CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   STAR-CCM+ (http://www.cfd-online.com/Forums/star-ccm/)
-   -   Heat exchanger (http://www.cfd-online.com/Forums/star-ccm/100071-heat-exchanger.html)

sundar April 19, 2012 11:47

Heat exchanger
 
Hello Everyone,
I'm new to this forum.I'm doing cfd analysis on heat exchanger in Star CCM+.While doing meshing i can't able to mesh the interface between the water domain and the inlet tube part containing air.Its a cross flow concentric tube heat exchanger.
can anyone please help me in this?

Thanks in advance.

iDeew April 23, 2012 12:36

heat exchanger
 
I'm trying to do almost the same thing in Star CCM+ too. My heat exchanger is a cross flow water heater. Hot air flows upwards heating a steel tube caring a turbulent flow of water. I just want to simulate the temperature distribution of the water flow inside the steel tube.

I first created 3 regions air,steel & water. then created the interfaces between air/steel & steel/water. and press Generate Vol. Mesh. I can create the mesh without any trouble. But once I defined all inlets & outlets and Run the simulation Star CCM+ gives me an error.

What is the error msg your getting?

sundar April 24, 2012 12:28

Heat exchanger
 
For me volume mesh itself showing error.When i ran the surface diagnostics it shows many free edges.I'm trying to reduce the errors showing.As i'm a learner i find it quiet difficult to mesh properly.
Thanks for your reply

madhuvc April 27, 2012 17:04

@sundar .. without a closed clean mesh you won't be able to generate the volume mesh. if you have large number of errors to fix it is in your best interest you can take that model to the pre-processor, clean it and bring it back in.

if you want to avoid this, you can try surface wrapper, but that's a whole different story.

if you have less no of errors to fix, it is easier to fix using surface repair tools in STAR-CCM+. let me know what exactly your issue is, so that I can be more specific.

abdul099 April 28, 2012 07:58

@sundar: Can you put your model here or at least some pictures? It's quite hard to give any advice without having seen the geometry and the issue at all.

@iDeew: I assume, your error message is related to your own post and therefore (according to your last post) no longer relevant? If not, please describe it in detail.

sitekss April 30, 2012 09:20

Try this:
-take one domain
-apply really fine surface mesh
-export the mesh
-repeat for all other domains
-open new Star-ccm+
-import surface meshes
-apply right sizes
-apply interfaces
-voila

abdul099 April 30, 2012 15:08

@sitekss
that's not necessary. Meshing all together in one simulation should be enough. And even with your approach, you still need the right setup.
Anyway, sundar has some errors in the surface, therefore he needs to fix that before he can mesh anything, with or without export / import.

sitekss May 1, 2012 13:35

Abdul,
I know what he wants to do. I suggested the 'trick' just to loose all invalidities comming from poor quality geometry. It is one of the way how it may be done.
I do not recommend fixing geometry at interfaces because it result in gaps in the interface (some of the surface is te interface, some is not).

BR
Adam

abdul099 May 1, 2012 16:44

The issue is not just a poor surface quality, since the surface remesher will take care of this. The issue is the surface which is not error free (free edges, mentioned by sundar some posts ago). Once he fixed that (on part level, like CD-adapco recommends it), it should not be an issue to generate a valid volume mesh. When you fix the surface on part level, you will never need to fix something on the interfaces since the interfaces are not existing at this point.

sitekss May 2, 2012 02:07

You've right. At the beginning he should repair the surface, then think about volume mesh and interfaces. However, the mesher works better with fine triangulate representation of surfaces.

abdul099 May 2, 2012 14:41

Please don't misunderstand me, I don't want to criticise you or the way you're working. I just want to provide some information to you, so you can make founded decision how to continue.

For the surface remesher, it doesn't matter since it's the remesher's job to provide this surface.
Sure, for the volume mesher it's necessary to provide an appropriate triangulated surface. But before volume meshing, you usually should run the surface remesher anyway, especially when interfaces are used (otherwise you will never have conformal interfaces). And also the surface sizes dictate the resulting volume cell size, so it might be a good idea to remesh it all together, so proximity refinement, curvature and growth rates can work in the designated way.
So although your approach might be the only one in some special cases, at least in usual cases it doesn't give you an advantage to follow your approach.

That doesn't mean, your approach doesn't work. It does. But it contains some unnecessary steps and therefore is inefficient. And further is has some disadvantages. You will loose all CAD information. Usually all vertices will be projected to the CAD surface during surface remeshing (as far as CAD information are available). When applying a finer mesh, it will better follow the CAD surface. But after the export, all CAD information is lost, so all you will remesh after importing it again is the already triangulated surface.


All times are GMT -4. The time now is 23:16.