About this Abstract |
| Meeting |
2026 TMS Annual Meeting & Exhibition
|
| Symposium
|
Solid-State Processing and Manufacturing for Nuclear Applications: Integrating Insights and Innovations
|
| Presentation Title |
Machine Learning Driven Optimization of 3D-Printed Advanced Materials for Radiation Shielding |
| Author(s) |
Johannes Huurman, Kunal Mondal, Oscar Martinez, Rigoberto Advincula |
| On-Site Speaker (Planned) |
Kunal Mondal |
| Abstract Scope |
Innovations in advanced manufacturing technology present significant opportunities for enhancing safety and efficiency in the nuclear industry. Functionally graded materials (FGMs) in particular offer a new paradigm for developing radiation shielding with an optimal balance of effectiveness and weight. We present a computational study designing additively manufacturable polymer composites to attenuate 1.25 MeV gamma photons. The concentration of dopants was varied across the materials' thickness according to distinct gradient profiles. Our analysis confirms that tungsten is the most effective shielding additive. While gradient profiles that concentrate the dopant, such as a parabolic function, yield the highest linear attenuation coefficient (μ), they also yield the heaviest material. By analyzing the trade-off between attenuation and mass per unit area, we demonstrate that tungsten-doped FGMs with sigmoid and step-function gradients provide a superior balance of properties compared to uniformly doped materials.
|
| Proceedings Inclusion? |
Planned: |
| Keywords |
Additive Manufacturing, Nuclear Materials, Machine Learning |