| About this Abstract |
| Meeting |
2024 TMS Annual Meeting & Exhibition
|
| Symposium
|
Additive Manufacturing: Advanced Characterization with Synchrotron, Neutron, and In Situ Laboratory-scale Techniques III
|
| Presentation Title |
Improving Porosity Characterization and Analysis in Additive Manufacturing Through 3D X-ray Microscopy Coupled with Deep Learning-based Resolution Recovery |
| Author(s) |
Nathan Johnson, Hrishi Bale, Yulia Trenikhina, Stephen Kelly |
| On-Site Speaker (Planned) |
Hrishi Bale |
| Abstract Scope |
The ultimate performance of the parts produced by metal powder-bed fusion additive manufacturing (AM) processes is dominated by porosity and flaws that emerge in the production process making it is critical to study their effects on the physical and mechanical properties of the produced parts. Laboratory X-ray microscopes enable access to higher resolution 3D imaging through the use of two-stage objective-based magnification approaching sub-micron resolutions. However, the age-old resolution vs field-of-view challenge makes it harder to fully characterize sufficiently large representative volumes in AM, especially in parts with widely distributed pore sizes ranging from nm-to-sub mm and in-homogenously distributed over a large volume. Latest advances in deep-learning based 3D image reconstruction and imaging optics enable mapping porosity at high resolution and over large area using the concept of resolution recovery. We present results demonstrating improvements in pore analysis by combining software and hardware solutions for characterizing complex AM architectures. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Characterization, Additive Manufacturing, Machine Learning |