Using Hybrid values in expressions

AliMadayen · May 1, 2014, 11:58

Hi everyone

In my mass transfer model, I have to define a boundary condition in a way that wall flux is related to velocity.
But my results are not correct, since CFX by default uses conservative values of wall velocity. Taking the plot in CFX post I have clearly observed that Conservative wall velocity is wrong, while hybrid values are valid.

Now my question is this:
How can I define an expression, in which I can use hybrid values of velocity in it?

A side question too:
My wall velocity is constant, But When I feed the constant value to the boundary condition, the answer is wrong. Can anyone explain this?

An additional explanation:

By wall velocity I mean injection/suction velocity and not moving wall. As you can find in my previous thread, I define that by a source term in continuity equation on the wall boundary. It works well.

singer1812 · May 1, 2014, 13:18

Please read the help on the difference between hybrid and conservative.

For wall velocity look up the varible Wall U, Wall V, Wall X

AliMadayen · May 1, 2014, 18:20

thanks for your answer

I already know the difference between conservative and hybrid types of variable.
My problem is while I want to include the value of velocity in an expression, CFX takes the conservative value on the boundary, which in my case and most cases is not correct, so I'm looking for a way that I can use hybrid values in my expressions.

And about the second part, I only found wall normal velocity, which is only available when the boundary is moving. Did you mean this variable?

singer1812 · May 1, 2014, 18:24

If your boundary is not moving, then your answer is 0 m/s.

Or perhaps I am not understanding what you are trying to get to?

If you mean you really need to get velocities "close" to the wall, but not actaully at the wall (0m/s for no-slip), then use conservative.

If you find that the near wall conservative value is not accurate, then you havent resovled your boundary layer enough. Add more near wall mesh....

AliMadayen · May 1, 2014, 18:25

An additional explanation:

By wall velocity I mean injection/suction velocity and not moving wall. As you can find in my previous thread, I define that by a source term in continuity equation on the wall boundary. It works well.

singer1812 · May 1, 2014, 18:27

Ahh, I see. I wasnt following that. That will require more thought, but I still think resolving the near wall area more will help you.

AliMadayen · May 1, 2014, 18:28

you mean making the near-wall mesh finer will make conservative and hybrid values converge?

ftab · May 27, 2014, 08:43

Dear Ali,
I totally understand your question, as it is my own struggle at the moment.
I do not think resolving the geometry would solve the problem in your case. The difference between conservative and hybrid variables are actually the differences of integration point (ip) and nodes, and refining the mesh would help in a case that the BC on the wall is no slip.

In your case you are applying the constant filtration velocity and this is applied on the ip (not sure myself) and then you have the no slip boundary which will zero the boundary NODES. Thus I do not expect them to converge in your case.

As I said, This is also my problem at the moment, and would appreciate if you let me know your idea on this. I am applying this on the interface of fluid-fluid.

P.S., I have seen this quote somewhere:
"To make the solver use hybrid variables, use the below expert parameter:

old bcp aver = t"
Does this help you?

ftab · May 28, 2014, 07:00

There is a simpler way to do that.
You can export the hybrid info to a csv or text and then define a function for that. In your expression, you just call that function.
This way you are sure the numerical values you are using are the ones you would prefer.

AliMadayen · May 28, 2014, 17:46

Dear ftab, thanks for your reply.
I see you have truly understood the problem.
After trying so hard, I found the solution, but with using constant filtration velocity.
While using a constant value for the filtration velocity, I made the mesh finer and finer. Eventually using a very very fine mesh, I could get the correct results. But the mesh is a lot finer that what one could imagine to be physical for similar simulations.
So as you see I haven't done anything for this problem of hybrid values. In some post of this forum written by an expert I read that doing so (using hybrid values in an expression) is impossible. But I will try your suggested method.
And about the second method of yours, i think it is only valid when the wall velocity is negligible regarding main flow velocity and doesn't affect the main flow hydrodynamics, am I right? Otherwise using this method would require a trial and error approach.

ftab · September 24, 2014, 06:59

Hi Ali,
Could you update me on this?
Did you manage to handle the filteration velocity on the wall with a method other than refining the mesh?
Could you apply the hybrid velocity ultimately?

