Hi All

I'm trying to simulate the laminar flame front in tube. The geometry is 2D thin segment 2 meters long pipe and 25 mm radius. The problem is very, very long time of calculation because I must use very fine mesh in whole tube (0.25 x 0.25 mm). In fact, that mesh is only necessary in reaction area but not in all tube hence it would be useful some mesh adaption option. But this doesn't work for transient cases.

Does somebody know some tricks or hints to significant shorten time of calculation for this case?

Hi,

Why don't you use some mesh grading? Assuming you are using CFX-Mesh, set a coarser global mesh size and a mesh control with the 0.25mm mesh in the region of the flame. Alternately if you are using a quad/hex mesh set mesh seeds to bias the mesh towards the flame.

Glenn Horrocks

Hi Glenn

It is really good idea because there is no problem to do this grid. But I think the problem is not too easy to resolve because the flame front is mooving during combustion process and must have fine mesh all time. Therefore the area of fine mesh should displace too with the same rate. How to do this?

I have seen Tutoral 20 in manual, but I don't think so that is similar case to mine. In my case, to maintain the same number of nodes, the nodes of grid should be erasing in area before the flame and adding in region behind it. Any suggestions?

Regards

Hi,

It's possible to use mesh adaption in transient simulations. Take a look at the Ansys Solutions from the summer 2005, page 24: "Normally, the size of mesh required to do this would make a long transient simulation intractable but, by using <u>transient mesh adaption</u> to concentrate cells into the region around the interface, we were able to reduce the total mesh size by an order of magnitude."

I played a little bit with transient mesh adaption in CFX-5.7.1, but with no successful results at all. You had to have Service Pack 1 installed and set an environment variable to be able to use it. The guys from CFX support told me it wasn't a supported feature and due to the algorithm being not fully conservative it could be difficult to converge. I don't know if there has been any improvement in CFX-10, but if you want to try it, seek some help from the CFX-support.

Hi,

You could run a number of runs with graded meshes, and interpolate the results between them so the refined mesh region tracks the flame front. It is a bit cruder than transient mesh adaption but a LOT easier to implement.

Glenn Horrocks