| About this Abstract |
| Meeting |
2025 TMS Annual Meeting & Exhibition
|
| Symposium
|
Fatigue in Materials: Fundamentals, Multiscale Characterizations and Computational Modeling
|
| Presentation Title |
Predicting TMF Life of Single-Crystal Ni-Base Superalloys Using a Probabilistic Physics-Guided Neural Network |
| Author(s) |
Richard William Neu, Rohan Acharya, Alexander N. Caputo |
| On-Site Speaker (Planned) |
Richard William Neu |
| Abstract Scope |
Predicting the life under thermomechanical fatigue (TMF) is challenging because there are several parameters defining the mechanical and thermal cycles including dwell periods within the cycle that can impact life. In addition, the anisotropy associated with the loading and crystal orientation significantly influences life. The relationships between these TMF history parameters, crystal orientation, and fatigue life are complex and a general model to capture all these influences is lacking. A model based on a probabilistic physics-guided neural network was developed and trained to learn these relationships using life data extracted from the literature. The model can predict the cycles to failure including uncertainty for a wide range of possible thermomechanical creep-fatigue histories and crystal orientations. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Machine Learning, Other, Other |