A dataset of 49,660 parameterized 3-D computational fluid dynamics (CFD) surface meshes generated by IBM Research. Each sample is a distinct geometry from a 17-parameter family and includes per-vertex pressure and wall shear stress fields. All samples share the same triangulation topology of 9,600 triangles and 25,600 vertices.
Use Cases
- Training surrogate models for CFD simulations based on parameterized geometry variations.
- Analyzing wall shear stress (WSS) distributions for hemodynamic studies.
- Benchmarking mesh generation and processing algorithms on a large, consistent topology.
- Developing physics-informed neural networks (PINNs) for pressure field prediction.
Strengths
- Contains 49,660 distinct simulation samples, providing a substantial corpus for model training.
- All samples share a consistent mesh topology (9,600 triangles, 25,600 vertices), simplifying comparative analysis.
- Includes two key physical fields per vertex: pressure and wall shear stress (WSS).
- Geometries are generated from a defined 17-parameter family, allowing for controlled parametric studies.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Description metadata is limited; actual data quality and file formats require manual inspection after download.
Provenance
- Source
- IBM Research
- Collection Method
- Generated from a 17-parameter family of geometries via computational fluid dynamics simulation.
- Time Range
- null
- Freshness
- Last updated 2026-05-06 21:05:24; freshness should be verified.
- Geography
- null