Effect of Stairstepping on numerical results

  • Last Post 13 March 2019
CatiaV5 posted this 07 March 2019

Hi all,

I was wondering what impact it has on the numerical results if stair stepping is occurring in the inflation layer. Especially in the area with the reduced area. Will it have a significant impact on the results

Attached Files

Order By: Standard | Newest | Votes
rwoolhou posted this 07 March 2019

That will depend on what is happening in that region, and what the cell quality is. In itself stair-stepping simply means you've got fewer layers in the inflation zone. 

Please can you post the image in the text: staff are not permitted to open attachments. 

CatiaV5 posted this 07 March 2019

The question was just general as I see the warning messages and did wonder. So would you say it is better to have a lot of inflation layers with stair stepping or less inflation layers but no stair stepping.

Also I found a paper where Anysy Icem was used for the tet mesh generation with inflation layers. It stated that the 1st cell height is 1e-7m with the total extent of the 36 prismatic layers of 0.003m resulting in the non-uniform exponential expansion ratios normal to the wall of 1.21 near the wall and 1.35 at the edge of the prismatic layers.

I do not understand why two growth rates are mentioned. I thought the growth rate in ICEM was constant. Do you by any chance know what this refers to? Or shall I open up another post?

rwoolhou posted this 08 March 2019

In your case I'd question why the inflation stops at that position: did you miss an edge in the selection? Otherwise it's still possibly OK depending on what the flow is doing in that area. 

I suspect the two growth rates are for the inflation region and then in the tet region: but as I don't use ICEM CFD (I used to along time ago before tets and automatic blocking) I'm not sure. We're also assuming the paper terminology is correct! 

kkanade posted this 11 March 2019

do you have image of mesh from paper?

CatiaV5 posted this 11 March 2019


yeah I do (see attached))

kkanade posted this 12 March 2019

so it is continuous prism layer. it does not have stairstepping. 

it is recommended to avoid stairstep in such geometries. 

so please check if you have missed any face in inflation definition in your case. 

kkanade posted this 12 March 2019


CatiaV5 posted this 12 March 2019

If you look at the profile it does not look continuous to me though. Or are those bumps not due to stairstepping?

Is there a rule btw how to avoid stairstepping? I figured that if I decrease my face sizing and do not vhange the first layer height I will get stairstepping. I haven’t figured out yet what I will need to change to avoid this. Because when decreasing the 1st layer height I will need to increase the number of layers which then results often in stair stepping again.

kkanade posted this 13 March 2019

from the first image you have posted, i think you have missed some faces to define inflation. so you are getting stairstepping. 

please cross check.