# Deep Learning for Real-Time Aerodynamic Evaluations of Arbitrary Vehicle Shapes

arXiv: https://arxiv.org/abs/2108.05798 · HTML: https://mlsanity.com/paper/2108.05798/

## Key facts

- Cited by 18 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 12 Aug 2021.

## Built on by (most cited first, 18 of 18)

- [DrivAerNet++: A Large-Scale Multimodal Car Dataset with Computational Fluid Dynamics Simulations and Deep Learning Benchmarks](https://mlsanity.com/paper/2406.09624/) (2024)
- [DrivAerNet: A Parametric Car Dataset for Data-Driven Aerodynamic Design and Prediction](https://mlsanity.com/paper/2403.08055/) (2024)
- [DoMINO: A Decomposable Multi-scale Iterative Neural Operator for Modeling Large Scale Engineering Simulations](https://mlsanity.com/paper/2501.13350/) (2025)
- [X-MeshGraphNet: Scalable Multi-Scale Graph Neural Networks for Physics Simulation](https://mlsanity.com/paper/2411.17164/) (2024)
- [Surrogate Modeling of Car Drag Coefficient with Depth and Normal Renderings](https://mlsanity.com/paper/2306.06110/) (2023)
- [AhmedML: High-Fidelity Computational Fluid Dynamics Dataset for Incompressible, Low-Speed Bluff Body Aerodynamics](https://mlsanity.com/paper/2407.20801/) (2024)
- [WindsorML: High-Fidelity Computational Fluid Dynamics Dataset For Automotive Aerodynamics](https://mlsanity.com/paper/2407.19320/) (2024)
- [TripNet: Learning Large-scale High-fidelity 3D Car Aerodynamics with Triplane Networks](https://mlsanity.com/paper/2503.17400/) (2025)
- [Drag-guided diffusion models for vehicle image generation](https://mlsanity.com/paper/2306.09935/) (2023)
- [CarBench: A Comprehensive Benchmark for Neural Surrogates on High-Fidelity 3D Car Aerodynamics](https://mlsanity.com/paper/2512.07847/) (2025)
- [A Benchmarking Framework for AI models in Automotive Aerodynamics](https://mlsanity.com/paper/2507.10747/) (2025)
- [Benchmarking Convolutional Neural Network and Graph Neural Network based Surrogate Models on a Real-World Car External Aerodynamics Dataset](https://mlsanity.com/paper/2504.06699/) (2025)
- [DrivAer Transformer: A high-precision and fast prediction method for vehicle aerodynamic drag coefficient based on the DrivAerNet++ dataset](https://mlsanity.com/paper/2504.08217/) (2025)
- [A Geometry-Aware Triplane Field Network for Vehicle Aerodynamic Prediction](https://mlsanity.com/paper/2606.07724/) (2026)
- [Learning with Geometry: Including Riemannian Geometric Features in Coefficient of Pressure Prediction on Aircraft Wings](https://mlsanity.com/paper/2401.09452/) (2023)
- [Learning Mappings in Mesh-based Simulations](https://mlsanity.com/paper/2506.12652/) (2025)
- [PMRT: A Training Recipe for Fast, 3D High-Resolution Aerodynamic Prediction](https://mlsanity.com/paper/2509.17182/) (2025)
- [HGPTrans: Hierarchical Graph-Pooling Transolver for Automotive Aerodynamic Drag Coefficient Prediction](https://mlsanity.com/paper/2609.31765/) (2026)

## Abstract (arXiv)

The aerodynamic optimization process of cars requires multiple iterations between aerodynamicists and stylists. Response Surface Modeling and Reduced-Order Modeling are commonly used to eliminate the overhead due to Computational Fluid Dynamics, leading to faster iterations. However, a primary drawback of these models is that they can work only on the parametrized geometric features they were trained with. This study evaluates if deep learning models can predict the drag coefficient for an arbitrary input geometry without explicit parameterization. We use two similar data sets based on the publicly available DrivAer geometry for training. We use a modified U-Net architecture that uses Signed Distance Fields to represent the input geometries. Our models outperform the existing models by at least 11% in prediction accuracy for the drag coefficient. We achieved this improvement by combining multiple data sets that were created using different parameterizations, which is not possible with the methods currently used. We have also shown that it is possible to predict velocity fields and drag coefficient concurrently and that simple data augmentation methods can improve the results. In addition, we have performed an occlusion sensitivity study on our models to understand what information is used to make predictions. From the occlusion sensitivity study, we showed that the models were able to identify the geometric features and have discovered that the model has learned to exploit the global information present in the SDF. In contrast to the currently operational method, where design changes are restricted to the initially defined parameters, this study brings surrogate models one step closer to the long-term goal of having a model that can be used for approximate aerodynamic evaluation of unseen, arbitrary vehicle shapes, thereby providing complete design freedom to the vehicle stylists.