WindsorML: High-Fidelity Wall-Modelled LES Dataset for Automotive Aerodynamics

Provider
Neil Ashton et al. (caemldatasets.org)
License
CC-BY-SA-4.0 (source)
Size
100,000 files
Format
CSV (in sampled files)
Stored at
Hugging Face (https://huggingface.co/datasets/neashton/windsorml)
Access
Free and public. No account with the data provider is needed.
Category
Physics
Source
https://arxiv.org/abs/2407.19320

About WindsorML: High-Fidelity Wall-Modelled LES Dataset for Automotive Aerodynamics

WindsorML provides 355 geometric variants of the Windsor body, a simplified automotive shape, simulated with a GPU-native wall-modelled large-eddy simulation (WMLES) on Cartesian immersed-boundary meshes of about 300 million cells using the Volcano Platforms solver, each run for about 80 convective time units. It was published by Neil Ashton and colleagues to develop and test machine learning surrogates for external aerodynamics. Measured 2026-11-14, the repository root holds baseline_geometry/ and run_N folders; each run has STL and STEP geometry, geometry parameters, a boundary VTU with surface quadrature weights (NPY), a volume VTU of about 20 GB, time-averaged force and moment coefficients (Cd, Cs, Cl, Cmy) and slice images, and splits/ holds deterministic benchmark manifests. The publisher lists about 8 TB stored. Open VTU files with PyVista, VTK or ParaView and CSVs with pandas. Licensed CC-BY-SA-4.0; cite Ashton et al., arXiv:2407.19320.

How to load this dataset in Python

import pandas as pd
df = pd.read_csv("hf://datasets/neashton/windsorml/force_mom_all.csv")
df.head()

Data structure

File types in sampled files: CSV

Files

Example file: force_mom_all.csv

Analyze this dataset without downloading it

On Scigantic, WindsorML: High-Fidelity Wall-Modelled LES Dataset for Automotive Aerodynamics opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.

This page is also available as Markdown and as JSON.

Related Physics datasets