airfoil-generator

Airfoil-generator is a tool generating airfoil flow field data for deep learning research.

This project is inspired by Deep-Flow-Prediction(code)

Requirements

Linux, Ubuntu 18.04 is recommended.

OpenFOAM-5.x and gmsh, this project use gmsh to generate unstructed mesh and use OpenFOAM-5.x(other version is not supported) as simulator. NOTE: There is two ways to install OpenFOAM, install from apt and build from source, install from apt in recommended.

 sudo sh -c "wget -O - https://dl.openfoam.org/gpg.key > /etc/apt/trusted.gpg.d/openfoam.asc"
 sudo add-apt-repository http://dl.openfoam.org/ubuntu
 sudo apt update
 sudo apt install openfoam5 gmsh
 source /opt/openfoam5/etc/bashrc  # activate openfoam5
 simpleFoam # check if installed successfully

airfoil-generator

cd
git clone https://github.com/yao-weijie/airfoil-generator.git && cd airfoil-generator/
python3 -m pip install -U pip
python3 setup.py install
airfoil_generator --help # check if installed successfully

Usage

It's better to generate data in a directory seperate from this project.

cd
mkdir airfoil-workspace
cp -r airfoil-generator/airfoil_database* airfoil-generator/config.yml airfoil-generator/case* -d airfoil-workspace

Airfoil_generator is a command-line tool, and can pass parameters from cli and a config file, and cli parameters have higher priority.

airfoil-generator --n-samples=3 --config=/path/to/config.yaml

An example of config yaml file:

#seed: 123456
case-dir: "caseSteadyState" # case-dir
airfoil-database: "airfoil_database" # airfoil database directory
fixed-airfoil: True # True to use only one airfoil "airfoil-name", False to randomly sample from "airfoil-database"
airfoil-name: "falcon"

freestream-angle: # inlet freestream angle, uniform randomly sample
  - 0
  - 20
freestream-length: # inlet freestream velocity, uniform randomly sample
  - 0
  - 50

# fluid parameters
rho: 1 # fluid density kg/m^3
nu: 1e-5 # fluid viscosity

# parallel parameters
parallel-enable: True
subdomains: 7 # subdomains if use parallel computing

# post process
res: 128 # resolution of output image
output-raw-mesh: True
output-airfoil-boundary: True

# outputs
n-samples: 2
output-dir: "outputs"
output-prefix: "sample"