About this Abstract |
| Meeting |
2026 TMS Annual Meeting & Exhibition
|
| Symposium
|
2026 Technical Division Student Poster Contest
|
| Presentation Title |
SPG-91: Microstructure-Sensitive Fatigue Crack Nucleation and Short Crack Growth Modeling Using Parametrically Upscaled Models |
| Author(s) |
Kishore Appunhi Nair, Tawqeer Tak, Somnath Ghosh |
| On-Site Speaker (Planned) |
Kishore Appunhi Nair |
| Abstract Scope |
Fatigue failure of metals is a multiscale phenomenon where the microstructure strongly influences the crack nucleation and short crack growth stage. However, existing multiscale models face a trade-off, compromising either computational efficiency or mechanistic fidelity. To address this, a thermodynamically consistent multi-scale modeling framework called the Parametrically Upscaled Crack Nucleation/Continuum Damage Mechanics (PUCNM/PUCDM) model is presented. Key to this framework is the concept of Representative Aggregate Microstructural Parameters (RAMPs), which encode microstructure into macroscale constitutive equations. Machine learning is used to extract the functional relationship between model parameters and RAMPs from data generated by high-fidelity micromechanical simulations using a crystal plasticity model. The resulting macroscale model links microstructure directly to fatigue response and enables large-scale simulations of engineering components, where microstructural sensitivity and computational efficiency are critical. The proposed model is used to investigate the competing effects of extrinsic and intrinsic factors on fatigue behavior. |
| Proceedings Inclusion? |
Undecided |
| Keywords |
Modeling and Simulation, Computational Materials Science & Engineering, Machine Learning |