Dirichlet boundary condition for additional variable on the wall

ftab · July 23, 2012, 12:05

Hi,

In a simulation, I have been asked to put the unity Dirichlet Boundary condition on the wall for concentration.

I have defined volumetric additional variable(AV) with dimension of kg m^-3 and set up the case. The problem is how to set the boundary condition of unit concentration on the wall? CFX asks for constant value of AV (kgm^-3), but obviously the wall is a surface not a volume. Hence, I am puzzled with physical meaning of it and how to apply a boundary condition which would be equivalent to Dirichlet BC of unity on the wall.
In fluent I remember, we simply put the constant value of UDS equal to 1. What would be the setting to get similar results in CFX?

Thanks!

Felggv · July 23, 2012, 14:42

Have you tried creating a Source Point on the walls?

Don't know if it will serve you... just a hint.

ghorrocks · July 23, 2012, 19:39

Felipe is correct, you can do this using a source term on the boundary surface. That will work.

ftab · July 24, 2012, 08:12

Thanks Felipe and Glenn for reply.
If I need to use the source for this boundary, what should be the setting in Boundary details then? Value (kg m^-3) or Flux equal to zero?
Setting the source in dimension of kg m^-2 s^-1 in a steady solution will introduce huge amount of additional variable (concentration) inside the domain and having seudo transient solution approach of CFX for steady cases, the concentration goes up to 10^8 or so!!
Thanks for your help

Felggv · July 24, 2012, 11:37

Hello,

In my opinion, Steady State simulations are way too harder to model than transient ones.

Would it be valid to do a transient run and come up with an expression that would cease the source of your contaminant in time just to study this case in a simpler model before running it in steady state?

Just for us to understand, when you set your wall with boundary sources, which options do you have? Would you attach a picture?

What is you material?

Bye

ftab · July 24, 2012, 12:00

Hi Felipe,
It is a simple cylindrical artery with stent inside, which is eluting drug both inside the wall and lumen of the artery.
Check this image for instance:
http://www.vizworld.com/2010/06/mit-...create-stents/

or this one:
http://maths.dur.ac.uk/~dma0mpj/summ...lo_Zunino.html

To be honest, I have not set the boundary source, but the flux in boundary details, which has exactly the same dimension and got the stupid results as mentioned.

Felggv · July 24, 2012, 14:45

Well,

What are you trying to study? How it mixes with blood flow? The only information you have is the concentration of the drug and it only goes to the blood flow by concentration difference?

If you have more information you can model it some other way.

Tell us more about your problem, now that we see the picture we can understand better what you want to do.

Bye

ftab · July 24, 2012, 18:25

Felipe,
The purpose is simply to model the transport of drug(concentration AV) from the stent into blood flow and wall both with convection and diffusion.
And due to the concentration difference (1 on the stent and 0 in wall and lumen) the drug is transported. The equations which are being solved are Mass and momentum together with AV transport.
The flow is simply laminar, incomp and steady for now.

ghorrocks · July 24, 2012, 20:26

Are you sure the Dirichlet BC is the appropriate one? Would a constant flux be better? Or what about a convective type BC, but based on AV concentration, not temperature?

ftab · July 25, 2012, 07:11

Glenn,
you are right! If it was transient solution it would fit. Even for the steady one, but as I said I am trying to replicate a reference paper who has used this BC.

Felggv · July 25, 2012, 13:14

ftab,

As Glenn always recommends, begin with a simpler model, that's why I asked you if you only had the concentration, because if you had the flux it would be easier to replicate the situation!

Even if that's not what you want, other BCs should do the job.

Good luck

ghorrocks · July 25, 2012, 19:31

If you are sure you need a dirichlet condition on AV concentration then here's how to do it

* Set a source term for the AV equation on the boundary to impose the condition on.
* Set the source term equation to -1.0e3*(1.0-(AV concentration variable) and the source term coefficient to 1.0e3. Please note I am recalling this equation from memory so might have got some signs wrong or other little error, but that is the basic idea.

Where 1.0e3 is a "big" number relative to your AV concentration, 1.0 is the value you want for the dirichlet boundary condition and "AV concentration variable" is the AV concentration variable however you have defined it.

This uses a source term to pull the local value to the set value (1.0), and the source term coefficient helps convergence.

bitak · January 27, 2013, 13:23

Hello.
I have a problem like you.

I want to reproduce an article but I don't know how Should put a boundary condition for wall!!!!!!!!!!!

my geometry is a tube( d=7mm and L=15.4 cm ).

Boundary conditions for concentration of LDL at the inlet, outlet and wall, sequentially, are as follows:
C(0, y) = C0 ( inlet concentration)

@ outlet flux = 0

on the wal : CwVw (wall concentration * infilteration velocity)
I really be happy and will be thanks if you answer to me.

bitak · January 27, 2013, 13:24

How I can source term?

July 23, 2012, 12:05	Dirichlet boundary condition for additional variable on the wall	#1
ftab Member Ftab Join Date: Sep 2011 Posts: 87 Rep Power: 15	Hi, In a simulation, I have been asked to put the unity Dirichlet Boundary condition on the wall for concentration. I have defined volumetric additional variable(AV) with dimension of kg m^-3 and set up the case. The problem is how to set the boundary condition of unit concentration on the wall? CFX asks for constant value of AV (kgm^-3), but obviously the wall is a surface not a volume. Hence, I am puzzled with physical meaning of it and how to apply a boundary condition which would be equivalent to Dirichlet BC of unity on the wall. In fluent I remember, we simply put the constant value of UDS equal to 1. What would be the setting to get similar results in CFX? Thanks!

