AB-UPT: Scaling Neural CFD Surrogates for High-Fidelity Automotive Aerodynamics Simulations via Anchored-Branched Universal Physics Transformers
Benedikt Alkin, Maurits Bleeker, Richard Kurle, Tobias Kronlachner, Reinhard Sonnleitner, Matthias Dorfer, Johannes Brandstetter, * Equal contribution Emmi AI ELLIS Unit, LIT AI Lab, JKU LinzCorrespondence to [email protected] · 2025
Read AB-UPT: Scaling Neural CFD Surrogates for High-Fidelity Automotive Aerodynamics Simulations via Anchored-Branched Universal Physics Transformers by Benedikt Alkin, Maurits Bleeker, Richard Kurle, Tobias Kronlachner, Reinhard Sonnleitner, Matthias Dorfer, Johannes Brandstetter, * Equal contribution Emmi AI ELLIS Unit, LIT AI Lab, JKU LinzCorrespondence to [email protected] (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-09692.txt.
Introduction
Benedikt Alkin and co-authors present AB-UPT in this 2025 paper, a neural stand-in for costly airflow simulations around cars. Car firms test shapes with physics software, but fine meshes can hold a hundred million cells. Neural models promise fast answers, yet they strain at that scale and must respect physics rules. The team's design splits shape encoding from prediction across branches.
The model runs its simulation in a small latent space, then decodes to full detail at anchored points. A built-in rule keeps swirling flow fields free of sources, as physics demands. On car tests from tiny to 150-million-cell meshes, accuracy led rival methods. Training took under a day on one graphics card. Predictions took seconds and could start from a design file alone, skipping mesh building.
Catalog record
| CREATOR | Benedikt Alkin, Maurits Bleeker, Richard Kurle, Tobias Kronlachner, Reinhard Sonnleitner, Matthias Dorfer, Johannes Brandstetter, * Equal contribution Emmi AI ELLIS Unit, LIT AI Lab, JKU LinzCorrespondence to [email protected] |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 137 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2502-09692