| About this Abstract |
| Meeting |
2025 TMS Annual Meeting & Exhibition
|
| Symposium
|
Novel Strategies for Rapid Acquisition and Processing of Large Datasets from Advanced Characterization Techniques
|
| Presentation Title |
Efficient SEM Imaging Strategies for Microstructure Analysis in Metal Additive Manufacturing |
| Author(s) |
Christina Koenig, Andrei Tudor Durnescu, Sotero Romero, Laura Andrea Paz Salas, Joerg Jinschek |
| On-Site Speaker (Planned) |
Christina Koenig |
| Abstract Scope |
Plasma Arc Additive Manufacturing (PLAAM) results in complex metal microstructures due to rapid temperature changes during the manufacturing process, significantly affecting material properties. This study aims to link PLAAM parameters, microstructure, and final material properties. Here, a strategy for high throughput examining large-scale samples is developed using exclusively scanning electron microscopy (SEM) imaging techniques.
Our approach focuses on achieving high statistical relevance through global sample analysis while reducing SEM imaging times. We use image analysis algorithms for defect analysis to correlate these to PLAAM conditions as well as a deep learning segmentation model to analyze grain characteristics, using precise Electron Backscatter Diffraction (EBSD)-derived labels for grain segmentation.
We systematically evaluate the effect of SEM imaging dwell time on the methods accuracy, providing insights into optimizing parameters for desired detection rates. Initial findings highlight significant reductions in imaging time, demonstrating the capacity to adjust imaging parameters to meet specific accuracy requirements. |
| Proceedings Inclusion? |
Planned: |