CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   ANSYS Meshing & Geometry (http://www.cfd-online.com/Forums/ansys-meshing/)
-   -   [ANSYS Meshing] Problem with specifying first element size in 2D Mesh (http://www.cfd-online.com/Forums/ansys-meshing/109240-problem-specifying-first-element-size-2d-mesh.html)

 sigmundab November 13, 2012 08:04

Problem with specifying first element size in 2D Mesh

1 Attachment(s)
Hi

I'm trying to model turbulent flow in an axis symetric pipe using fluent. This requires high control of the first node from the wall (y ~10^-4). For the time being I have accomplished this using sizing with bias on the edges (in Ansys Meshing). When using this approach, I do not feel I'm in control of the first node, regardless of choosing element size or numbers of divisions.

I would like to specify first node size and a growth factor, is there a way to do that?

I have included an example mesh below.

Regards
Sigmund

 bansa9643 November 13, 2012 09:53

Hello Sigmund,

I have a similar problem.. I would like to control the first element size (i.e. boundary layer size) once bias has been used in Ansys Meshing for edge sizes. (in case of a structured mesh)

When face sizing is used, one can use a boundary layer(inflation) option where one can control the first element size. But in this case, the resulting mesh is unstructured!

Any suggestions where one can control the boundary layer element size while ensuring a structured mesh would be very helpful!

bansa9643

 sigmundab November 18, 2012 13:40

Nice to see that there are more people out there with similar problems. Hopefully there is someone out there with the answer to our problem as well. The floor is open :)

 diamondx November 18, 2012 14:52

i had never used ansys meshing to meshing a 2d case, i always use icem. the feature that can allow you to do so is by applying inflation. try this, select that edge, right button click, insert inflation. While setting up the parameter for inflation you can give it the size of the first element. i do this with surface in 3d, i guess it will work in 2d two.
I hope it can help

 sigmundab November 19, 2012 06:17

Thanks for the response Ali.

The inflation function works fine, the problem is what happens beyond the inflated area. I want that area to be structured, but it ends up being unstructured.

What I thought would be the solution, was to combine inflation with size functions and/or mapped face meshing. With size functions it was still unstructured and the combination of inflation and "mapped face meshing" is invalid (at least with the default settings).

The only way I've been able to create a structured mesh is by using sizing functions with bias (in combo with mapped face surface), but then again I cannot get the accurate first element size close to the wall.

Should I try a different meshing program, or do anyone know how to solve this in Ansys Meshing?

Regards
Sigmund

 diamondx November 19, 2012 10:36

as i never tried, it is hard to believe that it's impossible to do it with ansys meshing. What i can suggest you is ICEM CFD. very easy to use with 2d geometries and you have total control over your fist cell height ... we can help you with icem cfd more than ansys meshing. can you post some screenshot of your geometry ?

 sigmundab November 21, 2012 10:57

Not exectly a solution...

Today I received an answer from Ansys support center. They could inform me that in Ansys meshing one can not use inflation and get a structured grid (in 2D). The only solution is to use sizing with bias and mapped face meshing, and a great amount of trial and error to get desired result.

Sigmund

 diamondx November 21, 2012 11:32

Quote:
 Originally Posted by sigmundab (Post 393453) Today I received an answer from Ansys support center. They could inform me that in Ansys meshing one can not use inflation and get a structured grid (in 2D). The only solution is to use sizing with bias and mapped face meshing, and a great amount of trial and error to get desired result. Sigmund
thanks you for sharing this...

 Crank-Shaft November 25, 2012 02:29

I have thought about this problem in the past as well however, did not investigate any further and simply found that the y+ values were suitable for my analysis by SOME trial and error.

After having learnt about the definitions themselves however, I believe you should be able to do some simple, manual calculations and implement this using the Element Size option. Based on limited knowledge, you could firstly specify the largest cell size you want in the wall-normal direction and then using the definition of bias factor as simply a length ratio between the largest and smallest cells to calculate the required value.

I guess this technique is not exactly accurate but should save you a lot of iterations compared to a complete trial and error approach.

 All times are GMT -4. The time now is 06:59.