| About this Abstract |
| Meeting |
2023 TMS Annual Meeting & Exhibition
|
| Symposium
|
Additive Manufacturing: Beyond on the Beam IV
|
| Presentation Title |
Quantification of Defects in Binder-jet Printed Steel Parts Using Confocal Imaging and Machine Learning |
| Author(s) |
Pooja Maurya, P Chris Pistorius |
| On-Site Speaker (Planned) |
Pooja Maurya |
| Abstract Scope |
Defects like porosity and oxide inclusions greatly influence the quality of binder-jet printed (BJP) parts. Sintering reduces porosity, improving the integrity of BJP parts. Confocal laser scanning microscope (CLSM) assisted real time imaging, with a dew point generator (for controlling sintering atmosphere) greatly helps in understanding evolution of such defects during sintering. In the present work, an in-situ CLSM has been used to image the surface topography of 316L BJP coupons during sintering at 1380℃. Systematic experiments indicated that de-binding at 470℃ is complete within 90 minutes (under argon). The effect of critical parameters like sintering time and dew point / frost point (in an Argon-5%H2 atmosphere) on the evolution of porosity and oxide inclusions has been analyzed. Machine learning is used to process the in-situ CSLM images for quantifying porosity by correlating with ground truth (from optical microscopy). |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Additive Manufacturing, Machine Learning, Characterization |