| About this Abstract |
| Meeting |
2025 TMS Annual Meeting & Exhibition
|
| Symposium
|
Materials Aging and Compatibility: Experimental and Computational Approaches to Enable Lifetime Predictions
|
| Presentation Title |
Accelerating Computational Calculations of Galvanic Corrosion using Machine Learning |
| Author(s) |
David Montes De Oca Zapiain, Aditya Venkatraman, Matthew Roop, Demitri Maestas, Michael Melia, Philip Noell, Ryan Katona |
| On-Site Speaker (Planned) |
David Montes De Oca Zapiain |
| Abstract Scope |
Galvanic corrosion deteriorates the performance of metallic alloys exposed to different atmospheric conditions. Accurate predictions of corrosion are critical to determine the effect these conditions have on the degradation rate of the alloy. Existing computational frameworks based on the finite element method are ill-suited, despite their accuracy, for identifying factors that foment, or deter, corrosion given their iterative nature. Machine learning (ML) models are a viable solution given their ability to provide accurate predictions at a fraction of the cost. Nevertheles, ML-based models need to be trained on adequately calibrated simulations and struggle in provding accurate predictions in extrapolation. This work addresses these challenges by introducing an automated and data-driven framework capable of identifying and incorporating experimental measurements and their corresponding computational results that will yield the maximum information gain and best enhance performance of the ML model.
SNL is managed and operated by NTESS under DOE-NNSA contract DE-NA0003525. SANDNo.SAND2024-07981A |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Machine Learning, Mechanical Properties, Modeling and Simulation |