CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   How to create permeable surface on inlet boundary? (http://www.cfd-online.com/Forums/cfx/63349-how-create-permeable-surface-inlet-boundary.html)

Chilli83 April 6, 2009 09:33

How to create permeable surface on inlet boundary?
 
Hi all

I have to model a permeable surface upon an inlet boundary. From the surface I also need to specify a heat flux. The problem I have tried to illustrate on the figure at the link:
http://img201.imageshack.us/img201/1...lcfdonline.jpg

The way I have tried to handle it is to define the geometry of the room and define the inlet and outlet by the “Imprint Faces” command. Then I define the surface as a solid which I use the “Frozen” command and then I put the room into an enclosure so I have one part consisting of two bodies. But when I go to the pre in CFX I can not find out how to define it as a permeable surface without a thickness.

How can this be done?

Kind regards
Michael

Chilli83 April 8, 2009 04:25

Not ideas how to solve the problem?

mvoss April 8, 2009 09:19

hi,
did you get a "surface" in Pre on which you could apply a bc and your question is how to do that (create the bc for a permeable surface)?
OR
you didn't get a surface and you know how to setup a permeable surface?

neewbie

Chilli83 April 8, 2009 16:17

Hi Neewbie

I have the surfaces of the volume in Pre where I can apply bc but I have to define the heat flux by a subdomain with a heat load from the volume given in W/m^3 and I would like a surface without a thickness if possible.

I think this is caused by the way I have define the geometry of the room. First I have generated the room geometry and then use the imprint command where the inlet and outlet is located. Then I generate a solid volume with a small height which i define as frozen. Then I put the room to an enclosure and jump to CFX-mesh where I generate a grid.

So to sum up maybe I have problem both to generate the geometry correct and to get the surface into Pre. If possible I would like not to model the surface as a porus domain since this increase the computing time.

Kind regards
Michael

Quote:

Originally Posted by neewbie (Post 212334)
hi,
did you get a "surface" in Pre on which you could apply a bc and your question is how to do that (create the bc for a permeable surface)?
OR
you didn't get a surface and you know how to setup a permeable surface?

neewbie


mvoss April 9, 2009 07:37

hi,
so did you get a volume for the "quasi-surface"? I´m asking because with the imprint you won´t get a volume.
Try to sepparate the volumes by creating a multibody part in the DM (unfortunatly i don´t know the correct english option ) and make the "surface" one element over hight, mesh them separatly and connect them by ggi or 1to1.

neewbie

ps:When defining bc names in cfx-mesh all earlier component names (from DM) will be overwritten.

Chilli83 April 10, 2009 12:36

Hi neewbie

No I don't get a volume when I use the imprint faces command since this give me a 2D surface as you also talk about.
I think you are talking about to form one part from multiple bodies in DM? This I have done where I define the “surface” as a frozen solid volume which gives me the opportunity to make one mesh and don’t have to define permeable subdomain. But if I make to models and to mesh which I “melt” together will it then be possible to define on of the side of the volume as a permeable surface?

kind regards
Michael

Quote:

Originally Posted by neewbie (Post 212453)
hi,
so did you get a volume for the "quasi-surface"? I´m asking because with the imprint you won´t get a volume.
Try to sepparate the volumes by creating a multibody part in the DM (unfortunatly i don´t know the correct english option ) and make the "surface" one element over hight, mesh them separatly and connect them by ggi or 1to1.

neewbie

ps:When defining bc names in cfx-mesh all earlier component names (from DM) will be overwritten.


mvoss April 12, 2009 18:38

hi,
i would guess you have to modell the "surface" as a volume momentum source. So it would be a sepparate volume mesh.
Or how do you do it?

neewbie

Chilli83 April 13, 2009 10:49

Quote:

Originally Posted by neewbie (Post 212649)
hi,
i would guess you have to modell the "surface" as a volume momentum source. So it would be a sepparate volume mesh.
Or how do you do it?

neewbie

Hi neewbie

I have made a solid frozen volume which are enclosured bu a large room. The mesh is then also generated through the volume and then I can define a volume source in CFX-Pre. So to model it as an surface I have made the height "small". Is this the way to handle it?

Kind regards
Michael

Chilli83 April 14, 2009 13:51

Hi all

I have an extra question to my problem. The area where the inlet is model I have defined as a surface by the imprint command. In CFX-Pre I saw that it is possible to define a heat source on the “inlet surface”. The problem is then that the surface representing persons is around 10 m2 located in the middle of the inlet surface which has a total area 65.9 m2. So I was wondering if a way could be to define a surface surrounding the small surface as one part of the inlet and the enclosed surface as another part of the inlet which also release a heat source?

The next problem is then how to define the total incoming mass flow rate correct. If I know the total mass flow rate are 2.04 kg/s would it then give the “correct” flow if I divided the mass flow rate into two quantities of 1.728 kg/s and 0.312 kg/s for the two areas on 55.9 m2 and 10 m2 respectively?

Kind regards
Michael

mvoss April 15, 2009 04:35

hi,
by using the imprint command twice you should be able to get such a geometry/faces- and by applying the partial massflows you will end up with your desired totalmassflow.

neewbie

Chilli83 April 15, 2009 06:10

Quote:

Originally Posted by neewbie (Post 212958)
hi,
by using the imprint command twice you should be able to get such a geometry/faces- and by applying the partial massflows you will end up with your desired totalmassflow.

neewbie

Hi neewbie

Thank you for your reply. It seems that my idea was in the right way then :)

Kind regards
Michael


All times are GMT -4. The time now is 08:37.