About this Abstract |
| Meeting |
TMS Specialty Congress 2026
|
| Symposium
|
World Congress on Reproducibility, Qualification, and Standards Development of Additive Manufacturing and Beyond (RQSD 2026)
|
| Presentation Title |
Multi-scale Physics-Informed Surrogate Modeling for Additive Manufacturing of Metallic Alloys |
| Author(s) |
Dilip Kumar Banerjee, Daniel Wheeler, Shengyen Li |
| On-Site Speaker (Planned) |
Dilip Kumar Banerjee |
| Abstract Scope |
This study develops a high-fidelity digital twin for laser powder bed fusion (LPBF) by integrating multi-scale physics with efficient surrogate modeling. We present a comprehensive physics-based framework that couples macroscopic phenomena such as fluid flow, heat transfer with microscopic solidification microstructure predictions. This multi-scale model is rigorously validated against experimental LPBF data to ensure physical accuracy. To overcome the computational intensity of such simulations, we utilize the validated data to train a series of surrogate models, specifically exploring Fourier Neural Operators (FNO) for accelerated spatial-temporal predictions. These surrogates are benchmarked against simplified models like 3DThesis to evaluate their reliability for real-time, practical applications. The results will show that physics-informed surrogate models can accurately mimic complex melt-pool dynamics and microstructure evolution at a fraction of the computational cost. This provides a robust pathway for real-time process parameter optimization and uncertainty quantification to achieve prescribed build quality in additive manufacturing. |
| Proceedings Inclusion? |
Undecided |