Finite element library
Fedoo is an open source Finite Element library developed in Python. It is mainly dedicated for mechanical problems but is easily developpable for other kind of problems (thermal laws already includes). One of the main application of fedoo is to simulate the mechanical response of heterogeneous materials. For that purpose, fedoo is part of the 3mah set that also include microgen for the CAD and meshing of heterogeneous materials and simcoon for the non linear material constitutive models in finite strain.
Here are the main features:
- Entirely written in Python 3
- Implicit finite element Solver for Static and Dynamics poblems
- Finite strain constitutive laws based on the simcoon library (simcoon is developped in C++ allowing a fast execution)
- Geometrical non linearities based on the simcoon library
- Wide constitutive equation library including composites law, elasto-plastic law, ...
- Include many types of elements like 2D, 3D, beam, plate, cohesive elements, ...
- Homogeneisation: Easy application of periodic boundary conditions and fast automatic extraction of mean tangent matrices
- Embedded results visualization using the powerfull pyvista library
- Multi-point constraints
- Easy scripting
- Good compromise between a reasonable execution time and an open and lisible code. fedoo is not the fastest finite element software and doesn't intend to be, but a particular attention is paid to the computational cost.
- Contact in 2D and 3D, Self contact
- And many other....
| PyPI package |  |
| Conda package |  |
| Documentation | |
| License |  |
| Website |  |
The documentation is provided by readthedocs at https://fedoo.readthedocs.io.
Installation with pip including recommanded dependencies excepted simcoon:
pip install fedoo[all]
Minimal installation with pip:
pip install fedoo
Installation with conda with recommanded dependencices (including simcoon):
conda install -c conda-forge -c set3mah fedoo
In mac OS make sure to use scikit-umfpack, the best way to take the most of MacOs accelerate framework is to install numpy from sources first:
pip install cython pybind11
pip install --no-binary :all: numpy
The conda package is restricted to version 0.33, which is not compatible with the latest versions of numpy
pip install scikit-umpack>=0.41