Cheers,
Ftab

AliMadayen · September 30, 2014, 11:49

Hi ftab

No I did not pursue that problem any longer and moved on.
As I said I heard that CFX is not capable of using hybrid values in CEL.
I hope I do not encounter this problem again.
pls let me know if you find any solutions.

Best

P.S. I handled the filtration velocity by the mass source on the wall boundary, not by refining the mesh (tnx to the Mr. Ghorroks)
Let me know if you need any details regarding this.

Opaque · September 30, 2014, 15:49

Have any of you tried using "Velocity u.hybrid", or "Velocity u.Boundcon" ?

It is documented. However, mass sources are not boundary conditions in the sense of CFX; therefore, you may get a 0 [m/s] for hybrid, != 0 [m / s] as the control volume velocity, but not the injection velocity you specified.

Hope the above helps,

jaimesdiegop · May 31, 2015, 19:10

Hi Opaque I'm really interested in this CEL that you use "Velocity u.hybrid". saddly on my POST, when I use "function(Velocity u.hybrid)@location" CFX show:

"The following unrecognised name was referenced: Velocity u.Boundcon"
or
"The following unrecognised name was referenced: Velocity u.hybrid"

If there is some kind of trick to use, please let me know.

Thank you, this could be the solution of one big problem that I have.

May 1, 2014, 11:58	Using Hybrid values in expressions	#1
AliMadayen New Member Ali Madayen Join Date: Apr 2014 Posts: 22 Rep Power: 13	Hi everyone In my mass transfer model, I have to define a boundary condition in a way that wall flux is related to velocity. But my results are not correct, since CFX by default uses conservative values of wall velocity. Taking the plot in CFX post I have clearly observed that Conservative wall velocity is wrong, while hybrid values are valid. Now my question is this: How can I define an expression, in which I can use hybrid values of velocity in it? A side question too: My wall velocity is constant, But When I feed the constant value to the boundary condition, the answer is wrong. Can anyone explain this? An additional explanation: By wall velocity I mean injection/suction velocity and not moving wall. As you can find in my previous thread, I define that by a source term in continuity equation on the wall boundary. It works well. Last edited by AliMadayen; May 1, 2014 at 18:29.

May 27, 2014, 08:43		#8
ftab Member Ftab Join Date: Sep 2011 Posts: 87 Rep Power: 16	Dear Ali, I totally understand your question, as it is my own struggle at the moment. I do not think resolving the geometry would solve the problem in your case. The difference between conservative and hybrid variables are actually the differences of integration point (ip) and nodes, and refining the mesh would help in a case that the BC on the wall is no slip. In your case you are applying the constant filtration velocity and this is applied on the ip (not sure myself) and then you have the no slip boundary which will zero the boundary NODES. Thus I do not expect them to converge in your case. As I said, This is also my problem at the moment, and would appreciate if you let me know your idea on this. I am applying this on the interface of fluid-fluid. P.S., I have seen this quote somewhere: "To make the solver use hybrid variables, use the below expert parameter: old bcp aver = t" Does this help you? AliMadayen likes this.

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
cell values vs node values	reversemermaid	FLUENT	0	March 13, 2014 19:06
Local values at BC cel-expressions	Wonderz	CFX	7	October 31, 2013 11:50
ATTENTION! Reliability problems in CFX 5.7	Joseph	CFX	14	April 20, 2010 16:45
strange node values @ solid/fluid interface - help	JB	FLUENT	2	November 1, 2008 13:04
Hybrid mesh generation	Jake	Main CFD Forum	2	April 21, 2007 15:27

May 1, 2014, 13:18		#2
singer1812 Senior Member Edmund Singer P.E. Join Date: Aug 2010 Location: Minneapolis, MN Posts: 511 Rep Power: 22	Please read the help on the difference between hybrid and conservative. For wall velocity look up the varible Wall U, Wall V, Wall X

May 1, 2014, 18:20		#3
AliMadayen New Member Ali Madayen Join Date: Apr 2014 Posts: 22 Rep Power: 13	thanks for your answer I already know the difference between conservative and hybrid types of variable. My problem is while I want to include the value of velocity in an expression, CFX takes the conservative value on the boundary, which in my case and most cases is not correct, so I'm looking for a way that I can use hybrid values in my expressions. And about the second part, I only found wall normal velocity, which is only available when the boundary is moving. Did you mean this variable?

