| About this Abstract |
| Meeting |
2020 TMS Annual Meeting & Exhibition
|
| Symposium
|
Advanced Characterization Techniques for Quantifying and Modeling Deformation
|
| Presentation Title |
K-2: Propagating Uncertainty through ICME Modules and Machine Learning towards Quicker and Accurate Distortion and Residual Stress Predictions |
| Author(s) |
Brijesh Kumar, Piyush Ranade, Alonso Peralta, Mustafa Megahed |
| On-Site Speaker (Planned) |
Mustafa Megahed |
| Abstract Scope |
Distortion and residual stress models provide useful information to designers and process engineers to assess manufacturability, geometric compensation and required post treatment. The model inputs coming from micromodels or via experiments vary leading to an uncertainty in distortion predictions. This study demonstrates how uncertainties are propagated from lower scale models to part scale models to assess the final uncertainty in residual stress and strain predictions. The data generated is also processed through machine learning algorithms investigating the potential of using reduced order models for multi-objective process optimization simulations. Experimental results are used to assess the reliability of the procedure and judge improvement needed to the methodology. |
| Proceedings Inclusion? |
Planned: Supplemental Proceedings volume |