Boundary layer not forming

Hi all,

Recently, I’ve been trying to mesh a simple aircraft for transsonic simulation (500 mph at 40,000 feet). I’ve done the Y+ calculations (I’m using wall function, so targeting a Y+ of around 140), and I’ve entered in all the relevant information into the “inflate boundary layer” refinement. However, I’ve found that the boundary layer only forms for part of the wing, and is super jagged for the rest of the wing. “Layer size” has already been turned off.

I’ve seen some earlier forum posts that tell me to disable the min volume check, but I’ve tried that and it has no effect. I’ve checked the mesh logs and it appears that the cells are being generated, but then tend to be removed. Most of the problem cells are classified as such becomes of nonorthogonality or face twist, however, I’m hesitant to disable those checks in quality control. Anyone has any ideas?

Screenshot of cells being deleted:
image

Screenshot showing missing BL:


image

Thanks!

Hello 23,

Glancing at your project, one thing to possibly try is to combine the two inflate boundary layer refinements you have. (Wing BL and Fuselage BL). When running them separately they may be interfering with one another. Combining them into one boundary layer refinement should provide a better cell mesh in that region where those two boundary layers would be interacting. Let me know what you think and if you have any success.
Best,
Jake

1 Like

Hi Jake,
Thanks for the reply. I tried it, no effect. Boundary layer still doesn’t form. Mesh log shows that the full boundary layer is being generated initially, but most of the boundary layer over the wing is later being deleted for mesh quality reasons. I’ve played with those parameters and none of them have done anything, though.

image

Hey! @23ax0762 your mesh is looking a lot better.

Boundary layers can be quite annoying i know. From your last post i had asked if you were going to try and simulate for viscous sub-layer or log law. Why for such a high speed aircraft are you choosing log-law? are you doing a stall angle study?

If you are using 11 layers for the boundary layer anyways you could try to to get your Y+ under 1 and get more accurate data as its more tailored for shear force measurement.
image

As for the BL inflation problem there are a couple of things i noticed that could or could not be one of the problems.

Normally i stick close to the default range for mesh settings, unless ive done a study of their effects. So these three could be causing some problems (maybe). So try these at default.

image

image

I think a big part of the layer deflation is the your surface refinement on the wing, Since you are using absolute layering, and 11 layers, the inflation has to push the HEX cells very far away from the geometry surface and with such small cells at the leading and trailing edges, its causing them to distort so much they become illegal.

To compensate the mesh will delete these layers so that the mesh will actually mesh. So try :

  1. make a smaller jump in cells for you surface refinement level difference. so in stead of 5 to 9, try 7 to 8 or 7 to 9. You might have noticed that the fuselage had full layering with a 4-5 level selection.

  2. You dont need 11 layers to resolve the log-law region. By the 11th layer you are probably way over 300, which isnt so bad as under 30 but can still have negative effects (like causing layer deflation).

Also the layer thickness Section of the mesh log is very informative. Here you have what seems like the fuselage section achieving 8.79 of your target 11. Which is OK but its best to hit your target layer amount.

The wing section is obviously not good with only 0.117 layers of 11 achieved. In general you want the thickness % to be as close to 100% as possible. So 84.8% for the fuselage isnt bad but can be improved.

image

LASTLY

I am currently working on a big project for BL deletion and i have found that the Max non-orthogonality setting will cause BL deletion. Try setting this Higher first to make sure this isnt effecting results then tighten the quality check after you have all layers inflated.

So try MAX at 75 or 80 and relaxed 5 deg higher at first then when you have layers, reduce the max and relaxed slowly to improve qaulity, then, when the max value is to low and causes layer deflation, go back to the previous setting and you know the lower limit.

Dan

2 Likes

Hi,
Thanks for help! I changed the y+ value to be below 1 to simulate the viscous sublayer, and applied the rest of your fixes. At first the layer didn’t show up at all (probably cuz the layer got a lot thinner), but after playing with mesh quality I managed to get a few layers back. The highest layers I managed to get was 5 layers, and some faces had less than 1 layer (on average). Maybe adding a distance refinement to the wing, something like refine the nearest 0.01 meters to level 7 would help?

Yes a distance refinement would help… and you can slowly increase mesh size away from the wing to fit nicely with whatever level you have for the cartesian box surrounding the plane.

Hi,
From a bit of experimentation, I’ve found that distance refinement does indeed help! If I refine the area around the body sufficiently such that the final layer thickness is somewhat comparable to the local refinement cell size. I found 1 set of values that can generate 7 layers - I have that configuration saved, and I’m looking for better ones. Thanks for the help!

The mesh cell count is really high (>40 million), but from what I’ve found online, it’s difficult to get those layers to form without inflating the cell count, so I guess I can live with it.

Thanks,
23