| About this Abstract |
| Meeting |
2024 Annual International Solid Freeform Fabrication Symposium (SFF Symp 2024)
|
| Symposium
|
2024 Annual International Solid Freeform Fabrication Symposium (SFF Symp 2024)
|
| Presentation Title |
Toward Parametric Heat Transfer Solvers in Additive Manufacturing |
| Author(s) |
Akshay Jacob Thomas, Eduardo Barocio, Ilias Bilionis, R. Byron Pipes |
| On-Site Speaker (Planned) |
Akshay Jacob Thomas |
| Abstract Scope |
Physics-informed neural networks (PINNs) have recently been a popular framework to integrate experimental data and physics-based constraints specified via partial differential equations. However, the application of PINNs to additive manufacturing is limited since a suitable physics-based loss function is missing for geometries that evolve. The objective of this work is to address this gap. We propose a loss function for PINNs to solve the heat transfer equation on evolving geometries without mesh-based discretization. We use our methodology to predict the temperature evolution as a single bead is being deposited. We consider various cases of mixed Dirichlet, and Neumann boundary conditions and compare our results to finite element simulations. We also present guidelines to obtain consistent results using the proposed method. We finally discuss the potential of our method in solving parametric heat transfer problems in additive manufacturing. |
| Proceedings Inclusion? |
Definite: Post-meeting proceedings |