeting |
2026 TMS Annual Meeting & Exhibition
|
| Symposium
|
Additive Manufacturing Modeling, Simulation and Artificial Intelligence
|
| Presentation Title |
Machine Learning Guided Prediction of Electrohydrodynamic Jet Printing by Incorporating High Speed Video Data |
| Author(s) |
Stanford White, Prashant Ghimire, Yiwei Han, Samrat Choudhury |
| On-Site Speaker (Planned) |
Samrat Choudhury |
| Abstract Scope |
This work introduces a data-driven framework for modeling process-structure-property relationships during electrohydrodynamic (EHD) jet printing using high-speed photography and machine learning. High-speed imaging at 15,000 fps captured dynamic jetting behaviors, enabling quantitative analysis of droplet velocity, cone angle, and printing mode. A Gaussian Process classifier mapped observed printing regimes with 92% accuracy, and machine learning was utilized on a large dataset of functional materials to develop processing maps for new materials. A separate Gaussian Process regressor predicted droplet breakage velocities, showing micro-dripping regimes exhibited higher velocities due to extended elongation, with an R2 = 0.955. Cone angle was quantified using digital image correlation. A transformer-based deep learning model accurately simulated temporal droplet behavior, achieving accurate prediction based of the meniscus contour perimeter. This integrated approach provides a scalable surrogate for real-time control and optimization in nozzle-based AM systems. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Additive Manufacturing, Machine Learning, Electronic Materials |