adFVM is an explicit unsteady compressible fluid dynamics package with adjoint capability using automatic differentiation provided by the library adpy.

adFVM requires the adpy library. To install it, after cloning, execute the following

git submodule update --init --recursive

Next build and install it using the following (using python2)

cd adpy
python setup.py build
python setup.py install --prefix=/path/you/want
cd ..

adFVM requires the following python packages, they will be automatically installed when you execute the python install script or you can install them separately if you want.

numpy
scipy
mpi4py
cython
matplotlib

Build and install adFVM

python setup.py build
python setup.py install --prefix=/path/you/want

Additionally for HDF5 capability parallel h5py needs to be installed. For running the adjoint solver using artificial viscosity PETSc and lapack need to be installed. For computing statistics of long-time averaged quantities in the primal and adjoint solvers the library ar needs to be installed. For handling meshes OpenFOAM and Metis need to be installed. For Ubuntu, the following script can be used to install all these packages.

./install_deps.sh

Edit the variable PYTHONPREFIX to change the python libraries installation location.

To run unit tests for adpy

cd adpy/tests
./run_tests.sh
cd ../..

To run unit tests for adFVM

cd tests
./setup_tests.sh
./run_tests.sh
cd ..

To use adFVM on a flow problem, a python case file needs to be defined that contains information about the details of the problem, like case folder, thermodynamic constants, design objective, simulation length, etc. Many example case files can be found in the templates folder. For a short explanation of the various options read the templates/vane.py file. To run the primal flow solver, use the following

./apps/problem.py templates/vane.py -c

To run the adjoint flow solver,

./apps/adjoint.py templates/vane.py -c

The sensitivities obtained using the adjoint solver can be found in the "objective.txt" file in the case folder. The "-c" option is not needed if running a flow solver multiple times.

In order to run the flow solver on multiple cores, first the mesh needs to be decomposed. Edit the "system/decomposeParDict" file in the case folder to set the number of processors, for example 4. Then, execute

decomposePar

To run the flow solver on 4 cores,

mpirun -np 4 ./apps/problem.py templates/vane.py 

The "objective.txt" for parallel runs is in the "processor0" directory in the case folder.

Chaitanya Talnikar and Professor Qiqi Wang