mechanisms-behind-pvs-flow

Simulation code, meshes and associated data to reproduce numerical examples presented in "Mechanisms behind perivascular fluid flow", C. Daversin-Catty, V. Vinje, K-A. Mardal, and M.E. Rognes (2020)

Running 2D axi and 3D models

The 2D axi and 3D simulations are performed using CFD models based on the FEniCS project.

The corresponding Python scripts can be run within the latest FEniCS Docker container with the last version of Scipy installed :

git clone https://github.com/cdaversin/mechanisms-behind-pvs-flow.git
docker run -it -v $(pwd)/mechanisms-behind-pvs-flow:/home/fenics/shared quay.io/fenicsproject/dev
sudo pip3 install scipy --upgrade
cd shared

2D axi model

The 2D axi model can be configured and run using the script 2Daxi/script_2D.py

cd 2Daxi
python3 script_2D.py

3D model

The 3D model can be configured and run using the script 3D/script_C0075.py

cd 3D
python3 script_C0075.py

Graphs

The graphs presented in the paper can be reproduced using Jupyter notebook, running the corresponding scripts in a Web browser.

cd mechanisms-behind-pvs-flow/notebooks
jupyter-notebook

Note : The data files used in the notebooks are present in the repository by default, and are re-generated when running the models as described in the dedicated section

Mesh generation

The generation of the 3D PVS mesh presented in the paper is performed using VMTK on a clipped geometry from Aneurisk dataset (case id C0075). The meshes in mechanisms-behind-pvs-flow/3D/C0075_fine to be used in our 3D model are generated using the dedicated bash script

cd mechanisms-behind-pvs-flow/MeshGeneration
bash generate_meshes.sh

This mesh generation tool requires both VMTK and FEniCS.

Reporting issues

Any questions regarding this repository can be posted as Issues.