CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   Problem with grid interface (http://www.cfd-online.com/Forums/fluent/30067-problem-grid-interface.html)

Anindya July 22, 2002 18:39

Problem with grid interface
 
I have an impinging jet problem. I have two parts where one is a box shape and the other part is the round nozzle. I create the

mesh separately.

The bottom surface of the box is set as a wall and the sides are pressure boundaries (open to atmosphere).

The top surface of the box is set as interface-1 boundary condition.

The bottom surface of the round nozzle is set as interface-2.

The sides of the nozzle are set as walls and the top surface as an inlet.

I use Tgrid to merge the two meshes.

I open the merged mesh file in Fluent6.0 (3D). I define the interface using grid interface and give it a name as inteface-1-2.

Then two walls are created (say wall-15 and wall-16). wall-16 is the non over-lapping part of interface-1 and interface-2

which is exposed to the atmosphere. I reset the wall-16 boundary condition as pressure-inlet (to simulate entrainment

boundaries).

Then when I run Fluent6, it gives the error

"Error: fluent.6.0.20 received a fatal signal (SEGMENTATION VIOLATION). Error Object: #f " .

I am unable to get rid of this error. This problem only runs when I set the wall-16 back to wall. Any other boundary condition

does not work. If I set it to wall then my simulation is not right as the top surface is exposed and not covered. So I need a

pressure boundary condition there.

Could someone tell me what the problem is and how I can bypass it ? Or is this a bug in Fluent6?

Do reply soon. Thanks a lot.

Anindya

david July 22, 2002 19:53

Re: Problem with grid interface
 
a) Why didn't you use one domain at the beginning? For different mesh size ? I did some impinging jet simulations using one domain !! b) I think, the non-overlapping interface is set as wall by default, if you change it, fluent will be confused. Did you try to use Fluent 5, just to checnk whether its a bug of F6 or not ?

By the way, what model are you using to capture entrainment behaviour?

DC

anindya July 22, 2002 20:40

Re: Problem with grid interface
 
Hi David, Thanks for your response. I did it with two domains as I need to model a moving jet, so that the nozzle will move with a specified speed relative to the box. I did model for a static case with one domain only as you said and that works fine. Alos having two domains helps a lot in meshing as I can mesh the box below with a hexa-mesh and the nozzle can be meshed separately with a combination of tet/hex.

I did not try the case with Fluent5 as I have Fluent6 only loaded now. So I cannot check if there is a bug with Fluent6 or not.

For entrainment I am using a pressure inlet b.c. at the top. Do you think it is ok? The sides are pressure outlet. What did you use for the entrainment b.c?

david July 22, 2002 21:21

Re: Problem with grid interface
 
Did you mean entrainment of air in the form of bubbles? DC

anindya July 23, 2002 03:25

Re: Problem with grid interface
 
No .. I meant that air can either enter or leave at that boundary surface. For the impingement jet case, the air can come into or leave the box also from the top surface just next to the nozzle . So I believe that the only possible boundary condition to simulate entrainment is a pressure bc.

david July 23, 2002 05:04

Re: Problem with grid interface
 
What about you creat only one domain where you keep the top of the box open to atmosphere and in fluent use rel velocity of the jet (horizontal ?).

Not sure though whether its possible or not. I modeled impinging jet and also moving wall using fluent and cfx, but not these two together !!

DC

hendawi July 23, 2002 08:01

Re: Problem with grid interface
 
Hi,

About the entrainment BC, you should use either Pressure, symmetrie or a velocity inlet with direction parallel to the boundary. Otherwise i'm studying nearly the same case as yours (sprinkler jet modelisation (with VOF)) and I'm encountering exactly the same interface problem. I don't know if fluent considers the boundary created as wall only or not (saying not possible to modify the wall (wall 16) type to velocity or pressure inlet!). please let me know if you found any solution to solve this problem.

Best regards, Mohamed

adrien July 23, 2002 08:25

Re: Problem with grid interface
 
Hi Anindya,

The case i study, is very near from yours, but i didn't manage to mesh correctly the cylindric nozzle and the box in one domain as to obtain convergence (VOF model). I also tried with an interface and i got the same problem as you and mohamed.

Would you be so kind as to send your mesh or piece of information to help me to mesh my case (type of mesh use for the nozzle and the box...).

Best regards, Adrien

anindya July 23, 2002 19:22

Re: Problem with grid interface
 
Hi David,

I have not tried that but will give it a try. If that works out it is fine. Else I have to go back to what I was doing before. I do not know how to contact the fluent guys as ours is University license and I cannot get into the fluent user forum.

Hendawi,

I will try to solve the problem in a different way. I am still not sure how, but I believe that I can do that. Else I will see if STARCD can do this simulation.

Adrien,

what problem do you have with the meshing? what type of mesh are you using for the box and the nozzle? did you try several grids to achieve convergence or manipulate the tolerance values for the solvers, etc?

Anindya

Anindya July 26, 2002 16:04

Re: Problem with grid interface
 
David,

I think that doing this will not work and simulate the real condition. I need to have two domains and move one domain(nozzle). Using a horizontal relative velocity will not work or simulate the real moving impinging jet problem.

Let me know if you could find out the solution for the interface condition.

have a nice weekend.

Anindya


All times are GMT -4. The time now is 06:17.