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 |
Thermal History Prediction of More Complex Geometries for DED Based on SciML |
Author(s) |
Bohan Peng, Ajit Panesar |
On-Site Speaker (Planned) |
Bohan Peng |
Abstract Scope |
Advancing from our PINN-based framework, we are proposing an eXtended physics-informed neural networks (XPINN)-based framework for thermal history prediction of parts during DED. Through a series of case studies against both the PINN solution as well as the benchmark (ANSYS solution), the XPINN-based framework has demonstrated significant improvement in accuracy (up to 50% reduction in RMSE and maximum absolute error) and more importantly, extended capability of thermal history prediction of more complex geometries with voids. The XPINN-based framework enables temperature history prediction of practical DED applications and brings process-aware design optimisation for real-life parts one step closer. |
Proceedings Inclusion? |
Planned: Post-meeting proceedings |