| About this Abstract |
| Meeting |
2024 TMS Annual Meeting & Exhibition
|
| Symposium
|
Accelerated Discovery and Insertion of Next Generation Structural Materials
|
| Presentation Title |
Physics-informed Creep Rupture Life Modeling of High Temperature Alloys for Energy Applications |
| Author(s) |
Madison Wenzlick, William Trehern, Anderson Soares Chinen, Anjana Talapatra, Michael Gao |
| On-Site Speaker (Planned) |
Madison Wenzlick |
| Abstract Scope |
Long-term mechanical properties are key benchmarks for designing new materials for advanced applications. However, properties such as creep rupture time are difficult to model empirically due to changing creep mechanisms over time, vast design spaces, and data scarcity from time-consuming and costly experimental testing. Integrated computational materials engineering (ICME) is therefore a promising approach for predicting creep behavior. Using a data-driven, physics-informed machine learning strategy, the model interpretability, physical relevancy, and extrapolative accuracy can be enhanced. In this work, physical features are incorporated into a predictive machine learning model based on experimental data to evaluate the creep life performance of 9-12% Cr and austenitic stainless-steel alloys. An active learning and optimization framework is deployed to efficiently guide experimentation for validation of selected candidate alloy predictions. Comparison is made against a state-of-the-art physics-based elasto-viscoplastic fast Fourier transform (EVP-FFT) constitutive creep model to evaluate the relative strengths of both techniques. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
High-Temperature Materials, Mechanical Properties, Machine Learning |