CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   mesh deformation (

haka November 24, 2010 07:38

mesh deformation
hello everyone,
i have problem involving cylindrical wafer getting polished by 2 plates(opposite rotational velocity) top and below. there is a water flow from top plates through nozzles to cool the process. i have to simulates this problem. with this wafer rotate with the plates with an angluer velocity.

i set this domain to mesh deformation ,
define the plates BC to rotating wall with + and - velocity.
and for wafer BC to mesh motion with specific displacement with cylindrical components .on cylindrical components i mention theta as distance between center of axis and center of wafer.

and getting a error with negetive volume.

please if some one can guide me ,if im on right direction or not.
i'm new to CFD im using this for my please explain me little bit in layman terms. i had read all the tutorial they dont have anythign for this perticular type of problem.
i want similar effect with the rotational velocity. and im learning CEL little bit.

ghorrocks November 24, 2010 17:49

Can you post a drawing of what you are doing?

visagant November 26, 2010 12:04


Originally Posted by ghorrocks (Post 284684)
Can you post a drawing of what you are doing?

HI ghorrocks,thanx for the reply.
here is ther drawing of similar polishing machine this is with less details, i hope you understand what i'm talkin about.

because of wafer rotaions ,im planning on using mesh deformation.
some suggest to use adaptive mesh refinment(but i dont know how to do this in ansys 12.1-cfx)

top and low plates have a 2m diameter and space between wafer and plates is 5e-6mm, because of this 2 extreme parameters some suggest it can't be done because of too many nodes to simulate.

in real case there will be more nozzle for fluid inlet to cool down the process.

please let me know if u need more details. sorry if bad technical wordings.
please help me out onthis.

ghorrocks November 27, 2010 05:44

Can you explain what you are trying to learn by doing a simulation? What is the important physics of what you want to do?

visagant November 27, 2010 18:21

thanx for the reply. This is a polishing tool for glass or plastic lens(wafer).because of the polishing at higher speed, the polishing surface will get heated ,this will make the unevenness in polishing plate surface, this will contribute to the unevenness in lens(wafer). so we try to learn the temperature distribution along the plate. this might help to control the speed of polishing and control the injection of cooling water to keep the temperature at desired amount.

ghorrocks November 27, 2010 20:53

OK, now I see what you are doing. So really all you want to know is how the coolant controls the temperature of the grinding.

I suspect the rotation does not play an important role. You can probably run this ignoring the rotation. Then it is just a normal CHT simulation.

visagant November 28, 2010 06:07

thanx for the quick reply. i have to ask some basic question please forgive me,ok but how temperature will increase if there no rotations are defines in grinding(friction between the plates and wafer cause the temperature increase) and my prof is looking forward to see the mesh deformation(AMR) type of simulation.

galap November 28, 2010 07:46

I also think there's no need for a mesh deformation simulation. In any case you have to apply a definition for the friction coefficent. You can set the heat production by a CEL expression as a function of the tangential velocity (again CEL). This heat production should be applied as energy source in a small subdomain adjacent to your interface. I never done this, but I think that should work. Don't set the heat source at your interface - that will result in a jump (saltus) of your temperature profile over your interface.

visagant November 29, 2010 06:42

thanx galap,
ok i will try you have a CEL manuel,similar to UDP manuel.there is non in ansys customer portal.

ghorrocks November 29, 2010 18:21

Sorry, yes I forgot to mention. No need for moving mesh and probably no need for rotating frames of reference either.

But the heat source can be applied to the outer face of the solid, just as a standard boundary condition. No need to apply it to a solid region.

The CEL functions are explained in the CFX reference manual.

visagant December 1, 2010 07:49

hi thanx ,for the replay. but CHT simulation what you suggest it works for single way rotation,mine is more like a earth rotating around the sun,with multiple self rotations.
i think this wont reflect in CHt simulation, please correct me if im wrong.

ghorrocks December 1, 2010 17:50

You will have to show a diagram which shows what you are talking about. Your previous diagrams did not suggest any counter rotation was required.

All times are GMT -4. The time now is 20:28.