CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   ANSYS Meshing & Geometry (https://www.cfd-online.com/Forums/ansys-meshing/)
-   -   [GAMBIT] Ask about how to set up geometry and mesh for moving mesh case (https://www.cfd-online.com/Forums/ansys-meshing/94595-ask-about-how-set-up-geometry-mesh-moving-mesh-case.html)

g-wagon November 20, 2011 22:13

Ask about how to set up geometry and mesh for moving mesh case
 
i'm newbie..i want to do dynamic analysis for my butterfly valve case 2D problem, as in real case i want to make a motion for my valve so i have to use moving mesh methode in FLUENT but i don't know how to set up geometry and mesh in GAMBIT for such case like this, please help me what i have to do first, here is my model
http://s8.postimage.org/c5th3hz11/katu_gambar.jpg
http://postimage.org/image/uldy0wd5d.jpg

-mAx- November 21, 2011 13:05

in your case, moving mesh (remeshing in fact) will be treated within FLuent solver.
So no need to do something in Gambit.
BUT I would suggest you to switch to sliding mesh.
Create a circle surrounding your butterfly valve and split your domain with the circle.
Because of sliding mesh, you need to create interfaces: your disk (valve included) will rotate and you need interfaces for interpolating data between rotor and stator.
FOr that: copy and translate anywhere your disk. Delete the original disk.
Assign remaining circle (stator-edge) as interface.
Assign also circle (rotor-edge from copy) as interface
Translate back your disk (rotor).
Now your 2 surfaces are disconnected and your rotor will be able to rotate free.
You can mesh both
Check tutorial on sliding mesh for set up in Fluent

g-wagon November 22, 2011 02:02

i just want to make a sure max, does my picture below right max??surface with two interface and then connect it in FLUENT
http://s7.postimage.org/4a1c9w2jv/setup.jpg

-mAx- November 22, 2011 03:21

Almost

-The interface has to be the full circle.In your case interface1 &2 belong to the same.
You will not able to see both interfaces separately since they are superposed.

-Then you may try to change your circle radius for preventing small angles as it seems you have in your picture (circle & top edge are tangential and will generate skewness problem). Try a very small change in radius (tolerance) just for creating a gap between entities

g-wagon November 22, 2011 03:26

okey i understand max what you want to talk about, the important things to create two interface between stator and rotor then connect it , right? okey max, thank for your assistance. owh max can you help me with the UDF? :)

-mAx- November 22, 2011 03:30

No you don't have to connect them. You need to couple the interfaces but lately in Fluent.
forget UDF with me... ;)

g-wagon December 15, 2011 02:44

hi max
how much surface i should have for this geometry??is it two?just the stator and for the rotor?

-mAx- December 15, 2011 02:55

at least 2.
Same topic here: http://www.cfd-online.com/Forums/ans...s-meshing.html

g-wagon December 15, 2011 03:08

http://postimage.org/image/4zqa8dg63i have my own geometry based on your guide, here is my geometry, is it correct??
http://s9.postimage.org/4zqa8dg63/katup222.jpg
http://www.cfd-online.com/Forums/%3C...5/katup222.jpghttp://postimage.org/image/4zqa8dg63/
http://www.cfd-online.com/Forums/%3C...atup222.th.jpghttp://imageshack.us/photo/my-images/100/katup222.jpg/
http://imageshack.us/photo/my-images/100/katup222.jpg/
http://imageshack.us/photo/my-images/100/katup222.jpg/

-mAx- December 15, 2011 03:53

I don't see your pictures

g-wagon December 15, 2011 07:03

http://s7.postimage.org/qo5sfh4kb/katup222.jpghttp://postimage.org/image/ytnudmst3/

-mAx- December 15, 2011 07:55

yes but your circle (interface) is too big.
Thus you will have mesh issue especially where the circle is tangential to your pipe.
So reduce the circle radius

g-wagon December 15, 2011 07:57

okey i got it max..thanks max. now my problem is to define the angular velocity for that thing using UDF

g-wagon December 20, 2011 00:41

hi max.. would you tell me about procedure for simulating this in fluent using sliding mesh approach and then tell me what should i do when i want to use UDF for the motion in this simulation using sliding mesh, thanks

-mAx- December 20, 2011 02:16

Prior to the sliding mesh, I would test if the interfaces are ok.
Juste define grid interfaces, and compute with your inlet velocity.
If everything is ok (understand fluid flows through the interfaces), then you can switch to sliding mesh.
Just enable dynamic mesh and set your angular velocity to your rotor domain. Since it is sliding mesh, no mesh option is required (smoothing, remeshing or layering).
For time dependant angular velocity, you can use UDF (maybe profile), but I am not so experienced in this field

g-wagon December 20, 2011 02:26

the first plan i want to use sliding mesh for this dynamic analysis, i have prepared with the geometry as you taught me last day.. this day when i'm trying to do the simulation i could not fix for the motion especially because i cannot input my UDF in sliding mesh approach.. based on your words, i can use dynamic mesh approach... is there any change in geometry preparation when i use dynamic mesh??or still be rotor and stator as you said,

