1. Deep learning approaches for instantaneous laser absorptance prediction in additive manufacturing.
- Author
-
Jiang, Runbo, Smith, John, Yi, Yu-Tsen, Sun, Tao, Simonds, Brian J., and Rollett, Anthony D.
- Subjects
DEEP learning ,CONVOLUTIONAL neural networks ,ENGINEERING models ,LASERS ,X-ray imaging - Abstract
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 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 vapor depression images and corresponding laser absorptance. In this work, we propose two different approaches to predict instantaneous laser absorptance. The end-to-end approach uses deep convolutional neural networks to learn implicit features of X-ray images automatically 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. While having distinct advantages, both approaches achieved a consistently low mean absolute error of less than 3.3%. [ABSTRACT FROM AUTHOR]
- Published
- 2024
- Full Text
- View/download PDF