About this Abstract |
| Meeting |
2026 TMS Annual Meeting & Exhibition
|
| Symposium
|
Forming and Joining of Advanced Sheet Metal Materials
|
| Presentation Title |
Multiscale modeling of incremental sheet forming using machine learning of crystal plasticity |
| Author(s) |
John S. Weeks, Aaron Stebner |
| On-Site Speaker (Planned) |
John S. Weeks |
| Abstract Scope |
Multiscale modeling of metal forming processes enables optimal design of engineering components with local material properties and microstructures, however, conventional concurrent techniques are typically too expensive for design iteration. In this work, we develop a multiscale workflow using low-cost recurrent neural network surrogate models for crystal plasticity that capture both constitutive response and texture evolution during an incremental sheet metal forming process with Al5052. This work couples multiple surrogate models for plastic and elasto-plastic behavior and prediction of texture evolution which is then demonstrated on truncated-pyramid and cone forming paths. We show the elasto-plastic model shows a decreased computational cost with good accuracy for full-field texture predictions and macroscopic parameters such as forming force. Performance across multiple scales establishes this approach as efficient multiscale workflow for incremental deformation processes which can facilitate design iteration considering both macro- and microscale behavior. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
ICME, Modeling and Simulation, Shaping and Forming |