logo made by Steph Olinger
finite element method (FEM) codes to calculate glacier crevasse stress intensity factors
Most of the important stuff is in crevasse.py
The notebooks contain:
- An example single-parameter run crevasse-model-single-parameter-set.ipynb
- An example parameter space study, crevasse-model-parameter-space.ipynb
- An effort to debug the non-smooth results, test-swell-phase-picker.ipynb
- Clone this repository.
- Run a Docker container with FEniCS (the FEM backend).
- Install a few things in the container.
The easiest option to get started is to use Jupyter. The following docker command will launch a Jupyter Notebook server where you can run the notebooks. For the following docker run commands, run these from the directory that contains the cloned repo.
docker run --name crevasserator -w /home/fenics -v (pwd):/home/fenics/shared -d -p 127.0.0.1:8888:8888 quay.io/fenicsproject/stable 'jupyter-notebook --ip=0.0.0.0'
We have run parameter studies with up to 1e4 individual runs. Scaling in this way can't be achieved using Jupyter since the overhead is too high and the sessions don't persist. Instead, for this application, we want to create a container with an interactive terminal:
docker run --name crevasserator_it -it -w /home/fenics -v (pwd):/home/fenics/shared quay.io/fenicsproject/stable
After creating the container, one useful workflow is to start screen (to ensure persistence) and then start and attach the container,
screen
docker start crevasserator_it
docker attach crevasserator_it
Finally, you'll need to install some things in the container:
sudo apt-get update -y
sudo apt-get install -y python-gmsh
pip install --user gmsh h5py meshio