| About this Abstract |
| Meeting |
2023 Annual International Solid Freeform Fabrication Symposium (SFF Symp 2023)
|
| Symposium
|
2023 Annual International Solid Freeform Fabrication Symposium (SFF Symp 2023)
|
| Presentation Title |
Deep Learning Approaches for Time-resolved Laser Absorptance Prediction |
| Author(s) |
Runbo Jiang, Brian Simonds, Anthony Rollett |
| On-Site Speaker (Planned) |
Runbo Jiang |
| Abstract Scope |
The quantification of absorbed light is essential for understanding laser-material interactions and melt pool dynamics in order to minimize defects in additively manufactured metal components. The geometry of a vapor depression in melt pools formed during laser melting is closely related to laser energy absorption. This relationship has been observed by the state-of-the-art in situ high speed synchrotron x-ray visualization and integrating sphere radiometry. These two techniques create a temporally resolved dataset consisting of keyhole images and the corresponding laser absorptance. In this work, we propose two different approaches to predict laser absorptance. The end-to-end approach uses convolutional neural networks to learn features of unprocessed x-ray images automatically without human supervision and predict the laser energy absorptance. The two-stage approach uses a semantic segmentation model to engineer geometric features and predict absorptance using classical regression models. Though with different advantages, both approaches reached a smooth L1 loss less than 6%. |
| Proceedings Inclusion? |
Definite: Post-meeting proceedings |