DrivAerCHX: Automotive Aerodynamics Dataset

Four datasets of DrivAer vehicle variants with surface pressure fields from LES simulations, generated using the ShapeCH method.

Description

This repository contains 4 datasets (DrivAerCH3, CH4, CH5, CH6) of geometric variants of the DrivAer automotive reference model with surface pressure distributions from Large Eddy Simulation (LES). The datasets were generated using the ShapeCH systematic sampling method developed in the referenced paper.

Each sample includes:

• VTP file: VTK PolyData mesh with per-triangle pressure values
• PNG images: 6 geometry views + 6 pressure field views

Structure

DrivAerCH3/    820 samples (3-corner convex hull)
DrivAerCH4/    814 samples (4-corner convex hull)
DrivAerCH5/    715 samples (5-corner convex hull)
DrivAerCH6/    792 samples (6-corner convex hull)

Each sample:
├── 0.vtp              # surface mesh with pressure (CellData/pressure)
└── pngs/
    ├── x_front.png, x_back.png, x_left.png, x_right.png, x_top.png, x_bott.png
    └── p_front.00150000.png, p_back.00150000.png, ...

Files

• VTP: Binary VTK PolyData, viewable in ParaView
• PNG: x_*.png are geometry views; p_*.png are pressure fields

For extracting quantitative data from the pressure images, you may find the CharLES PNG tools useful. Contact markben at stanford dot edu for more information.

Related Dataset

• mulligatawnysoup/DrivAerCH3

Reference

@article{benjamin2025systematic,
  title={A systematic dataset generation technique applied to data-driven automotive aerodynamics},
  author={Benjamin, Mark and Iaccarino, Gianluca},
  journal={APL Machine Learning},
  volume={3},
  number={1},
  year={2025},
  publisher={AIP Publishing}
}

Contact

Mark Benjamin (markben at stanford dot edu)