[Crank-Shaft]

Hello CFDOnline,

Hope you are all going well with your individual CFD challenges.

I am currently involved in the development and use of a research CFD code which relies on a compressible flow solver on multiblock, structured grids with a curvilinear coordinate system. This requires a CGNS file input and I have access to ICEM for the meshing and BC definitions.

The geometry in question is a flow over a periodic hill from NASA Turbulence Model Validation, which has also been mentioned in this previous thread for OpenFOAM. Thread

Unfortunately, I have not been able to convert the geometry and the grid files provided as .dat files. Hence, I had to resort to recreating the geometry in Solidworks and then attempting to mesh using ICEM. For the purposes of validation of our solver, it is crucial that identical grids be used and hence, I would like to know whether such conversion processes exist.

I look forward to your suggestions. If it is not possible to convert this into a format readable by ICEM, then please provide some guidance on the blocking and meshing strategy. I have tried to block this previously however, I faced problems with the splines around the hills.

Thank you.

[Crank-Shaft]

Has anyone encountered this geometry before?

I still haven't been able to correctly make the end blocks
follow the curvature of the hills.

I look forward to any suggestions you have.

[Crank-Shaft]

Hello CFDOnline,

I had a question regarding the periodic hill domain outlined above. Having made numerous attempts at blocking the curved hill geometry at the inflow and the outflow, I just couldn't manage to replicate the Tecplot 2D grid shown in the attached image.

Can someone please help with the blocking of the two zones at the end? I am not sure how to make the edges follow the curvature and even though I associated the edges to the curves, it doesn't appear to do it correctly.

Thanks in advance.

[bluebase]

Hi Crank-Shaft,

try the following to correct your associations.

Associate the "hanging" vertices to the corner points of your bottum part. And then reassociate the edges of the hill zone to the hill curves. Have a look at the attached picture.

Periodic_Hill_Geom2-NoEqn.jpg

[diamondx]

You associations were not good, did you fix the problem, let me know if not, I will upload my project for you

[Crank-Shaft]

Thanks for your comments everyone.

Bluebase, your blocking looks great. I will need to try this some time soon.

I have tried to break this into several blocks at each end of the domain. This was the only way I could manage to get a robust, readable mesh without a lot of errors.

Here are some of the screenshots from my last attempt.
The first mesh which follows the curvature of the hill was made using ANSYS workbench mesher and the second mesh was generated using ICEM CFD with 7 different blocks which is not elegant. ​NASA_LARC_PeriodicHill

It would like to be able to do this with just 3 blocks (similar to Bluebase's suggestion) since that would make it much more elegant and easier for grid refinement studies.

Thanks

[flotus1]

Maybe I got the whole thing wrong, but I managed to get a decently looking mesh with one block only.
I just had to associate all 8 vertices of the block aswell as the lower edges to the curved walls.
Geometric distribution in wall-normal direction, uniform distribution in streamwise direction.

The only difference from the tecplot image you posted seems to be that the grid lines are not perpendicular to the wall.

periodic_hill.jpg

[bluebase]

There is a smoothing tool which allows to improve orthogonality and the perpendicularity of elements to the first layer on surfaces.

You can find this feature under: Blocking / Pre-Mesh Smooth
Change the Method to orthogonality.
Probably you need to play with the settings a bit. My suggestion is to increase the grid expansion rate to 2 and maybe increse the on-surface-iteration.

Hope this tip helps =)

Though, sometimes, the smoothing tool is a bit tricky. In complex grids it often mess my grids up.

[flotus1]

Originally Posted by bluebase

the smoothing tool is a bit tricky

Absolutely. I would have recommended the smoothing tool in the previous post if I had been able to produce a better mesh with it. The tool needs a more experienced user.