Datasetpublic

giovannicatalani/RAE2822_Airfoil_Dataset

2D Transonic RAE2822 Airfoil Dataset This dataset contains computational fluid dynamics (CFD) simulations of the transonic RAE2822 airfoil, as presented in the paper: "A comparative study of learning techniques for the compressible aerodynamics over a transonic RAE2822 airfoil"Computational Fluid Dynamics Journal, 2022DOI: 10.1016/j.compfluid.2022.105759 Dataset Description This dataset provides high-fidelity CFD simulation results for the RAE2822 airfoil across… See the full description on the dataset page: https://huggingface.co/datasets/giovannicatalani/RAE2822_Airfoil_Dataset.

sourceHugging Faceapache-2.0updated 1y agoView on Hugging Face

1likes149downloads

Dataset Card

2D Transonic RAE2822 Airfoil Dataset

This dataset contains computational fluid dynamics (CFD) simulations of the transonic RAE2822 airfoil, as presented in the paper:

"A comparative study of learning techniques for the compressible aerodynamics over a transonic RAE2822 airfoil" Computational Fluid Dynamics Journal, 2022 DOI: 10.1016/j.compfluid.2022.105759

Dataset Description

This dataset provides high-fidelity CFD simulation results for the RAE2822 airfoil across various flight conditions in the transonic regime. The data includes pressure and velocity fields computed on structured grids, making it suitable for machine learning applications in computational aerodynamics.

Key Features

—Airfoil: RAE2822 (classic transonic airfoil benchmark)
—Flow Regime: Transonic (Mach 0.0 to 0.9)
—Angle of Attack Range: 0° to 9°
—Grid Type: Structured computational mesh
—Variables: Pressure, velocity components (Vx, Vy), coordinates

Files Description

1. `airfoil.npy`

Contains the RAE2822 airfoil coordinates defining the geometry.

2. `db_random.npy`

Dataset generated using random sampling of flow conditions:

—Mach number: 0.0 to 0.9 (uniform random distribution)
—Angle of attack: 0° to 9° (uniform random distribution)
—Sampling method: Monte Carlo random sampling

3. `db_cyc.npy`

Dataset generated using Clenshaw-Curtis quadrature rule:

—Mach number: 0.0 to 0.9 (structured sampling)
—Angle of attack: 0° to 9° (structured sampling)
—Sampling method: Clenshaw-Curtis quadrature for better parameter space coverage

Data Structure

Each dataset file (.npy) contains a dictionary with the following keys:

Key	Description	Shape
`Pressure`	Pressure field on computational grid	`[n_samples, grid_x, grid_y]`
`Vx`	X-component of velocity field	`[n_samples, grid_x, grid_y]`
`Vy`	Y-component of velocity field	`[n_samples, grid_x, grid_y]`
`Xcoordinate`	X-coordinates of grid points	`[grid_x, grid_y]`
`Ycoordinate`	Y-coordinates of grid points	`[grid_x, grid_y]`
`Vinf`	Freestream Mach number for each sample	`[n_samples]`
`Alpha`	Angle of attack for each sample (degrees)	`[n_samples]`
`idx`	Sample indices	`[n_samples]`

Usage Example

python

import numpy as np
import os

# Load the datasets
data_directory = "path/to/dataset"
db_random = np.load(os.path.join(data_directory, 'db_random.npy'), allow_pickle=True).item()
db_cyc = np.load(os.path.join(data_directory, 'db_cyc.npy'), allow_pickle=True).item()

# Load airfoil coordinates
airfoil_coords = np.load(os.path.join(data_directory, 'airfoil.npy'))