ROM-assisted Predictors For Accelerated Convergence of Partitioned Fluid-Structure Interaction Simulations.

Online adaptive ROMs with local dynamic subspaces

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License.

This repository contains reproducible material for a new ROM-assisted predictor for partitioned FSI, using online adaptive ROMs, with local dynamic subspaces.

Online adaptive non-intrusive model reduction via manifold interpolation and subspace updates: application to FSI convergence acceleration (2026)

• Preprint / DOI: TODO, incoming ..
• Authors: Azzeddine Tiba, Florian De Vuyst, Iraj Mortazavi

Reproducibility (2026 paper)

The goal of this repository is reproducibility across multiple FSI settings.

What is required

1. A Kratos build compatible with the ROM-assisted coupling workflow used in this project:

• The new ROM-assisted predictors are implemented in a forked version of Kratos (v.0n.Adptv) from the v9.4.2 release.

2. The rom_am Python package version pinned in each example requirements.txt.
3. Example-specific Python environments (one per example is recommended).
4. Access to the files of the saved ROMs from Zenodo (see Data availability below).

Data availability (Zenodo)

The used ROMs from the paper, for the two new examples, are saved in pickle files and are hosted externally.

• Zenodo record: FSI-Online-ROM-Predictor-Data
• Zenodo DOI: https://doi.org/10.5281/zenodo.22261698
• The Zenodo repo has the same file tree as the current repo.

Reproduce the two new examples

1) lid_driven

Folder: examples/lid_driven

Install dependencies:

python -m venv .venv-lid
source .venv-lid/bin/activate
pip install -U pip
pip install -r examples/lid_driven/requirements.txt

Run workflow:

cp ini_common/* .
cp ini_testing_phase/* .
python MainKratos.py

Relevant outputs will be in the CoSimData/ folder.

2) fsi_turek

Folder: examples/fsi_turek

Install dependencies:

python -m venv .venv-turek
source .venv-turek/bin/activate
pip install -U pip
pip install -r examples/fsi_turek/requirements.txt

Run workflow:

cp ini_common/* .
cp ini_testing_phase/* .
python MainKratos.py

Relevant outputs will be in the CoSimData/ folder.

Legacy context (2024 paper and material)

The repository originally presented a similar approach, with the difference being the use of ROMs with global bases, and with the use of ROM-FOM coupling for additional acceleration:

• See [2, 3].
• The associated, older examples and notebooks remain available.
• The associated data repository is https://doi.org/10.5281/zenodo.13909576.

Reproducibility

To reproduce the results presented in [1], the Comphy tag of this repository should be used, as well as installing the Comphy tags of ROM_AM and KratosMultiphysics. See the installation instructions at each repository.

git clone --depth 1 --branch Comphy https://github.com/FsiROM/Kratos.git

git clone --depth 1 --branch Comphy https://github.com/azzeddinetiba/ROM_AM.git

git clone --depth 1 --branch Comphy https://github.com/FsiROM/FSI-ROM-Predictor.git

Data files : The results already obtained by the author, and used in [1] are available in the form of .vtk files and numpy arrays stored in .npy files. These files are not available with the current repository due to their large volume. The files can be found in this Zenodo repository. The Zenodo repository contains the data folders organized using the same structure of this repository. The user then can just replace the empty folders in this repository with the Zenodo folders.

References

[1] Azzeddine Tiba, Florian de Vuyst, Iraj Mortazavi. Online adaptive non-intrusive model reduction via manifold interpolation and subspace updates: application to FSI convergence acceleration. 2026.

[2] Azzeddine Tiba, Thibault Dairay, Florian De Vuyst, Iraj Mortazavi, Juan-Pedro Berro Ramirez (2024). Machine-Learning Enhanced Predictors for Accelerated Convergence of Partitioned Fluid-Structure Interaction Simulations. arXiv preprint arXiv:2405.09941

[3] Azzeddine Tiba, Thibault Dairay, Florian De Vuyst, Iraj Mortazavi, Juan-Pedro Berro Ramirez, Non-intrusive reduced order models for partitioned fluid-structure interactions, Journal of Fluids and Structures, Volume 128, 2024, 104156, ISSN 0889-9746.