About this Abstract |
Meeting |
2022 Annual International Solid Freeform Fabrication Symposium (SFF Symp 2022)
|
Symposium
|
2022 Annual International Solid Freeform Fabrication Symposium (SFF Symp 2022)
|
Presentation Title |
Predicting Geometric Accuracy via Image-to-image Machine Learning |
Author(s) |
Daphne Lin, Tim Phillips, Jared Allison, Carolyn Seepersad |
On-Site Speaker (Planned) |
Daphne Lin |
Abstract Scope |
As AM technologies become more widespread, gaps in process control also become more apparent. While previous research primarily focuses on the effects of adjusting process parameters, there is a lack of process control tools for monitoring and ensuring accuracy in part geometry during a print. Our goal is to demonstrate the possible use of image-to-image machine learning algorithms to monitor and update process parameters in polymer selective laser sintering (SLS). We first train the image-to-image conditional generative adversarial network (cGAN) on a dataset of layer-by-layer thermal images and physical dimensions of the final part. Aftr the cGAN model is trained, it can accurately predict the physical dimensions of a layer based on input thermal images. The ultimate goal of the research is to integrate the model into the controls of the polymer SLS machine and update process parameters in real time to ensure accurate prints with little manual parameter adjustment. |
Proceedings Inclusion? |
Definite: Post-meeting proceedings |