[KateEisenhower]

Hello everyone,

I have a question concerning boundary layer creation in cfMesh. If Franjo, Tomislav or another developer is reading this, thank you for creating this tool! I worked with it for two days now and I enjoyed it pretty much.
Let me cite the official User Guide first:

In addition, their thickness is controlled by the cell size specified at the boundary and the mesher tends to produce layers of similar thickness to the cell size.

This corresponds with what I have experienced so far.
If I have a small cell size, the total boundary layer thickness is too small for my application.
If I go with a rather coarse grid at the boundary, I manage to get a big enough total boundary layer thickness. So far, so good, the problem is that I need around 25 layers to dissolve my boundary layer sufficiently. Then I get a thickness ratio (first cell of the farfield/last cell of the boundary layer) of at least 5 (please have a look at the attached picture). Please correct me if I'm wrong but I think for a proper mesh, the ratio should be between 1 and 1.5.
Is there any way to tweak this behaviour or is cfMesh simply not fit for my usecase?

Best regards,

Kate

[franjo_j]

Hello Kate,

Thank you for your feedback.

Originally Posted by KateEisenhower

Hello everyone,

I have a question concerning boundary layer creation in cfMesh. If Franjo, Tomislav or another developer is reading this, thank you for creating this tool! I worked with it for two days now and I enjoyed it pretty much.
Let me cite the official User Guide first:

This corresponds with what I have experienced so far.
If I have a small cell size, the total boundary layer thickness is too small for my application.
If I go with a rather coarse grid at the boundary, I manage to get a big enough total boundary layer thickness. So far, so good, the problem is that I need around 25 layers to dissolve my boundary layer sufficiently. Then I get a thickness ratio (first cell of the farfield/last cell of the boundary layer) of at least 5 (please have a look at the attached picture). Please correct me if I'm wrong but I think for a proper mesh, the ratio should be between 1 and 1.5.
Is there any way to tweak this behaviour or is cfMesh simply not fit for my usecase?

You can try to work around the problem by adjusting the number of layers and the thickness ratio. Currently, there exists no explicit control of the ratio between the last cell of the layer and the first far-field cell. This ratio is not a problem if there is no significant gradient variation there.

Originally Posted by KateEisenhower

Best regards,

Kate

Regards,

Franjo

[KateEisenhower]

Hello Franjo and thanks for the answer,

Originally Posted by franjo_j

Hello Kate,

Thank you for your feedback.



You can try to work around the problem by adjusting the number of layers and the thickness ratio. Currently, there exists no explicit control of the ratio between the last cell of the layer and the first far-field cell. This ratio is not a problem if there is no significant gradient variation there.



Regards,

Franjo

I tried this. I don't know if the results are sufficient for my application but I'll just give it a try with checkMesh and simpleFoam.
However I have two more problems:

1) When I go for a rather big layer, the layers go up und down along the boundary (boundaryLayer_waves.png).
2) The mesh around the trailing edge doesn't look good at all. I tried with a cut trailing edge (trailingedge_cut.png) and an uncut version (trailingedge_uncut.png). As recommended in the guide, I already lowered the cell size in this region below the feature size.

Maybe you have some advice for me.

Kind regards,

Kate

[franjo_j]

Hello Kate,

Originally Posted by KateEisenhower

1) When I go for a rather big layer, the layers go up und down along the boundary (boundaryLayer_waves.png).

The layer goes up and down because the goal is to generate a mesh without any inverted elements, and this requirement could not be satisfied when the layer has constant thickness.

Originally Posted by KateEisenhower

2) The mesh around the trailing edge doesn't look good at all. I tried with a cut trailing edge (trailingedge_cut.png) and an uncut version (trailingedge_uncut.png). As recommended in the guide, I already lowered the cell size in this region below the feature size.
Kate

I am of the impression that you have not defined feature edges or patches, which enforce the mesher to capture the trailing edge. Does the situation get better with mesh refinement?

Regards,

Franjo

[KateEisenhower]

Hello Franjo,

Originally Posted by franjo_j

Hello Kate,


The layer goes up and down because the goal is to generate a mesh without any inverted elements, and this requirement could not be satisfied when the layer has constant thickness.

Can you explain in detail what you mean with inverted elements in this context?

Originally Posted by franjo_j

I am of the impression that you have not defined feature edges or patches, which enforce the mesher to capture the trailing edge. Does the situation get better with mesh refinement?

Regards,

Franjo

Yes the situation was a lot worse with a coarser mesh. I defined my patches within the *.stl file. Now I added one for the trailing edge and it now gets dissolved properly as you can see in the picture. That was very helpful. Thank you for that.

Best regards,

Kate

[franjo_j]

Hello Kate,

Originally Posted by KateEisenhower

Hello Franjo,

Can you explain in detail what you mean with inverted elements in this context?

All cells shall be convex. There shall be no concave edges in any cell.

Regards,

Franjo

[francesco.giacobone]

Hello Franjo,

I'm currently using cf-mesh and I have experienced the same problem as Kate regarding the total thickness of the inflation layers from the surface. Is there now any method to control this thickness or is still automatically set by the cell size at the farfield?

Thanks,

Francesco

[craven.brent]

I realize this is a somewhat dated thread, but I am also trying to increase the thickness of the near-wall boundary layer mesh from cfmesh. I realize there is not an option for this in the meshDict in the open-source version. But, similar to Kate, I am trying to create a coarse mesh at the boundary, which can then hopefully be used to increase the total thickness of the boundary layer mesh.

But, I have not been able to create a coarser boundary mesh because "maxCellSize" always restricts the size of the boundary cells. Has anyone come up with a way to go about doing this: create a slightly coarser boundary mesh than than the interior so that the total boundary layer mesh thickness will be larger?

Thanks!

Brent

[Gang Wang]

Originally Posted by craven.brent

I realize this is a somewhat dated thread, but I am also trying to increase the thickness of the near-wall boundary layer mesh from cfmesh. I realize there is not an option for this in the meshDict in the open-source version. But, similar to Kate, I am trying to create a coarse mesh at the boundary, which can then hopefully be used to increase the total thickness of the boundary layer mesh.

But, I have not been able to create a coarser boundary mesh because "maxCellSize" always restricts the size of the boundary cells. Has anyone come up with a way to go about doing this: create a slightly coarser boundary mesh than than the interior so that the total boundary layer mesh thickness will be larger?

Thanks!

Brent

Hi!

I also have this problem when using cfmesh to draw the boundary layer mesh. Do you have any solved plans?

Best,
Gang

[Harshal.Suthar]

Hello guys
I am trying to generate a good boundary layer around my rocket model. I want to do a 2d simulation so using the cartesian2Dmesh command. As per the inlet velocity and the length of the rocket my first cell height comes out to be 1.95e-4 with a gradient of 1.4.

but when I am trying to give those values to the meshdict it creates the mesh but during the check mesh it fails all 6 checks. ex. negatives cells are very highly skewed cells and the aspect ratio is in the order of e+295. which is insane. can anyone please guide me where I am making mistake during boundary layer generation