CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   ANSYS Meshing & Geometry (http://www.cfd-online.com/Forums/ansys-meshing/)
-   -   [ANSYS Meshing] How fine must a mesh be for flow over an airfoil (http://www.cfd-online.com/Forums/ansys-meshing/101427-how-fine-must-mesh-flow-over-airfoil.html)

 nottern May 3, 2012 16:55

How fine must a mesh be for flow over an airfoil

Hi,
I've been working in a 3D simulation about a crossflow water turbine. Due to restricted RAM resources, my mesh was limited to a specific size.
I ran some 2D cases and with finer meshes there are still changes.

I'd be pleased to find some references about how fine a mesh might be when modelling flow over airfoils (even if in this case are hydrofoils)

 PSYMN May 4, 2012 12:54

There are too many variables for a rule of thumb... (Although if someone has some ideas, please share)

What you should do is a mesh refinement study. Create several meshes, each one finer than the next, run them all and see if you start to get the same answer between any two meshes...

For instance, if you have several meshes where 1 is the most coarse and 5 is the most fine, and if 4 and 5 agree on the answer (and are correct), then you know that you only need to refine up to the level you did in mesh 4. Any extra refinement (to get up to the level of mesh 5) was not helpful.

Try this refinement study on a simple model (just a wing or something) under similar operating conditions to your real model, then you can apply that ideal refinement to the real model...

 nottern May 4, 2012 13:12

I tried that in 3D

I did a mesh refinement study in 3D. It was a pain in the neck, because I work with a complicated geometry and the RAM limited my finer mesh size. But, as I wrote first, looks like then when I compared in 2D with a finer mesh (I couldn't made that kind of mesh in 3D) I just noticed it wasn't enough.

I was thinking someone may have tried to make a correlation between some adimensional numbers and number of divisions, something like that. It's pretty difficult to find the info about mesh validation in papers (I'd try do the correlation if I had time and info)

Thanks anyway :)

 All times are GMT -4. The time now is 08:12.