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Water subcooled boiling	Attesz	CFX	7	January 5, 2013 04:32
vorticity boundary condition	bearcharge	Main CFD Forum	0	May 14, 2010 12:32
Variable temperature wall boundary condition	Virginie	FLUENT	2	October 2, 2007 10:23
UDF hook on the WALL boundary condition	Luke	FLUENT	0	June 7, 2006 12:54

July 23, 2012, 14:42		#2
Felggv Member Felipe Gobbi Join Date: Apr 2012 Location: Brazil Posts: 76 Rep Power: 14	Have you tried creating a Source Point on the walls? Don't know if it will serve you... just a hint.

July 23, 2012, 19:39		#3
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	Felipe is correct, you can do this using a source term on the boundary surface. That will work.

July 24, 2012, 08:12		#4
ftab Member Ftab Join Date: Sep 2011 Posts: 87 Rep Power: 15	Thanks Felipe and Glenn for reply. If I need to use the source for this boundary, what should be the setting in Boundary details then? Value (kg m^-3) or Flux equal to zero? Setting the source in dimension of kg m^-2 s^-1 in a steady solution will introduce huge amount of additional variable (concentration) inside the domain and having seudo transient solution approach of CFX for steady cases, the concentration goes up to 10^8 or so!! Thanks for your help

July 24, 2012, 11:37		#5
Felggv Member Felipe Gobbi Join Date: Apr 2012 Location: Brazil Posts: 76 Rep Power: 14	Hello, In my opinion, Steady State simulations are way too harder to model than transient ones. Would it be valid to do a transient run and come up with an expression that would cease the source of your contaminant in time just to study this case in a simpler model before running it in steady state? Just for us to understand, when you set your wall with boundary sources, which options do you have? Would you attach a picture? What is you material? Bye

July 24, 2012, 12:00		#6
ftab Member Ftab Join Date: Sep 2011 Posts: 87 Rep Power: 15	Hi Felipe, It is a simple cylindrical artery with stent inside, which is eluting drug both inside the wall and lumen of the artery. Check this image for instance: http://www.vizworld.com/2010/06/mit-...create-stents/ or this one: http://maths.dur.ac.uk/~dma0mpj/summ...lo_Zunino.html To be honest, I have not set the boundary source, but the flux in boundary details, which has exactly the same dimension and got the stupid results as mentioned.

July 24, 2012, 14:45		#7
Felggv Member Felipe Gobbi Join Date: Apr 2012 Location: Brazil Posts: 76 Rep Power: 14	Well, What are you trying to study? How it mixes with blood flow? The only information you have is the concentration of the drug and it only goes to the blood flow by concentration difference? If you have more information you can model it some other way. Tell us more about your problem, now that we see the picture we can understand better what you want to do. Bye

July 24, 2012, 18:25		#8
ftab Member Ftab Join Date: Sep 2011 Posts: 87 Rep Power: 15	Felipe, The purpose is simply to model the transport of drug(concentration AV) from the stent into blood flow and wall both with convection and diffusion. And due to the concentration difference (1 on the stent and 0 in wall and lumen) the drug is transported. The equations which are being solved are Mass and momentum together with AV transport. The flow is simply laminar, incomp and steady for now.

July 24, 2012, 20:26		#9
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	Are you sure the Dirichlet BC is the appropriate one? Would a constant flux be better? Or what about a convective type BC, but based on AV concentration, not temperature?

July 25, 2012, 07:11		#10
ftab Member Ftab Join Date: Sep 2011 Posts: 87 Rep Power: 15	Glenn, you are right! If it was transient solution it would fit. Even for the steady one, but as I said I am trying to replicate a reference paper who has used this BC.

July 25, 2012, 13:14		#11
Felggv Member Felipe Gobbi Join Date: Apr 2012 Location: Brazil Posts: 76 Rep Power: 14	ftab, As Glenn always recommends, begin with a simpler model, that's why I asked you if you only had the concentration, because if you had the flux it would be easier to replicate the situation! Even if that's not what you want, other BCs should do the job. Good luck

July 25, 2012, 19:31		#12
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,871 Rep Power: 144	If you are sure you need a dirichlet condition on AV concentration then here's how to do it * Set a source term for the AV equation on the boundary to impose the condition on. * Set the source term equation to -1.0e3*(1.0-(AV concentration variable) and the source term coefficient to 1.0e3. Please note I am recalling this equation from memory so might have got some signs wrong or other little error, but that is the basic idea. Where 1.0e3 is a "big" number relative to your AV concentration, 1.0 is the value you want for the dirichlet boundary condition and "AV concentration variable" is the AV concentration variable however you have defined it. This uses a source term to pull the local value to the set value (1.0), and the source term coefficient helps convergence.

January 27, 2013, 13:23		#13
bitak Member bitak Join Date: Jul 2009 Posts: 56 Rep Power: 17	Hello. I have a problem like you. I want to reproduce an article but I don't know how Should put a boundary condition for wall!!!!!!!!!!! my geometry is a tube( d=7mm and L=15.4 cm ). Boundary conditions for concentration of LDL at the inlet, outlet and wall, sequentially, are as follows: C(0, y) = C0 ( inlet concentration) @ outlet flux = 0 on the wal : CwVw (wall concentration * infilteration velocity) I really be happy and will be thanks if you answer to me.

January 27, 2013, 13:24		#14
bitak Member bitak Join Date: Jul 2009 Posts: 56 Rep Power: 17	How I can source term?