[Matlab69]

Hi guys,

I am trying to model a methane flow around a bluff body.

The geometry of the bluff body results in two methane inlets at the bottom, with pressure outlets at the top.

The problem I am having is that the exit flow is not symmetric, even though as you can see, the geometry is perfectly symmetric.

I believe that the asymmetric flow is caused by the generated mesh.

Any idea how I could tackle this problem to achieve a symmetric recirculation zones?

THANKS ALOT

[diamondx]

What type of mesh did you use. i will go and try another mesh to see if results are different !

[Matlab69]

This is what the actual geometry looks like and the mesh.

I have messed about with meshing, and all of the setups result in similarly assymetric exit flow....

Thanks in advance

[diamondx]

why don't you try an hexa mesh using icem. can you attach a geometry file? see what i can do about it

[Matlab69]

I am new to this program and do not clearly understand what you mean by that..? Could you explain how I could hexa mesh it?

The geometry file is too big too upload, perhaps I could send you it through skype?

Thanks

[diamondx]

I assume you have Ansys, if so, it comes with beautiful programs to create your mesh. You can use either ICEM CFD or ansys meshing. Here, take a look at two videos made by Mr Simon showing how two perform 2D hexa meshing for an airfoil wich can be applied to your case since your geometry is easier.
Using Ansys Meshing : http://www.youtube.com/watch?v=tYrbScUH9RE
Using ICEM CFD: http://www.youtube.com/watch?v=tYrbScUH9RE
Do you have dropbox, maybe you can drop you geometry there and send me the link.
If not get an account there . How can you not use DROPBOX in university ?? just kidding. anyway good luck with that and let me know if you have more questions

[Matlab69]

I didn't have it, but I do now

http://dl.dropbox.com/u/73065297/Geom.agdb

It does work, right?

Thanks!

[Matlab69]

...............

[Matlab69]

I have tried looking at the videos, but my ANSYS interface is completely different...



The options I have to mess about with the mesh:


THANKS

[diamondx]

ok first of all , i mistook the link i had to send you. Yes one of the link has Mr simon performing meshing on the airfoil, and if you said that the interface is not the same as yours, it means that you are only familiar to Ansys meshing and not ICEM CFD. Icem CFD is another program that allows you to create hexa meshs, so if you want to give it a try. just follow Mr simon in that video. One of the link i wanted to send you instead of the second video is a tutorial of Cornell university. here is the link
https://confluence.cornell.edu/pages...geId=144976434
In that tutorial, the guy will make a geometry of an airfoil in and meshs it in ansys meshing. it's the same program as your so you shoud not worry about it.
The file you sent me is not good. When you save a project, you should found the project extension wich is .adgb an the file that goes with it -when all the geomtry and meshs are stored-. to sum it up, if you wanna send a project, you need to send the .agdb and the file together zipped, it's better. OR click on export geometry and save it as an IGES. i hope i was clear, give me a feedback whenever you get a chance

[Matlab69]

Could you redraw the geometry (specifications in one of the print screens), mesh it and send it to me..?

I just can't do it.

Do I have to have a symmetric mesh or will be final result will converge to symmetry despite assymetric mesh?

[diamondx]

i need information about your geometry !!! length, height ??

[Matlab69]

So,

Draw this shape

cut out this shape

and you shoul get this shape


It is a flow through bunsen burner with the triangular bluff body fitted into it. I am trying to study the recirculation past the bluff body due to stagnation pressure differences. I am trying to achieve 2 symmetric recirculation zones. My problem is that the zones are assymetric. Could you try it yourself..? Draw, mesh, and model it? I am including turbulence with a k-e model.

Could you explain me one thing...is it crucial to have a symmetic mesh? I have not really let any of the calculations converge, I would stop my simulation after about 200 iterations and find out that the velocity vector distribution is not symmetic.
Could it also be caused by the turbulence parameters?

Thanks alot....

[diamondx]

i will see what i can do about it. The purpose of symmetry is to reduce the time of calculation. if you draw your geometry like this, it's too late you can't assume symmetry. what you need to do is to draw just half of your geometry, left or right then the mid-axis has to be called symmetry. Anyway i will see what i can do about it

[Matlab69]

Do you mean like this?


Highlighted areas mean are named selections.

I have tried that, but when I specify the model to be axis symmetric it does not work...I would prefer it this way though.

[Hybrid]

not axi-symmetric? it's mirror symmetry.

[Matlab69]

I don't think I can use symmetry due to the nature of the problem.

I am studying two independent circulation zones which have an effect on each other and only modelling one of them will not approxite the real thing..am I correct in saying is?


Thanks

[diamondx]

i finished your mesh sorry for the timing.
Here is the link: http://dl.dropbox.com/u/35161486/bluff%20body.rar

In the zip file file you find the mesh file named bluff.msh. i also joined the bluff.prj so you can open it in icem and play with it (learn )

[Matlab69]

Hey, thanks alot for that!

I will try it as soon as I can open it. Can you explain how I can do that please?

I opened the workbench, created a geometry and mesh tabs and tried to import geometry but nothing happened.

[diamondx]

just open fluent directly and then open the mesh file inculded in the zip i send you.
The file name is bluff.msh