About this Abstract |
| Meeting |
MS&T25: Materials Science & Technology
|
| Symposium
|
Understanding and Mitigating High Temperature Corrosion Processes Through Synergistic Integration of Experimental, Computational and Manufacturing Techniques
|
| Presentation Title |
Discovering mappings between thermodynamic states in metal oxidation using machine learning |
| Author(s) |
Nathan Bianco, Scott Monismith, Remi Dingreville |
| On-Site Speaker (Planned) |
Nathan Bianco |
| Abstract Scope |
Zirconium alloys are widely utilized in nuclear reactors for structural components and fuel cladding, primarily due to their exceptional corrosion resistance under operating conditions and capacity to capture neutrons. However, the oxidation equivalence of zirconium alloys compared to other alloys remains largely unexplored. Developing a framework for equivalence in corrosion behavior is essential for extending the lifespan of materials and enhancing flexibility in material selection. This study presents a novel approach that employs machine learning to identify equivalences between zirconium alloys and various surrogate alloys. By integrating diverse data sources, including CALPHAD simulations and phase-field modeling, machine learning techniques are employed to generate low-dimensional embeddings that capture the underlying characteristics of different alloys. These embeddings are analyzed to uncover equivalences in the oxidation behavior of metal oxides, providing valuable insights for material optimization and selection in nuclear applications. SNL is managed and operated by NTESS under DOE NNSA contract DE-NA0003525. |