Wall Boundary Condition for UDS

ftab · July 23, 2012, 11:40

HI all,
Could any one please elaborate for me what is the meaning for constant value of UDS on the wall boundary?
What I do not get is the dimension. Putting the constant value of say 1 for UDS on the wall, for instance for concentration, will put the concentration equal to 1 kg/m^3 volumetrically? or it will simply put it equal to 1 kg/m^2 since the wall is surface? I do not get the physical meaning properly and it is not clear also in the documentation.

Thanks for your comments!

Amir · July 24, 2012, 10:01

Quote:

Originally Posted by ftab

Could any one please elaborate for me what is the meaning for constant value of UDS on the wall boundary?
What I do not get is the dimension. Putting the constant value of say 1 for UDS on the wall, for instance for concentration, will put the concentration equal to 1 kg/m^3 volumetrically? or it will simply put it equal to 1 kg/m^2 since the wall is surface? I do not get the physical meaning properly and it is not clear also in the documentation.

Dear Farhad,

The UDS is a non-dimensional variable which is evident from its conservative equation! As you know this is a linear PDE, so you can multiply the results by a number to achieve your purpose. (note that you can set a Dirichlet or Neumann BCs at your boundaries whether you know the UDS value or flux there). Don't hesitate to ask if it's not clear ....

Bests,

ftab · July 24, 2012, 10:37

Dear Amir, Thanks a lot for the prompt reply.
My main difficulty is coming from the fact that I am trying to duplicate a test case of Fluent in CFX. The reference case is simply putting the Dirichlet boundary condition of 1 for UDS which is really easy to apply in fluent. I also did the same and got the similar result in Fluent ( Still not converging, and increasing the Under-relaxation diverges the simulation immediately, any hint on this?).
But, in CFX, similar to UDS we have additional variable, and there we have dimensions and hence, I can put Value of concentration (kg m^-3) or flux (kg m^-2 s^-1) or a boundary source flux (kg m^-2 s^-1) or total value (kg s^-1). And this is exactly where I am puzzled and I cannot fix the boundary condition to be exactly similar to Fluent case. Is my question clear now?
It might not be clear for you if you are not working with CFX, but your general comment on this dimensional consistency would be of great help.

Amir · July 24, 2012, 11:11

Quote:

Originally Posted by ftab

My main difficulty is coming from the fact that I am trying to duplicate a test case of Fluent in CFX. The reference case is simply putting the Dirichlet boundary condition of 1 for UDS which is really easy to apply in fluent. I also did the same and got the similar result in Fluent ( Still not converging, and increasing the Under-relaxation diverges the simulation immediately, any hint on this?).

Check whether you may solve UDS equation in a de-coupled manner. Also choosing proper convective scheme or resolving the grid in critical regions may help.

Quote:

Originally Posted by ftab

But, in CFX, similar to UDS we have additional variable, and there we have dimensions and hence, I can put Value of concentration (kg m^-3) or flux (kg m^-2 s^-1) or a boundary source flux (kg m^-2 s^-1) or total value (kg s^-1). And this is exactly where I am puzzled and I cannot fix the boundary condition to be exactly similar to Fluent case. Is my question clear now?
It might not be clear for you if you are not working with CFX, but your general comment on this dimensional consistency would be of great help.

Note that UDS equation is a general tool as scalar equation. FLUENT also has specific tool for mass transfer named species transport where you can find desired dimensions but has few restrictions. (which is a special case of UDS equation)
There shouldn't be any contradiction if you exactly know what's the governing equation and BCs. you can consider the UDS as a normalized concentration if you think it may help. but just note that the flux BC is:
$(\rho D \nabla C-\rho U C).\vec{n}$
where rho is the base fluid density not the UDS's.

Bests,

ftab · July 24, 2012, 12:07

Thanks a lot for your answer Amir,
Comparing the theory of Fluent and CFX, I can see that the PDE is exactly identical in both softwares, which heartens me that they will lead to similar results.
The only difference is that in CFX you also specify the dimension, and in this case the constant value BC in CFX would be identical to constant UDS value in Fluent, and the same for the flux I suppose.
Your comments really helped me to clarify it a lot, and I do appreciate that.

July 23, 2012, 11:40	Wall Boundary Condition for UDS	#1
ftab Member Ftab Join Date: Sep 2011 Posts: 87 Rep Power: 15	HI all, Could any one please elaborate for me what is the meaning for constant value of UDS on the wall boundary? What I do not get is the dimension. Putting the constant value of say 1 for UDS on the wall, for instance for concentration, will put the concentration equal to 1 kg/m^3 volumetrically? or it will simply put it equal to 1 kg/m^2 since the wall is surface? I do not get the physical meaning properly and it is not clear also in the documentation. Thanks for your comments!

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Natural convection in a closed domain STILL NEEDING help!	Yr0gErG	FLUENT	4	December 2, 2019 01:04
Applying Rotational Boundary Condition on Surface with Radial Wall	tango711	CFX	16	October 9, 2018 08:01
Domain Imbalance	HMR	CFX	5	October 10, 2016 06:57
isothermal wall boundary condition	Neil	Main CFD Forum	3	November 9, 2015 03:34
CFX13 Post Periodic interface	EtaEta	CFX	7	December 8, 2011 18:15

July 24, 2012, 10:37		#3
ftab Member Ftab Join Date: Sep 2011 Posts: 87 Rep Power: 15	Dear Amir, Thanks a lot for the prompt reply. My main difficulty is coming from the fact that I am trying to duplicate a test case of Fluent in CFX. The reference case is simply putting the Dirichlet boundary condition of 1 for UDS which is really easy to apply in fluent. I also did the same and got the similar result in Fluent ( Still not converging, and increasing the Under-relaxation diverges the simulation immediately, any hint on this?). But, in CFX, similar to UDS we have additional variable, and there we have dimensions and hence, I can put Value of concentration (kg m^-3) or flux (kg m^-2 s^-1) or a boundary source flux (kg m^-2 s^-1) or total value (kg s^-1). And this is exactly where I am puzzled and I cannot fix the boundary condition to be exactly similar to Fluent case. Is my question clear now? It might not be clear for you if you are not working with CFX, but your general comment on this dimensional consistency would be of great help.

July 24, 2012, 12:07		#5
ftab Member Ftab Join Date: Sep 2011 Posts: 87 Rep Power: 15	Thanks a lot for your answer Amir, Comparing the theory of Fluent and CFX, I can see that the PDE is exactly identical in both softwares, which heartens me that they will lead to similar results. The only difference is that in CFX you also specify the dimension, and in this case the constant value BC in CFX would be identical to constant UDS value in Fluent, and the same for the flux I suppose. Your comments really helped me to clarify it a lot, and I do appreciate that.