Machine Learning in OpenFOAM

Contributors

• Tomislav Marić - Maintainer/Developer - Contact
• Andre Weiner - Developer - Contact
• Antonio Martín-Alcántara - Developer - Contact
• Niklas Bönisch - Developer/Tester - Contact

Surface Approximation

Neural Networks for approximation of implicit surfaces on unstructured meshes.

Deep Reinforcement Learning

Deep Reinforcement Learning algorithms for solution control of coupled PDEs.

Getting Started

These instructions will get you a copy of the project up and running on your local machine for development and testing purposes. See deployment for notes on how to deploy the project on a live system.

Prerequisites

What things you need to install the software and how to install them

• Python: Python 3.9.1, python-pandas, python-jupyter, python-matplotlib, python-numpy
• OpenFOAM: v2012, v2006
• libtorch: C++ API for pytorch, tested with 1.7.1

Installing

Install libtorch PyTorch C++ API either using a package manager, or download the archive from the PyTorch website and extract the archive in the openfoam-ml folder.

For example, get the latest CPU distribution of libtorch,

    openfoam-ml> wget https://download.pytorch.org/libtorch/cpu/libtorch-cxx11-abi-shared-with-deps-1.7.1%2Bcpu.zip 
    openfoam-ml> unzip libtorch-cxx11-abi-shared-with-deps-1.7.1+cpu.zip

Configure variables used to include and link PyTorch files and libraries,

    openfoam-ml> source setup-torch.sh

If not already set up, set up the OpenFOAM Environment

    openfoam-ml> source /path/to/openfoam/etc/bashrc

Build openfoam-ml by calling

?> ./Allwmake

Running the tests

TODO: Explain how to run the automated tests for this system

Break down into end to end tests

Surface approximation

unit_box_domain test

?> blockMesh 
?> aiFoamLearnSurface

Fields with nn in their name are approximated by Neural Networks and can be visualized in ParaView.

Contributing

Contact the maintainer for the access to the project, report bugs and request new features here.

License

This project is licensed under the GPL 3.0 License.