About this Abstract |
Meeting |
2025 TMS Annual Meeting & Exhibition
|
Symposium
|
Additive Manufacturing Modeling, Simulation and Machine Learning
|
Presentation Title |
Enabling Part-Scale Microstructure Modeling in Powder Bed Fusion |
Author(s) |
Michael Paleos, Shawn Hinnebusch, Albert To |
On-Site Speaker (Planned) |
Michael Paleos |
Abstract Scope |
Cellular automata (CA) models have been successful in approximating the true physics of melt pool solidification in powder bed fusion processes, but they are typically confined to relatively small spatial domains. This work focuses on scaling CA simulations to part-scale while achieving satisfactory fidelities through several computational schemes. On the thermal side, we use both layer- and scan-resolved simulations based on the matrix-free, finite element method-based PAMSIM process simulator to estimate the melt pool temperature history. A deep learning model then refines the layerwise-simulated data predicting the necessary scan-resolved thermal signatures, acting as a surrogate for the scan-resolved model. Following the time-efficient capture of the thermal signature data, the as-built microstructure can be predicted by leveraging the open-source ExaCA software in a decoupled manner. Ultimately, by circumventing the computational cost of thermal process simulations, this framework enables part-scale simulations and, thus, facilitates process optimization. |
Proceedings Inclusion? |
Planned: |
Keywords |
Additive Manufacturing, ICME, Machine Learning |