| About this Abstract |
| Meeting |
MS&T24: Materials Science & Technology
|
| Symposium
|
Computational Materials for Qualification and Certification
|
| Presentation Title |
Machine Learning Enabled Parametrically Upscaled Constitutive Models for Fatigue Simulations: A Data-Driven Multiscale Modeling Approach |
| Author(s) |
Somnath Ghosh |
| On-Site Speaker (Planned) |
Somnath Ghosh |
| Abstract Scope |
This talk will give an overview of the development of the Parametrically Upscaled Constitutive Model (PUCM) and the Parametrically Upscaled Crack Nucleation Model (PUCNM) for fatigue modeling of metallic materials like Ti alloys. These thermodynamically consistent constitutive models bridge multiple spatial scales through the explicit representation of representative aggregated microstructural parameters (RAMPs), representing statistical distributions of morphological and crystallographic descriptors of the microstructure. They enable computationally efficient simulations with significant speedup over detailed lower-scale models. A host of computational tools and machine learning (ML) algorithms are developed to create an automated pipeline for parametric upscaling. The novel algorithms used include genetic programming symbolic regression (GPSR) for functional representation. The computational tool chain outputs the highly efficient PUCM/PUCNM, which are invaluable tools for multiscale analysis of deformation and fatigue failure with implications in location-specific design. |