How to control the prism layer thickness in scFLOW?

Hi Rodrigo, thanks for your question on prism layer insertion.

I understand what you're saying about the 1st prism layer getting smaller when you increase the number of layers.  And I can see how this is not what you expected.  The reason is that it's actually the outermost prism layer that has its size set to the fixed fraction of the local octant size.  - With the default prism insertion settings, the size of the outermost layer is calculated first, then the sizes of the lower layers are calculated one at a time, successively applying the [Variation rate of thickness] setting (default value = 1.1) to reduce the prism thickness as we move towards the wall.

So, if you have 2 prism layers, the 1st prism layer (the prism layer on the wall), should be 1/1.1 times the size of the local octant.  But if you specify 5 prism layers, the 1st prism layer thickness will be reduced to 1/(1.1^4) of the octant size.  - Going from 2 layers to 5 layers, that overall ratio is reduced from 0.909 to 0.683, meaning that the thickness of the 1st prim layer will be noticeably smaller.

You said that your goal was to increase the total thickness of prism layer coverage.  Given the simple example above, just increasing the number of prism layers isn't going to give you a straightforward way to control the total coverage, especially if you want to maintain the same 1st prism layer thickness when you vary the number of layers. - You probably want to keep that "y1" size, to maintain the same y+ value on the wall, while increasing the total coverage to take in more of the boundary layer.

A more effective way to increase the total prism coverage will be to deselect the checkbox for [Calculate thickness automatically from octant size] and then manually set [Thickness of 1st layer] = 0.2 times the local octant size.  With this approach, the size of each prism layer will then be successively calculated from the innermost layer (layer-1) to the outermost layer.  - The opposite order to the default settings order.  When you increase the number of prisms, you will see a much larger (and more easily predictable) increase in their total coverage.  Here's a screenshot showing this modified type of prism layer setting:  (This assumes a local octant size of 0.01 [m] and [Thickness coefficient] = 0.2))

Your question is actually quite timely as Cradle CFD 2025.1 has just been released and the scFLOW Preprocessor Reference manual now includes a more detailed explanation of how the prism layer thicknesses are calculated.  You can find that in this section of the manual: [Condition] - [Mesh Parameter] - [Parameters for Prism Layer Insertion] - (Supplement 2).

I hope my comments are helpful.  Please let me know if you have any follow-up questions.

Hello David,

Thank you for your detailed explanation. Your explanation of how the prism layer thickness is controlled makes a lot of sense. As you say, since I am interested in controlling the total thickness height, I will specify the first layer thickness and ensure that there are enough layers to cover the total height.

I have a follow up question related to the [Timing of Prism Layer Insertion] setting. Does this play a role in either the first layer height or the total thickness?

I see that there are 3 options: Before volume meshing, after volume meshing, and after polyhedral conversion – how do these affect the outcome of the mesh, and do you have recommendations to pick one over the other?

Hello David,

Thank you for your detailed explanation. Your explanation of how the prism layer thickness is controlled makes a lot of sense. As you say, since I am interested in controlling the total thickness height, I will specify the first layer thickness and ensure that there are enough layers to cover the total height.

I have a follow up question related to the [Timing of Prism Layer Insertion] setting. Does this play a role in either the first layer height or the total thickness?

I see that there are 3 options: Before volume meshing, after volume meshing, and after polyhedral conversion – how do these affect the outcome of the mesh, and do you have recommendations to pick one over the other?