About this Abstract |
| Meeting |
2025 Annual International Solid Freeform Fabrication Symposium (SFF Symp 2025)
|
| Symposium
|
2025 Annual International Solid Freeform Fabrication Symposium (SFF Symp 2025)
|
| Presentation Title |
Born Qualified Additive Manufacturing - In-Situ Prediction of Microstructure in
Laser Powder Bed Fusion Using Physics and Data-Driven Machine Learning |
| Author(s) |
Kaustubh Deshmukh, Prahalada K. Rao |
| On-Site Speaker (Planned) |
Kaustubh Deshmukh |
| Abstract Scope |
This work concerns the prediction of flaw formation and microstructure in Inconel 718 parts made using LPBF. Microstructure features, such as primary dendritic arm spacing, grain size, shape, morphology, and orientation, are key determinants of functional mechanical properties. To predict the aforementioned multi-scale quality metrics, we developed and implemented a physics and sensor data integrated machine learning approach. In this work, we manufactured parts of varying geometry and processing parameters on an SLM 280 system. Real-time process data was acquired via a heterogeneous sensor suite consisting of a long-wave infrared camera, optical tomography and powder bed imaging. This sensor data was used along with thermal history predictions from a computational model as inputs to a machine learning model trained to predict flaw formation and microstructure. The approach predicts solidified microstructure with accuracy exceeding 95%. This work takes the first step toward in-situ qualification of LPBF part quality. |
| Proceedings Inclusion? |
Planned: Post-meeting proceedings |