New submodule for contributions from the data-driven modelling SIG
Functionality to add/problem to solve
Combining flow simulations and machine learning (ML) workflows has gained a lot of traction in recent years. The special interest group for data-driven modelling is very active in this field and recently published prototypical workflows combining OpenFOAM and ML via SmartSim and SmartRedis. The accompanying journal article is available as preprint.
Target audience
The entire OpenFOAM community will benefit from this contribution since ML is a flexible tool applicable across many disciplines of flow simulations and development (e.g., turbulence modelling, modal analysis, reduced-order modelling, to name a few).
Proposal
After a short exchange with @kuti and the turbulence TC, we decided to bundle contributions in a dedicated community repository to mimic the two-tier system suggested by the turbulence TC, i.e., a repository to bundle general community contributions as links (tier-1) and actively maintained contributions as submodules (tier-2). To create a submodule of submodules, we propose to add the community repository as a submodule to the official OpenFOAM repository.
Links / references
Funding
We do not have a dedicated funding source, but we receive active support and resources from industry, e.g., HPE, AWS, Intel, and Nvidia.
Best regards,
Andre W. on behalf of the data-driven modelling SIG