May 1, 2014, 18:24		#4
singer1812 Senior Member Edmund Singer P.E. Join Date: Aug 2010 Location: Minneapolis, MN Posts: 511 Rep Power: 22	If your boundary is not moving, then your answer is 0 m/s. Or perhaps I am not understanding what you are trying to get to? If you mean you really need to get velocities "close" to the wall, but not actaully at the wall (0m/s for no-slip), then use conservative. If you find that the near wall conservative value is not accurate, then you havent resovled your boundary layer enough. Add more near wall mesh....

May 1, 2014, 18:25		#5
AliMadayen New Member Ali Madayen Join Date: Apr 2014 Posts: 22 Rep Power: 13	An additional explanation: By wall velocity I mean injection/suction velocity and not moving wall. As you can find in my previous thread, I define that by a source term in continuity equation on the wall boundary. It works well.

May 1, 2014, 18:27		#6
singer1812 Senior Member Edmund Singer P.E. Join Date: Aug 2010 Location: Minneapolis, MN Posts: 511 Rep Power: 22	Ahh, I see. I wasnt following that. That will require more thought, but I still think resolving the near wall area more will help you.

May 1, 2014, 18:28		#7
AliMadayen New Member Ali Madayen Join Date: Apr 2014 Posts: 22 Rep Power: 13	you mean making the near-wall mesh finer will make conservative and hybrid values converge?

May 28, 2014, 07:00		#9
ftab Member Ftab Join Date: Sep 2011 Posts: 87 Rep Power: 16	There is a simpler way to do that. You can export the hybrid info to a csv or text and then define a function for that. In your expression, you just call that function. This way you are sure the numerical values you are using are the ones you would prefer.

May 28, 2014, 17:46		#10
AliMadayen New Member Ali Madayen Join Date: Apr 2014 Posts: 22 Rep Power: 13	Dear ftab, thanks for your reply. I see you have truly understood the problem. After trying so hard, I found the solution, but with using constant filtration velocity. While using a constant value for the filtration velocity, I made the mesh finer and finer. Eventually using a very very fine mesh, I could get the correct results. But the mesh is a lot finer that what one could imagine to be physical for similar simulations. So as you see I haven't done anything for this problem of hybrid values. In some post of this forum written by an expert I read that doing so (using hybrid values in an expression) is impossible. But I will try your suggested method. And about the second method of yours, i think it is only valid when the wall velocity is negligible regarding main flow velocity and doesn't affect the main flow hydrodynamics, am I right? Otherwise using this method would require a trial and error approach.

September 24, 2014, 06:59		#11
ftab Member Ftab Join Date: Sep 2011 Posts: 87 Rep Power: 16	Hi Ali, Could you update me on this? Did you manage to handle the filteration velocity on the wall with a method other than refining the mesh? Could you apply the hybrid velocity ultimately? Cheers, Ftab

September 30, 2014, 11:49		#12
AliMadayen New Member Ali Madayen Join Date: Apr 2014 Posts: 22 Rep Power: 13	Hi ftab No I did not pursue that problem any longer and moved on. As I said I heard that CFX is not capable of using hybrid values in CEL. I hope I do not encounter this problem again. pls let me know if you find any solutions. Best P.S. I handled the filtration velocity by the mass source on the wall boundary, not by refining the mesh (tnx to the Mr. Ghorroks) Let me know if you need any details regarding this.

September 30, 2014, 15:49		#13
Opaque Senior Member Join Date: Jun 2009 Posts: 1,929 Rep Power: 34	Have any of you tried using "Velocity u.hybrid", or "Velocity u.Boundcon" ? It is documented. However, mass sources are not boundary conditions in the sense of CFX; therefore, you may get a 0 [m/s] for hybrid, != 0 [m / s] as the control volume velocity, but not the injection velocity you specified. Hope the above helps,

May 31, 2015, 19:10		#14
jaimesdiegop New Member Diego Jaimes Join Date: Feb 2015 Posts: 8 Rep Power: 12	Hi Opaque I'm really interested in this CEL that you use "Velocity u.hybrid". saddly on my POST, when I use "function(Velocity u.hybrid)@location" CFX show: "The following unrecognised name was referenced: Velocity u.Boundcon" or "The following unrecognised name was referenced: Velocity u.hybrid" If there is some kind of trick to use, please let me know. Thank you, this could be the solution of one big problem that I have.