Metal additive manufacturing (AM) experiments are slow and expensive. Engineers are using physics-informed neural networks to predict the outcomes of complex processes involved in AM. The team trained the model on supercomputers using experimental and simulated data. They recreated the dynamics of two benchmark experiments in metal AM. The method could lead to fast prediction tools for AM in the future.
Using AI to predict 3D printing processes
- ScienceDaily
- 1. 7 2021 (20:09)