CFD Online Discussion Forums

CFD Online Discussion Forums (
-   ANSYS Meshing & Geometry (
-   -   Prism mesh away from the flow domain (

saisanthoshm88 October 10, 2010 05:59

Prism mesh away from the flow domain
Hello every body , im a novice user to ICEM CFD 12.1. I found an option called Extrude into orphan region which shall facilitate the growth of prisms away from the flow domain. Can some one give me an example where in this option will be actually useful.

saisanthoshm88 October 19, 2010 02:05

Multiple Prism Layers
Hello every body , I have another question. I see that the mesh quality goes down when a single prism layer is split into multiple layers is this usually expected.

saisanthoshm88 October 19, 2010 07:05

Number of Prism layers required
I found some formulae in the icem cfd tech notes at , to compute the first prism layer height is there any basis to do a little math and find out the number of layers required if I know the first layer height and the growth ratio. I'm aware of the options like Max height over base and Prism height limit factor but my intention in particular is to find the number of prism layers required by doing some math , having known the first layer height and the growth ratio some one help me out regarding this.

dst October 20, 2010 04:08


I'll try to make some hints to make the problem more narrow.
Please, feel free to correct me if you are interested in other aspects and the following information has nothing to do with your question.

Required number of prism layers depends mostly on the flow properties. If you are describing the boundary layer with prisms you need to check the turbulence model (wall function) requirements for exact turbulence model you are planning to use.

If you are simulating the boundary layer separation from the airfoil surface - it is one problem and approach can (in most of cases - will) be absolutely different if you are simulating the flow in a long tube where you have boundary layers' connection point.

In comparison to structure block meshes where all the cell structure is predictable, we have some limitations with prisms when we are doing unstructured meshes.

Difficulties in this case can occur due to geometry variations or due to tetra mesh which you need to move in order to generate the prism layer.

As for a study-case, I would suggest to try to use 3-5 prism layers - just to make it work and get feeling when it is working or when it is not.
You can estimate somehow the initial thickness of the first layer depending on the simulation.

To have a prism thickness variation with a variation of a cell size, you can limit the prism size. In general, you can use 0.1-0.3 size limitation factor. As a result, in the places where you have refined mesh with a smaller sizes, ICEM can reduce the thickness following the limitation factor. Otherwise, prism generation can be aborted when there is no possibility to move some small tetras out from the wall to such a distance (prism thickness).

If you are interested in some exact problem (simulation) - just describe it. This would allow others to give you some exact advices.

Best regards,

All times are GMT -4. The time now is 18:22.