HiLiftAeroML: High-Fidelity CFD Data for High-Lift Aircraft

1,800 high-fidelity computational fluid dynamics samples for the NASA Common Research Model high-lift variant. The dataset was created by NVIDIA for AI surrogate model development and includes 180 geometry variants across 10 angles of attack. It was last updated on May 15, 2026.

Use Cases

• Train surrogate models for aerodynamic force prediction based on geometry and angle-of-attack variations.
• Validate machine learning methods for computational fluid dynamics based on high-fidelity simulation data.
• Benchmark AI performance in predicting complex flow fields around high-lift aircraft configurations.
• Explore the design space of high-lift devices using data-driven methods based on parametric geometry changes.

Strengths

• 1,800 high-fidelity CFD samples provide a substantial foundation for model training.
• Covers 180 distinct geometry variants of a standard aerospace benchmark model.
• Includes data across 10 angles of attack, from 4° to 22°, capturing a range of flight conditions.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment for specific model architectures.

Provenance

Source

NVIDIA

Collection Method

High-fidelity computational fluid dynamics simulations of the NASA Common Research Model high-lift variant.

Freshness

Last updated 2026-05-15 09:54:18