# The Well: Rayleigh-Benard Convection (2D)

## Facts

- Provider: Polymathic AI (The Well); Rudy Morel, Flatiron Institute CCM
- License: CC-BY-4.0 (https://arxiv.org/html/2412.00568)
- Size: 113 files
- Format: HDF5 (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/polymathic-ai/rayleigh_benard
- Access: public, readable without an account
- Category: Physics
- Source: https://arxiv.org/abs/2412.00568
- Page: https://scigantic.com/archive/9d88b098-517e-5822-a2eb-6c06d3e60d6c
- Keywords: Physics, PDE simulation, HDF5, Hugging Face

## About

2D horizontally periodic Rayleigh-Benard convection: a fluid layer heated from below and cooled from above, forming convection cells, plumes and boundary layers, solved with Dedalus under the Boussinesq approximation. Each simulation has 200 time steps of 512 by 128 images on a 4 by 1 domain over t from 0 to 50, periodic horizontally with Dirichlet conditions at the plates. Fields are buoyancy, pressure and velocity. The datasheet lists 1,750 simulations: 35 PDE parameter pairs (Rayleigh 1e6 to 1e10, Prandtl 0.1 to 10) times 50 initial conditions (5 buoyancy perturbation amplitudes times 10 noise seeds). Listing data/ (measured 2026-11-13) shows train/ (35 files, 294 GB, one per Rayleigh-Prandtl pair), valid/ and test/ (35 files, 37.9 GB each); about 370 GB in the repository. Files are HDF5 in the common The Well format; open them with the_well package (WellDataModule with base path hf://datasets/polymathic-ai/) or directly with h5py. Part of The Well (Ohana et al., NeurIPS 2024 Datasets and Benchmarks, arXiv:2412.00568); released under CC-BY-4.0, cite The Well and the simulation's source papers.

## Load the data

```python
from huggingface_hub import hf_hub_download
path = hf_hub_download(repo_id="polymathic-ai/rayleigh_benard", filename="data/test/rayleigh_benard_Rayleigh_1e10_Prandtl_1.hdf5", repo_type="dataset")
print(path)
```

## Data structure

File types in sampled files: HDF5

Example file: `data/test/rayleigh_benard_Rayleigh_1e10_Prandtl_1.hdf5`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/9d88b098-517e-5822-a2eb-6c06d3e60d6c
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/9d88b098-517e-5822-a2eb-6c06d3e60d6c
- Over MCP: `get_schema_card` with id `9d88b098-517e-5822-a2eb-6c06d3e60d6c` at https://api.scigantic.com/mcp

## Related topics

- [HDF5 datasets](https://scigantic.com/topics/hdf5)
- [Physics datasets](https://scigantic.com/topics/physics)
