AutoHood3D

A Multi‑Modal Benchmark for Automotive Hood Design and Fluid–Structure Interaction

AutoHood3D is an open‑source dataset and framework comprising over 16 000 parametric hood geometries with coupled high‑fidelity LES–FEA simulations. It supports CAD‑driven generative‑AI, surrogate modeling, physics‑informed ML, and LLM fine‑tuning.

---

🛠️ REPOSITORY UPDATES

2025‑05‑10 v1.0 initial public release: 16 000+ geometries, full end‑to‑end pipeline

---

⚖️ License

Shield:

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. Licensing note: The dataset is licensed under CC BY-NC 4.0, and the code is licensed under the PolyForm Noncommercial License 1.0.0. For commercial usage, please contact the authors.

---

📥 Dataset Access

• Harvard Dataverse:

  • Base Hood Skins: https://doi.org/10.7910/DVN/9268BB

  • Dataset 4k Hoods STLs: https://doi.org/10.7910/DVN/HEILMB

  • Dataset 4k Hoods Sim Data: https://doi.org/10.7910/DVN/VCKOK5

  • ML-PreSplit for 4k Hoods: https://doi.org/10.7910/DVN/6OAFF8

  • ML (Graph)-PreSplit for 4k Hoods: https://doi.org/10.7910/DVN/WODNWY

  • Test ML Workflow for 100 Hoods (from 4k set): https://doi.org/10.7910/DVN/FSYRJA

  • Dataset 12k Hoods STLs: https://doi.org/10.7910/DVN/Z0VXLI

  • Dataset 12k Hoods SimData A (around 0.7TB) : https://doi.org/10.7910/DVN/BVPATN

  • Dataset 12k Hoods SimData B (around 0.7TB) : https://doi.org/10.7910/DVN/UDXEG9

  • Dataset 12k Hoods STLs (random): https://doi.org/10.7910/DVN/OJXIS1

• LLM SFT Prompts for Point Clouds:

  • 06_LLM_Generation/Final_consolidated_prompts.jsonl

---

📂 Directory Structure

• 00_Base_Hoods_and_Curves # Raw CAD and curve libraries
• 01_Generating_Hoods # Convex‐hull, segmentation & shell reconstruction
• 02_HPC_Automation # Scripts for case setup & SLURM orchestration
• 03_FSI_Solvers # Custom UM_pimpleFoam & UM_solidDisplacementFoam
• 04_Preprocessing_data_for_ML # Mesh‐to‐point/cloud conversion & feature extraction
• 05_ML_Framework # Training & evaluation pipelines for surrogate models
• 06_LLM_Generation # Prompt generation & LLM fine‑tuning scripts
• 07_Postprocessing # Visualization and benchmark plots
• CroissantData # JSON descriptors for Dataset Metadata

---

🚀 Quick Start

1. Clone the repo

  git clone https://github.com/YourOrg/AutoHood3D.git
  cd AutoHood3D

2. Install dependencies

pip install -r 05_ML_Framework/package.list
# plus OpenFOAM v2312, preCICE v3.1.2, preCICE OpenFOAM adapter v1.3.0

3. Run

• Generate shell variants: python 01_Generating_Hoods/...
• Launch FSI co‑simulation: see 02_HPC_Automation/...
• Preprocess for ML: python 04_Preprocessing_data_for_ML/...
• Train surrogate models: python 05_ML_Framework/...
• Fine‑tune LLM: python 06_LLM_Generation/...
• Plot results: python 07_Postprocessing/viz...

NOTES : Each folder contains separate instructions, please check README files.

---

🤝 Contributing

Issues and pull requests welcome via GitHub.

---

📬 Contact

Authors: - Vansh Sharma, Harish Jai Ganesh, Maryam Akram, Wanjiao Liu and Venkat Raman - Email at: vanshs@umich.edu and ramanvr@umich.edu

Research Group: - Advanced Propulsion Concepts Lab
- Department of Aerospace Engineering, University of Michigan, Ann Arbor

---