-thanks a lot-

-mAx- December 20, 2011 02:36

dynamic mesh literally means dynamic update from mesh, but in you case you don't need to update your mesh since they are disconnected.
Understand you can rotate your rotor zone without any effect on stator mesh (thanks to interfaces)
Basically now you only need to define your rotor domain as rigid body, and defined a angular velocity.
Then you can preview the mesh motion, and you should see your rotor domain rotating

g-wagon December 20, 2011 02:46

so the first thing that i've to do is...checking the dynamic mesh parameter in fluent right??

-mAx- December 20, 2011 02:58

no
go to define/dynamic mesh/zones...
select your rotor zone, define it as rigid body, and set your angular-velocity.
I don t remember if you have to set your velocity in dynamic mesh panel, or in the boundary condition panel (you have to try)
(I don't have Fluent, I am basing on User Guide)

g-wagon December 20, 2011 03:05

i think you wrong max, we have to enable the dynamic mesh parameter and then we go to dynamic mesh zone :D

-mAx- December 20, 2011 03:08

yes enable the dynamic mesh, else you cannot access to dynamic mesh zone.
But don't change anything, since your mesh will not be updated (no layering, no smoothing and no remeshing)

g-wagon December 20, 2011 03:13

thanks boss, really appreciate it. btw come visit my country indonesia:)

g-wagon December 21, 2011 03:10

hi max i have problem again..
i have two interface in my geometry in fluent i tried to connect them using coupled but the contour of velocity was so weird check this out
http://www.cfd-online.com/Forums/%3C...pvl79/aneh.jpghttp://s10.postimage.org/ckww3r5zt/aneh.jpg

that is contour for initial solution..the wall is not rotating..but whay the fluid didnt pass the valve??

-mAx- December 21, 2011 03:13

yes you have problem at interfaces
Check if they are well defined.
I am in vacation untill January the 4th.

g-wagon December 21, 2011 03:19

but why??i think it had been defined well.. i have two surface not connected, i give that surface interface rotor and stator after that i use FLUENT to defined the interface type as coupled...

-mAx- December 21, 2011 03:24

no you need to define the edges as interface.
Between rotor and stator, only edges are common

g-wagon December 21, 2011 03:26

sorry i was wrong..i mean the edge of both stator and rotor

-mAx- December 21, 2011 03:28

no you mean one interface for the stator edges, and another interface for the rotor edges.
In Fluent select both interfaces to define the grid interfaces

g-wagon December 21, 2011 03:30

i've done that, and for the interface type i use coupled type

g-wagon January 2, 2012 21:13

hi max..me again hope you not bored with me :)
i'm facing a problem..yes again...
in my dynamic analysis using moving mesh, i'm facing the negative cell problem.. can you analysis what was wrong in my simulation, why did this comes up, here is what error like to be

http://s13.postimage.org/xe3sun41j/negative_cell.jpg

-mAx- January 4, 2012 01:47

something is wrong since you are not suppose to change your mesh.
Check where are the negative cells

ftimis February 3, 2012 06:33

Hi my friend, I am currently working on a similar project. I have to design a check valve. Is that working for you? Shall I follow the steps of your discussion??

Rahul123 September 8, 2012 11:26

I am also plannin on using Sliding mesh to rotate a square Cylinder 2D....can ny1 help in defining interfaces.....I am not able to understand how we have two interfaces from d same circle...Do I have 2 make two circles nd define dem as interfaces

-mAx- September 10, 2012 00:55

imagine a surface with a hole (disk) > you are agree that in this surface you have a circle.
imagine now a disk not located on the surface mentionned earlier, but anywhere > you are agree that a circle (edge) is attached to this disk
You have then 2 circles.
Now move your disk in such way, that it will cover the hollow from first surface.
Finally you have defined a stator (first surface), and the rotor (disk) which are disconnected. And you also have 2 superposed interfaces (one belongs to stator surface and the other belongs to the rotort (disk)) Here the interfaces are the circles (edges)
I Hope I was clear
;)

Rahul123 September 10, 2012 08:01

thanks max....How do I couple the two interfaces
1.coupled
2.periodic
or
3.periodic wall boundary.
I found these 3 options in fluent

-mAx- September 10, 2012 08:03

don't remember anymore, but coupled should be ok


All times are GMT -4. The time now is 16:51.