About this Abstract |
| Meeting |
MS&T26: Materials Science & Technology
|
| Symposium
|
Additive Manufacturing Modeling, Simulation, and Machine Learning: Microstructure, Mechanics, and Process
|
| Presentation Title |
Efficient Melt-Pool Modeling Informed by Computational Fluid Dynamics for Part Scale Porosity and Microstructure Prediction |
| Author(s) |
Evan Adcock, Joshua Pribe, George Weber |
| On-Site Speaker (Planned) |
Evan Adcock |
| Abstract Scope |
In additive manufacturing (AM) processes like powder bed fusion, modeling accurate melt-pool behavior is the first step toward reliable process-structure-property (PSP) predictions for AM parts. Simple analytical models for the melt-pool such as the Rosenthal solution do not account for the physics of solidification and lack variability in melt-pool dimensions. However, more accurate models like computational fluid dynamics (CFD) are computationally expensive and thus impractical for full part-scale simulations. This work presents a middle-ground approach utilizing a Gaussian-filter based analytical thermal model built in the open-source, PSP modeling Python package, Materialite, and calibrated against the commercial CFD software FLOW-3D AM. This presentation shows melt-pool geometry and resulting porosity predictions results from Materialite for three sets of processing parameters ranging from nominal process window to lack-of-fusion regimes in nickel alloy 718. Porosity predictions are compared against characterization data to support melt-pool model validation and improve PSP simulation capabilities. |