[Gang Wang]

Hi formers!

I'm using SST-IDDES model in OpenFOAM to simulate the flow around circular cylinder at Re=3900, quite typical topics for the benchmark test of your code. I went through all the materials and tutorials about OF, in divSchemes, if you want to use LUST scheme, it would be two choice, Gauss LUST grad(U) and Gauss LUST unlimitedGrad(U). Another point I noticed is in the pisoFoam, LES, motorbike tutorials, they used snappyHexMesh to draw the hex mesh, and it's unstructured, and they used Gauss LUST unlimitedGrad(U), I don't know how to choose Gauss LUST grad(U) and Gauss LUST unlimitedGrad(U).

Does anyone know the difference of these two schemes? unlimited and no unlimited, Do you choose that on the basis of grid shape (Hex and Tet?), structured grid and unstructured grid?

Best regards,
Gang

[syavash]

Originally Posted by Gang Wang

Hi formers!

I'm using SST-IDDES model in OpenFOAM to simulate the flow around circular cylinder at Re=3900, quite typical topics for the benchmark test of your code. I went through all the materials and tutorials about OF, in divSchemes, if you want to use LUST scheme, it would be two choice, Gauss LUST grad(U) and Gauss LUST unlimitedGrad(U). Another point I noticed is in the pisoFoam, LES, motorbike tutorials, they used snappyHexMesh to draw the hex mesh, and it's unstructured, and they used Gauss LUST unlimitedGrad(U), I don't know how to choose Gauss LUST grad(U) and Gauss LUST unlimitedGrad(U).

Does anyone know the difference of these two schemes? unlimited and no unlimited, Do you choose that on the basis of grid shape (Hex and Tet?), structured grid and unstructured grid?

Best regards,
Gang

Hi Gang,

Did you get descent results with kOmegaSST-IDDES? I tried this model for an internal flow problem but all I get is damped solution similar to RANS. Instead I used SA-IDDES and it looks good in terms of flow oscillation and turbulent content. I wonder if any adjustment is required for the kOmegaSST-IDDES to obtain proper turbulent flow. I am using OF-v2012 FYI.

Kind regards,
syavash

[Gang Wang]

Originally Posted by syavash

Hi Gang,

Did you get descent results with kOmegaSST-IDDES? I tried this model for an internal flow problem but all I get is damped solution similar to RANS. Instead I used SA-IDDES and it looks good in terms of flow oscillation and turbulent content. I wonder if any adjustment is required for the kOmegaSST-IDDES to obtain proper turbulent flow. I am using OF-v2012 FYI.

Kind regards,
syavash

Hi,

I think it might be many factors contributing to the errors, but I suppose closing to RANS might be attributed to your mesh, as a LES based simulation, your mesh needs to be very fine. Otherwise the modelled turbulent kinetic energy would be in a large proportion, instead of resolved.

Hope it helps

Best regards,
Gang Wang

[Gang Wang]

Originally Posted by Gang Wang

Hi,

I think it might be many factors contributing to the errors, but I suppose closing to RANS might be attributed to your mesh, as a LES based simulation, your mesh needs to be very fine. Otherwise the modelled turbulent kinetic energy would be in a large proportion, instead of resolved.

Hope it helps

Best regards,
Gang Wang

Well until now I still didn't know what about the difference between unlimitedGard(U) and grad(U). If anyone has ideas, feel free to discuss together.

Best,
Gang