[Fool]

Hi together,

I've just started to learn about CFD simulations, especially with OpenFOAM and I asked myself what happens if I use a turbulence module like simpleFOAM for a geometry where the flow isn't turbulent at every time. I'll probably get shitty results but how do I avoid it?

The case is that I have to simulate the distribution of water in a retention filter basin. The water comes to the filter basin by a special building where turbulent conditions exists but as soon as the water leaves this building and flows into the basin the conditions get laminar because of the larger area in that basin. Has anyone an idea how to handle situations like this?

Thanks
David

[RobertHB]

The turbulence model and your solver, simpleFoam, are two different things. Using a certain solver, says nothing about whether you simulate turbulence or not, and more about if you are able to resolve the spacial and temporal scales of your flow.
simpleFoam is a steady-state solver, which is perfectly fine for flows which do not change over time, e.g. a laminar one, or large vortexes that achieve a steady motion. If you want to resolve the turbulences of your flow dependent on time, you might want to use a transient solver like pisoFoam or pimpleFoam. Solvers are chosen in the system/fvSchemes dictionary.

Each of the three solvers can use a turbulence model to calculate key variables like the turbulent kinetic enery k or the turbulence dissipation epsilon. This is where turbulence models come into play. Turbulence models are chose in the constant/turbulenceProperties dictionary and they require you to solver for different variables dependent on the model used.

Originally Posted by Fool

I asked myself what happens if I use a turbulence module like simpleFOAM for a geometry where the flow isn't turbulent at every time.

You can certainly run a simpleFoam simulation with a RANS turbulence model. No harm will come from solving a laminar flow with a turbulence model. But if you allready know that your flow will be laminar, why do so? You will solve for variables you don't need and will waste time doing so.

I'll probably get shitty results but how do I avoid it?

Not if your flow is laminar and you use a turbulence model. But if your flow is turbulent and you do not use a turbulence model you will loose important information of your flow and, most likely, not see any turbulence.

Has anyone an idea how to handle situations like this?

Transition of trubulent to laminar flow? The quickest solution: simpleFoam + RANS turbulence model. To get a first idea of your flow conditions, maybe its even enough to resolve some turbulence, but don't count on it.
Then use a pimpleFoam + RANS turbulence model. Map the velocity field, pressure and turbulence variables of your simpleFoam simulation to get a head start. Use an adjustable timestep to keep the maximum Courant number below 4.0. Give it alot of simulation runtime. Write timesteps regularily to check how and if your simulation changes from the simpleFoam solution.

[Fool]

Thanks a lot!