velocity-dependent diffusivity

porterma · August 17, 2009, 15:01

I am using star-cd (version 4.06). I need to define a velocity-dependent diffusivity coefficient for a passive scalar. I have worked out that I need to write a subroutine using pordif.f. What I can't figure out is how I access the magnitude of the velocity that has already been calculated by star-cd. My calculation is very simple:

PORODF = 0.0522*v + 1e-8

where v is the velocity magnitude at every location in my volume. It seems like should be very easy, but I can not find any documentation regarding this matter.

Thanks in advance.

Pauli · August 17, 2009, 19:58

At the top of pordif.f you should have this line:
INCLUDE 'usrdat.inc'
It will provide U,V,W in global cys. See nom.inc for details.

usrdat.inc & nom.inc will appear in your ufile directory when you compile (star -ufile).

porterma · November 5, 2009, 13:18

Thank you for the reply Pauli. I know this is very late, but we had some license issues here and I haven't been able to use star-cd in awhile. Thus, I now return to my problem. I implemented the velocity-dependent diffusivity for each region in my model domain as follows:

Code:

IF(NPPOR.EQ.1) THEN
  PORODF = 0.00152*SQRT(UL**2 + VL**2 + WL**2) + 3.6431E-10 
ENDIF

I have one question regarding this implementation. I wasn't sure if UL, VL and WL are the appropriate velocity components or if I should be using U, V and W. In porous media flows, there is a superficial velocity and a pore velocity; they are related by the porosity (i.e., pore velocity equals the superficial velocity divided by the porosity). Based on the description of each I could not determine if if U or UL was the pore velocity. Does anyone know which is which.

My other concern is that in porous media flow the diffusivity becomes a tensor (similar the alpha and beta tensors that are implemented in star-cd). But the diffusivity is treated as a scalar. In the code above I technically need a PORODF for each component (x,y,z) of flow. It appears that I need additional if statements that identify each component of flow. For example, the if statements should look something like this:

Code:

IF(???.EQ.1.AND.NPPOR.EQ.1) THEN
  PORODF = 0.00152*UL + 3.6431E-10 
ENDIF
IF(???.EQ.2.AND.NPPOR.EQ.1) THEN
  PORODF = 0.00152*VL/10 + 3.6431E-10 
ENDIF
IF(???.EQ.3.AND.NPPOR.EQ.1) THEN
   PORODF = 0.00152*WL/10 + 3.6431E-10 
ENDIF

The problem here is that I can't find the variable (the ??? in the above pseudo code) used to identify each component of flow?

Thanks in advance.
porter

Pauli · November 5, 2009, 14:03

U is velocity in the global CSYS. UL is the velocity in the porous media CSYS. I don't know if it is superficial or otherwise.

It looks like diffusivity is implemented as isotropic (i.e. a scalar). I have no idea how you can make it otherwise. I don't think your psuedo-code would work.

Both are good questions for CD-adapco support.

August 17, 2009, 15:01	velocity-dependent diffusivity	#1
porterma New Member Mark Porter Join Date: Aug 2009 Posts: 2 Rep Power: 0	I am using star-cd (version 4.06). I need to define a velocity-dependent diffusivity coefficient for a passive scalar. I have worked out that I need to write a subroutine using pordif.f. What I can't figure out is how I access the magnitude of the velocity that has already been calculated by star-cd. My calculation is very simple: PORODF = 0.0522*v + 1e-8 where v is the velocity magnitude at every location in my volume. It seems like should be very easy, but I can not find any documentation regarding this matter. Thanks in advance.

November 5, 2009, 13:18	direction-dependent difusivity	#3
porterma New Member Mark Porter Join Date: Aug 2009 Posts: 2 Rep Power: 0	Thank you for the reply Pauli. I know this is very late, but we had some license issues here and I haven't been able to use star-cd in awhile. Thus, I now return to my problem. I implemented the velocity-dependent diffusivity for each region in my model domain as follows: Code: IF(NPPOR.EQ.1) THEN PORODF = 0.00152SQRT(UL2 + VL2 + WL2) + 3.6431E-10 ENDIF I have one question regarding this implementation. I wasn't sure if UL, VL and WL are the appropriate velocity components or if I should be using U, V and W. In porous media flows, there is a superficial velocity and a pore velocity; they are related by the porosity (i.e., pore velocity equals the superficial velocity divided by the porosity). Based on the description of each I could not determine if if U or UL was the pore velocity. Does anyone know which is which. My other concern is that in porous media flow the diffusivity becomes a tensor (similar the alpha and beta tensors that are implemented in star-cd). But the diffusivity is treated as a scalar. In the code above I technically need a PORODF for each component (x,y,z) of flow. It appears that I need additional if statements that identify each component of flow. For example, the if statements should look something like this: Code: IF(???.EQ.1.AND.NPPOR.EQ.1) THEN PORODF = 0.00152UL + 3.6431E-10 ENDIF IF(???.EQ.2.AND.NPPOR.EQ.1) THEN PORODF = 0.00152VL/10 + 3.6431E-10 ENDIF IF(???.EQ.3.AND.NPPOR.EQ.1) THEN PORODF = 0.00152WL/10 + 3.6431E-10 ENDIF The problem here is that I can't find the variable (the ??? in the above pseudo code) used to identify each component of flow? Thanks in advance. porter

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
time dependent velocity input by user subroutine	shohel mahmud	Siemens	4	December 18, 2007 11:40
Velocity in Porous medium : HELP! HELP! HELP!	Kali Sanjay	Phoenics	0	November 6, 2006 06:10
A velocity dependent kepsilon boundary condition	tangd	OpenFOAM Pre-Processing	2	September 15, 2006 07:04
Variables Definition in CFX Solver 5.6	R P	CFX	2	October 26, 2004 02:13
Terrible Mistake In Fluid Dynamics History	Abhi	Main CFD Forum	12	July 8, 2002 09:11

August 17, 2009, 19:58		#2
Pauli Senior Member Pauli Join Date: Mar 2009 Posts: 189 Rep Power: 17	At the top of pordif.f you should have this line: INCLUDE 'usrdat.inc' It will provide U,V,W in global cys. See nom.inc for details. usrdat.inc & nom.inc will appear in your ufile directory when you compile (star -ufile).

November 5, 2009, 14:03		#4
Pauli Senior Member Pauli Join Date: Mar 2009 Posts: 189 Rep Power: 17	U is velocity in the global CSYS. UL is the velocity in the porous media CSYS. I don't know if it is superficial or otherwise. It looks like diffusivity is implemented as isotropic (i.e. a scalar). I have no idea how you can make it otherwise. I don't think your psuedo-code would work. Both are good questions for CD-adapco support.