The topic of the project is simulation of fluid dynamics with deep generative models. Specifically simulation of bubbles with a Front Tracking approach [1] using Graph Neural Networks [2].
Supervisor: Vlado Menkovski, Ivo Roghair
[1] Lau, Y. M., et al. “Numerical investigation of the drag closure for bubbles in bubble swarms.” Chemical engineering science 66.14 (2011): 3309-3316.
[2] Pfaff, Tobias, et al. “Learning Mesh-Based Simulation with Graph Networks.” arXiv preprint arXiv:2010.03409 (2020).