π Paper: A Neural Surrogate Approach for Simulating Natural Convection
Problems (arXiv:2606.25259) β Nurshat Menglik,
Alex Shao, David Hyde.
10,000 matched pairs of 2D natural-convection simulations of the differentially
heated square cavity under randomized wall-temperature boundary conditions. Each
sample solves the same problem twice β once with the Boussinesq model and once
with the fully compressible model ββ¦ See the full description on the dataset page:
https://huggingface.co/datasets/NurshatMenglik/Paired_Compressible_Boussinesq_Flow_Simulation_with_Random_Temperature_BCs.