About this Abstract |
| Meeting |
2026 TMS Annual Meeting & Exhibition
|
| Symposium
|
2026 Technical Division Student Poster Contest
|
| Presentation Title |
SPU-41: Development of Long-Term Corrosion Predictions for Anoxic Steel Corrosion: Collection of Experimental Boundary Conditions for Physics-Informed AI/ML Models |
| Author(s) |
Selena Ganues, Matthew Roop, Ryan M. Katona, Rebecca F. Schaller |
| On-Site Speaker (Planned) |
Selena Ganues |
| Abstract Scope |
Physics-informed Finite Element Method (FEM) modeling using experimental parameters can predict materials' corrosive behavior. However, there is a need to accelerate predictions and extend them to long timeframes (i.e., > 10,000 years). Incorporating artificial intelligence/machine learning (AI/ML) techniques with an active learning framework into FEM models can achieve this. Long-term corrosion predictions are particularly necessary for carbon steels used in geologic repository storage for radioactive waste, but relevant measurements are lacking. Potentiodynamic data of carbon steels (ASTM A1008) in various concentrations and pH levels of representative repository brines under anoxic conditions (< 5 ppm oxygen) and atmospheric oxygen were collected to establish boundary conditions. Modeling this data will investigate how changes in environmental parameters affect the corrosion activity of carbon steel.
Acknowledgements: SNL is managed and operated by NTESS under DOE NNSA contract DE-NA0003525. This document is SAND2026-16026A. |
| Proceedings Inclusion? |
Undecided |
| Keywords |
Electrometallurgy, Machine Learning, Modeling and Simulation |