Skip to content

Dataset · ShapeBench

Aerodynamic simulations of aircraft geometry

ShapeBench/COCOANet

A tabular dataset with 3,570 flow simulations for 401 parametric aircraft geometries for aerodynamics tasks.

What's inside

The dataset from the ShapeBench project includes 3,570 computational fluid dynamics (CFD) runs in the Flow360 solver for 401 aircraft geometry variants created in the nTop CAD kernel. Each record is described by 16 geometric parameters and 3 flight parameters (angle of attack, speed, and altitude) selected via Latin hypercube sampling. The total volume ranges from 1k to 10k rows.

What you can do with it

  • Training surrogate models for fast prediction of aerodynamic characteristics
  • Tabular regression and wing/fuselage aerodynamic shape optimization tasks
  • Benchmarking machine learning algorithms in engineering design
Tasks
Tabular regression
Size
1K–10K rows
Downloads this month
31.8K
Likes
1

How to load it

from datasets import load_dataset

ds = load_dataset("ShapeBench/COCOANet")