discreteAdjointOpenFOAM
Description
We develop a discrete adjoint version of OpenFOAM (currently based on OpenFOAM(.com) v1812). It can be applied to topology optimization cases, but a wide range of other use-cases can also be tackled due to the flexibility of the discrete adjoint approach. The following features are available:
- Full first order adjoint of the whole OpenFOAM package
- Equidistant or Revolve Checkpointing
- Symbolically differentiated linear solvers
- Parallelization with adjoint-MPI (AMPI)
- Tangent Linear and Finite Difference Implementations for verification
The source is available upon request. Please drop us a mail and we'll grant you access to the repository (please include your public ssh key).
A work in progress wiki with usage instructions can be found here.
Shape Sensitivities of Sonnenwagen / Image Credit: Lennart Moltrecht, sonnenwagen.org
Publications
- 2013 A Discrete Adjoint Model for OpenFOAM Proceedings of ICCS 2013
- 2014 A Discrete Adjoint Version of an Unsteady Incompressible Solver for OpenFOAM using Algorithmic Differentiation Eccomas 2014
- 2015 MPI-Parallel Discrete Adjoint OpenFOAM Proceedings of ICCS 2015
- 2018 SIMPLE adjoint message passing Optimization Methods and Software
- 2019 Discrete Adjoint Optimization with OpenFOAM
- 2021 Discrete Adjoint Approaches for CHT Applications in OpenFOAM, Advances in Evolutionary and Deterministic Methods for Design, Optimization and Control in Engineering and Sciences
Workshop Materials
- Introduction to Algorithmic Differentiation available here
- Advanced Topics in Algorithmic Differentiation available here
- Introduction to Discrete Adjoint OpenFOAM available here
- Workshop Materials from OpenFOAM Workshop 2019: slides handout
Authors
Contact
Disclaimer
This offering is not approved or endorsed by OpenCFD Limited, producer and distributor of the OpenFOAM software and owner of the OPENFOAM® and OpenCFD® trade marks.