[Bal21]

Hi,

The Y+ is <1, but I want to achieve a smooth transition between prismatic and core mesh layers.. see pictures..
Which parameters should I be adjusting? Thanks!

[LuckyTran]

Increase the overall prism layer thickness and increase the stretch ratio. If both of those sound like things you don't wan to do, it's because they are. Because the third thing that you also don't want to do is make the core mesh finer.

[Bal21]

But I've not defined a stretch ratio because I've selected the distribution mode as prism layer near wall thickness

[cwl]

It is a geometric progression anyway (unless you've selected hyperbolic distribution) - and you set two out of three parameters defining it.

There is also a Near Core Layer Aspect Ratio setting, but that is more like for cells of varying size.

[Bal21]

I have selected geometric progression and wallet thickness but kept everything else according to default values.. see pictures

I tried changing near core mesh aspect ratio but it didn't have effect

[cwl]

What are the desired near wall prism layer thickness, near core (farthest) prism layer thickness and amount of prism layers?

[Bal21]

Thank you. I want to have a Y+<1
So I tested a near wall thickness 5e-4mm, 15 prismatic layers and prism layer total thickness of 0.07 mm and they gave the required Y+ value. The base size is 0.75 mm, and min surface size is 0.0375 mm

I want to also optimise the number of cells as much as possible.

[cwl]

So - what is the expected thickness of the outer prism layer (away from the wall)?

[LuckyTran]

I woke up today and realized I forgot to mention you can also reduce the number of prism layers. So that's another thing you can do.


It's just a geometric series. The 1st layer thickness times the stretch ratio gives you the thickness of each layer. Sum them up and they have to add to the total thickness. You choose which 2 of the 3 settings you want to apply and the 3rd is constrained by this simple algebraic formula.


I know it's very frustrating not having the mesh you desire but the rules for meshing are really that simple. There is no complicated optimization strategy. It is blatantly clear what the last cell height is going to be.


fyi if your total prism layer thickness is less than a tenth the base cell size, that means the largest prism possible is smaller than the core mesh by a factor of 10. You'll almost certainly never get a gradual, smooth transition.

[Bal21]

Thank you,

the problem is that that the near core mesh appears to be of varying size, so the size of some near core cells might be comparable to the last prismatic layer while others might be larger.. I hope that I clarified my concern.

I tried to implement your recommendations (see attached image), does this show a smooth transition between prismatic and core meshes (your opinion is appreciated)?

[cwl]

Looks perfect to me.
Except maybe for a bit coarse mesh on the circles - but that completely up to you.

[Bal21]

Thanks.
I tried to adjust the curvature of the circles using number of points per circle and sat it to 80 (it was 50 in the above image) but the curvature didn't improve. Then I tried to reduce the min surface size to gradually up until 0.01 mm where it captured the circle smoothly, but this is horrible in terms of the number of cells. Any other suggestion to capture the curvature of the circle without getting an insane number of cells?

[Bal21]

Thanks.
I tried to adjust the curvature of the circles using number of points per circle and sat it to 80 (it was 50 in the above image) but the curvature didn't improve. Then I tried to reduce the min surface size to gradually up until 0.01 mm where it captured the circle smoothly, but this is horrible in terms of the number of cells. Any other suggestion to capture the curvature of the circle without getting an insane number of cells?

[Bal21]

Any suggestions?