This repository contains the source code and data to reproduce the results presented in the paper:
"Reduction of Outflow Boundary Influence on Aerodynamic Performance using Neural Networks"
by Mario C. Bedrunka, Dirk Reith, Holger Foysi, Łukasz Łaniewski-Wołłk, and Travis Mitchell.
The accurate treatment of outflow boundary conditions remains a critical challenge in computational fluid dynamics when predicting aerodynamic forces and/or acoustic emissions. This is particularly evident when employing the lattice Boltzmann method (LBM) as the numerical solution technique, which often suffers from inaccuracies induced by artificial reflections from outflow boundaries. This paper investigates the use of neural networks (NNs) to mitigate these adverse boundary effects and enable truncated domain requirements. Two distinct NN-based approaches are proposed: (1) direct reconstruction of unknown particle distribution functions at the outflow boundary; and (2) enhancement of established characteristic boundary conditions (CBCs) by dynamically tuning their parameters. The direct reconstruction model was trained on data generated from a 2D flow over a cylindrical obstruction. The drag, lift, and Strouhal number were used to test the new boundary condition. We analyzed results for various Reynolds numbers and restricted domain sizes where it demonstrated significantly improved predictions when compared with the traditional Zou & He boundary condition. To examine the robustness of the NN-based reconstruction, the same condition was applied to the simulation of a NACA0012 airfoil, again providing accurate aerodynamic performance predictions. The neural-enhanced CBC were evaluated on a 2D convected vortex benchmark and showed superior performance in minimizing density errors compared to CBCs with fixed parameters. These findings highlight the potential of NN-integrated boundary conditions to improve accuracy and reduce computational expense of aerodynamic and acoustic emissions simulations with the LBM
The repository is organized to separate code, data, models, and figures for clarity and reproducibility.
This is a template repository for a overlay repository for TCLB Solver. Kind of a layer you can put on top of the TCLB's repository to allow you to develop and share your models or cases in your own repo, without the "baggage" of the main repo.
After you create your own repository from this template you can:
git clone git@github.com:your_user/your_repo.git project
cd project
./update.shThe update script.sh will checkout all the files of the TCLB solver and the two repositories (TCLB, and your overlay) will share the same directory. You can operate on your repository with the standard git commands with:
cp examples/flow/2d/karman.xml ./
## edit karman.xml with your favorite editor
git add karman.xml
git commit -m "My karman case"
git pushYou can select the repo and branch with the --tclb url@branch option:
./update.sh --tclb git@github.com/my_user/TCLB.git # the my_user's fork
./update.sh --tclb git@github.com/my_user/TCLB.git@develop # develop branch from my_user's fork
./update.sh --tclb @develop # develop branch from the main repositoryYou can create a file .overlay.config with the default TCLB fork and branch to use when ./update.sh is run:
> cat .overlay.config
TCLB_FORK='my_user/TCLB'
TCLB_BRANCH='develop'
> git add .overlay.config
> git commit -m "Adding default TCLB repo and branch"You can also use the commandline option from the previous section with --save to create this file.
And can operate on the TCLB repository with git tclb command:
git tclb remote add myfork git@github.com:my_user/TCLB.git
git tclb fetch myfork
git tclb switch -c fix/stuff
## fixing stuff in TCLB
git tclb commit -a -m "I fixed stuff"
git tclb push myfork fix/stuff
## Open pull request at:
## https://github.com/my_user/TCLB/pull/new/fix/stuff