SHIFT-Wing-sample: High-Fidelity Computational Fluid Dynamics Dataset for Transonic Wing External Aerodynamics

We're excited to introduce the SHIFT-Wing dataset — a high-fidelity aerodynamic simulation dataset developed as part of the Luminary SHIFT Models initiative. This dataset enables the training and benchmarking of real-time physics AI models for transonic wing external aerodynamics.

Website: www.luminarycloud.com/models

Contact: [email protected]

Note: this is a sample dataset of the first ~100 samples in the full SHIFT-Wing. It provides the same information as provided in the full, except that no volume solutions are provided per-sample. This should be sufficient to assess pipelines and test whether you want access to the full dataset.

Summary

Physics AI models can transform early stage plane design by giving users real-time feedback on the physics-based performance implications of design decisions. However, the lack of a sufficiently large high-quality training dataset has been a barrier to their development. Luminary SHIFT Models provide access to both high-quality datasets and pretrained models for a variety of applications and industries.

SHIFT-Wing is a massive step forward in this direction: purpose-built for high-fidelity aerodynamic inference of non-linaer flow fields, without requiring CFD expertise or meshing. Developed in collaboration with Otto Aviation and NVIDIA, SHIFT-Wing is based on thousands of parametric variants of the high speed NASA Common Research Model (CRM), developed as an open-source generic high speed transport aircraft geometry.

This dataset supports training surface-based or volume-based aerodynamic surrogate models, real-time inference systems, and exploring shape-performance correlations for aerospace design.

Applications

• Rapid aerodynamic prototyping and shape optimization
• Research in aero-inference, point cloud learning, or physics-aware generative models
• Training and fine-tuning Physics AI models

Attribution

Please attribute NASA for the original CRM model, and Luminary Cloud for the SHIFT-Wing model and dataset.

An article is being prepared so users can cite this dateset - we will update this accordingly when available. Until then you can use this citation:

@misc{shift_wing_2025,
  author = {Luminary Cloud},
  title = {SHIFT-Wing: High-Fidelity Computational Fluid Dynamics Dataset for Transonic Aerospace External Aerodynamics},
  year = {2025},
  url = {https://huggingface.co/datasets/luminary-shift/Wing}
}

Contents

This repository contains the SHIFT-Wing dataset. We will continue to push newly computed samples to this repository periodically (approximately monthly). The data generation and organization within the repository is described below.

Geometry Variation

Currently only one geometry parameterization is provided. The organization supports future samples where input samples are parameterized differently.

OnShape: Luminary DPW4, wing-only

A fully parameteric NASA CRM model was constructed; the reference model distributed by NASA is used as a starting point. The parameterization is divided between the wing and fuselage.

For the wing, the reference CAD model wing is intersected at 6 spanwise stations and the resulting airfoil is extracted as a 2D profile. The position, scaling, and rotation of these profiles are parameterized with several common planform and wing design parameters, and re-lofted to produce new variations of the wing.

For the fuselage, the tube and juncture are defined separately. The tube is parameterized by its diameter and length, while the juncture shape and position is based on the wing parameters. This is to ensure the juncture appropriately mates the wing to the fueslage, regardless of the wing parameter values used.

The OnShape document leverages the Variable Studio table to implement the parameterization. While a specific subset of design parameters are exposed in this dataset, the approach (define arbitrary number of airfoils profiles, perform transformations of these profiles, loft the wing) allows users to define their own parameterization of the model quite easily.

CFD Solver

All cases were run using the Luminary Cloud platform. The cases were simulated using simulation practices honed during our participation in the NASA HLPW and leverages our automated adaptive solution technology: Luminary Mesh Adaptation (LMA). This ensures thin and sharp features, such as transonic shocks, are accurately captured by the solver with no user input and across the wide range of flow conditions and geometries.

Files

At the top level of the repository you will see directories for each of the Geometry parameterizations described above (currently only one):

OnShape_luminary_crm_version001

Within this directory you will see folders simply indexed by sample_xxxxxx, where the indices are a zero-padded six digit integer. Each of these samples contains a set of files describing their geometry and simulation settings, among other artifacts.

In each directory you will find the following files:

• merged_surfaces.stl: STL file with the crm geometry. This represents the surface tessellation from the final adapted mesh in the LMA sequence
• merged_surface_inter_mesh_{index}.stl: the mesh adaptation procedure creates a series of successively adapted and refined meshes. Here we provide several, though not all, intermediate surface meshes in the default 13-mesh sequence (ranging from ~250k cells to 30M cells). The merged_surfaces.stl dscribed above is the final surface mesh in ther sequence.
• merged_surfaces_onshape_200k.stl: an STL of the same geometry exported from OnShape directly, containing only 200k vertices to describe the geometry. This representation is also more isotropically sampled.
• merged_surfaces.vtp: surface field solution file with pressure and wall shear-stress fields at the final iteration
• viz: directory of images showing the geometry and flow fields
• forces.json: file containing drag, lift, normalized coefficients, and moments from the final iteration
• params.json: description of geometry parameters and flow parameters defining both the geometry sample and the flow condition

Downloading

You can use HuggingFace to gain access to the entire repository, but will require the associated TBs of storage available locally. Note you will need to have git lfs installed first, then run

git clone [email protected]:datasets/luminary-shift/Wing-sample

If you will access only a subset of the data, or wish to interact in a staged manner, you can clone the repository where the LFS files are not checked out (simply pointers):

GIT_LFS_SKIP_SMUDGE=1 git clone [email protected]:datasets/luminary-shift/Wing-sample .

# to ensure future `git pull` commands won't checkout full files, you'll want to ensure the skip is active in this repo
cd <path/to/repo>
git lfs install --skip-smudge --local

You can the pull down files you want to interact with in multiple ways:

# pull a specific file
git lfs pull --include="path/to/your/file"

# pull a directory
git lfs pull --include="path/to/file1,path/to/dir/*"

# pull, but exclude certain paths
git lfs pull --exclude="**/*.mp4"

and remove those files and reset them to pointers when done using them:

rm path/to/your/file
git checkout -- path/to/your/file

Credits

Tom Wayman, Otto Aviation

Guidance on geometry variants and industry relevant parameterization of the geometry was provided by Otto Aviation.

NVIDIA

NVIDIA provided tools and infrastructure for model training via PhysicsNeMo and DoMINO. This is relevant if accessing any pretrained models.

License

This dataset is distributed under the CC-BY-4.0 license, which is also included in the dataset itself. By downloading the dataset you acknowledge the terms of this license.