CFD Online Discussion Forums

CFD Online Discussion Forums (
-   OpenFOAM (
-   -   Dissimilar meshes with chtMultiRegionFoam (

crmccreary September 19, 2011 21:51

Dissimilar meshes with chtMultiRegionFoam
I've been using chtMultiRegionFoam with good success for some time now. To date, all of the cases I've solved have been "consistent" meshes in that the solid/fluid interface patches share the same points. Is there a mechanism in OpenFOAM in which I can use completely different meshes for the solid and the fluid in which no points are shared?

My objective is to have a rather coarse mesh for the solid and a very fine mesh for the surrounding fluid.

mabinty September 21, 2011 03:03

thats a very good question I d be also interested in! what I m currently trying is to refine the base mesh where later on the individual regions should be placed in order to achieve different mesh sizes in the regions.


vkrastev September 21, 2011 08:34

The chtMultiRegionFoam (and chtMultiRegionSimpleFoam as well) solver implemented in the OpenCFD OpenFOAM releases (at least till the 1.7.1 one) needs a consistent mesh at the separation surface between different regions. By the way, I saw recently a presentation from Prof. Jasak in which he explained quite clearly that the GGI (Generic Grid Interface) implemented in the -dev or -ext OpenFOAM releases is (in principle) capable of handling any type of flux exchange between two adiacent non-conformal interfaces: thus, if you are interested in this topic, I can advice you to deeply investigate the capabilities of the OpenFOAM-1.6-ext release (personally I havent't had sufficient time to do it in the last period, so I will not be able to give you any practical further support about this matter).



chandramurthy September 21, 2011 15:02

It must be possible with ggi feature of 1.6-ext. The non-conformal fluid-solid coupled BC need to be conservative and at-least 2nd order accurate to get better results. I think it can work out, if you can write a new BC inheriting the classes of ggi and coupled bc of chtmultiregionfoam. It appears that the present implementation of ggi in 1.6-ext is conservative due to face-cutting method. I think it is possible to implement it.
I have done a similar type work, which takes the nearest 3 points and calculates new face field using inverse distance weight function method. hope this is useful
Chandra Murthy

All times are GMT -4. The time now is 07:57.