Reproduce “Sample-Efficient Control of Magnetoconvective Flows via Differentiable Simulation” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Jannis Becktepe, Dmitry Krasnov, Nils Thuerey, Jörg Schumacher, Sebastian Peitz
The authors develop a differentiable simulator for low-magnetic-Reynolds-number magnetohydrodynamic liquid-metal flows, enabling gradients to guide control and inverse design. The solver adds electric-potential, current, and Lorentz-force physics to a fluid solver. In simulation, a gradient-based policy improves heat transfer by 28.59% at Ra=20Q; on a separate synthetic task, gradient-based optimization reduces probe mismatch. These results are limited to tested settings.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.