| Field | Response |
|---|---|
| Intended Application & Domain: | Automotive Aerodynamics |
| Model Type: | Factorized Implicit Global Convolutional U-Net |
| Intended User: | Automotive engineers and researchers accelerating CFD predictions using AI. |
| Output: | Tensor (4 variables - pressure and wall shear stress components on vehicle surface). |
| Describe how the model works: | FIGConvUNet uses factorized implicit global convolutions in a U-Net architecture to predict aerodynamic surface fields from 3D point cloud geometry. |
| Name the adversely impacted groups this has been tested to deliver comparable outcomes regardless of: | N/A |
| Technical Limitations: | The model may perform poorly for vehicle geometries significantly different from the training data or for flow conditions outside the training regime. Physical consistency is not explicitly enforced. |
| Verified to have met prescribed NVIDIA quality standards: | Yes |
| Performance Metrics: | Relative Root Mean Square Error (RRMSE), Drag coefficient. |
| Potential Known Risks: | This model may inaccurately predict aerodynamic fields for vehicle designs outside the training distribution. |
| Licensing: | Use of this model is governed by the NVIDIA Open Model Agreement. |