FSGe: A fast and strongly-coupled 3D fluid-solid-growth interaction method

Reference

arXiv:2404.14187

Quickstart

1. Install docker and pull the svMultiphyiscs image from

docker pull simvascular/solver:latest

Make two directories, one for svMultiPhysics(svfsi), one for svFSGe:

mkdir svFSGe && mkdir svMultiPhysics

Start a new svMultiphysics container by:

docker run -it --user $(id -u):$(id -g) \
           -v ./svMultiPhysics:/svfsi \
           -v ./svFSGe:/svFSGe        \
              simvascular/solver:latest /bin/bash

2. Inside the container, install svFSIplus from this branch, to build svfsi, you need to run(inside docker, command line begins with #):

cd /svfsi
git init
git remote add origin https://github.com/Eleven7825/svMultiPhysics.git
git fetch origin FSGe
git checkout -b FSGe origin/FSGe

At the /svfsi directory, build it with

bash makeCommand.sh

3. Install required Python packages (inside the container):

# Option 1: Install from requirements.txt (recommended)
pip install -r /svFSGe/requirements.txt

# Option 2: Install individually
pip install numpy vtk matplotlib scipy xmltodict distro

4. Still inside the container, you need to adapt paths in /svFSGe/in_sim/partitioned_full.json to make sure they are correct

5. Run the simulation:

cd /svFSGe
python3 ./fsg.py in_sim/partitioned_full.json

Quick Setup Script

For convenience, use the provided setup script to automate Docker environment setup:

# From the svFSGe directory
./scripts/setup_docker.sh

This script will:

• Create necessary directories
• Start Docker container with proper mounts
• Clone and build svFSIplus
• Install Python dependencies
• Provide an interactive shell ready to run simulations

Continuous Integration

GitHub Actions automatically tests FSGe on every pull request and branch push:

• Test configuration: in_sim/partitioned_test.json (nmax=2 for faster testing)
• Workflow: .github/workflows/test-fsg.yml
• Comparison: Validates convergence metrics against reference baseline

See test_reference/nmax_2/README.md for details on reference data.

File overview

• cylinder.py generates structured FSI hex-meshes with configuration files in in_geo
• fsg.py runs partitioned FSGe coupling using svFSIplus with
  • in_sim FSGe configuration files
  • in_svfsi_plus svFSIplus input files
  • in_petsc PETSc linear solver settings
• post.py generate line plots from FSGe results
• svfsi.py sets up, executes, and processes svFSIplus simulations
• utilities.py IQN-ILS filtering
• vtk_functions.py useful VTK functions for file IO
• scripts/ utility scripts including:
  • compare_results.py - CI test comparison script
  • setup_docker.sh - Automated Docker environment setup