| About this Abstract |
| Meeting |
2025 TMS Annual Meeting & Exhibition
|
| Symposium
|
Additive Manufacturing Modeling, Simulation and Machine Learning
|
| Presentation Title |
A Deep Learning Framework for Predicting Surface Deformation of Alloys Under Uniaxial Tensile Loading at Microscopic Length Scale |
| Author(s) |
Kavindu Wijesinghe, Steven Arnold, Ajit Achuthan |
| On-Site Speaker (Planned) |
Kavindu Wijesinghe |
| Abstract Scope |
Understanding mechanical behavior at both macroscopic and microscopic scales is crucial for analyzing structural performance and designing advanced alloys. Traditional in-situ tensile testing, while effective in capturing microstructure-influenced deformation, is time-consuming and requires high expertise. To address this, we introduce a Deep Learning (DL) assisted virtual in-situ tensile testing system that predicts deformation behavior from microstructures. This adaptable framework enables rapid and accurate characterization of new material. Our system uses a high-throughput in-situ tensile testing device for capturing the training data. To reduce data complexity and minimize the need for extensive training data, we use dynamic mode decomposition, retaining crucial information for accurate predictions. The system utilizes comprehensive input data on undeformed microstructures and potential deformation mechanisms to predict the deformation and evolution of grain and sub-grain features. This approach advances materials science research by providing a robust method for understanding microstructural behavior under tensile conditions, expediting next-generation alloy development. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Machine Learning, Characterization, Other |