help me to set suitable outlet boundary condition

shaswat · July 20, 2013, 05:06

Quote:

Originally Posted by ghorrocks

That has to start with you - you say fluid goes through the porous wall all over. So what controls it? You need some function to drive it. A defined flux? Maybe a concentration gradient? Maybe a constant value? So "some function" is a vague reference to the wide variety of functions you can use to define this flow.

Thank you Dr

I got the simulation what I expected. Now the the simulation is running

Thank you for your kind advice and help

Thank you

shaswat · July 28, 2013, 11:28

Dear all

Please find the attached cross sectional view of artery.The artery length is 100mm
The outer part is porous domain. I introduced free slip between fluid and porous interface . I set fluid - porous interface by using GGI.when I run the simulation I saw momentum and mass -2 is not at all executed . please clarify

Thank you

ghorrocks · July 28, 2013, 20:43

What is momentum and mass -2? Is that fluid flow in the porous domain?

shaswat · July 28, 2013, 22:46

Quote:

Originally Posted by ghorrocks

What is momentum and mass -2? Is that fluid flow in the porous domain?

I think it is fluid flow in the porous domain

ghorrocks · July 29, 2013, 01:13

So it looks like it is not solving the fluid equations in the porous region. Can you post your CCL?

shaswat · July 29, 2013, 05:22

Quote:

Originally Posted by ghorrocks

So it looks like it is not solving the fluid equations in the porous region. Can you post your CCL?

Dear Dr

Please find the attached my CCL .

I don't know what is the problem
I need to solve this as soon as possible . Please help me in this regards.

Thank you

Reagrds

ghorrocks · July 29, 2013, 09:04

Where do you get the time step size from? Did you actually do something to show that time step was required or did you just guess?

You have your artery wall set to solid morphology. You probably want this porous (I am not sure about that).

You have the mass momentum model as free slip on the interface. You will want this to be no slip.

Why have you set a max coeff loops of 3? And why a minimum of 1? Remove the min loops and make the max loops something like 10.

Do you need the expert parameter? Have you checked you need it?

I would simplify this model to get the components working. I would model the arterey only (fluid flow only, and the fluid is a newtonian fluid) to make sure the time step and boundaries are working. Then add the porous wall. When that works add the non-newtonian fluid model.

shaswat · July 29, 2013, 11:19

Quote:

Originally Posted by ghorrocks

Where do you get the time step size from? Did you actually do something to show that time step was required or did you just guess?

I just guess . I want to show how the shear stress and other parameters changes with respect to time

Quote:

Originally Posted by ghorrocks

You have your artery wall set to solid morphology. You probably want this porous (I am not sure about that).

Yes . I want this as a solid morphology having a porous in nature.
Is it wrong to define solid here? if I remove, will my result vary or not?

Quote:

Originally Posted by ghorrocks

You have the mass momentum model as free slip on the interface. You will want this to be no slip.

I have seen many articles they are using free slip boundary at the interface

Quote:

Originally Posted by ghorrocks

Why have you set a max coeff loops of 3? And why a minimum of 1? Remove the min loops and make the max loops something like 10.

Ok I will do

Quote:

Originally Posted by ghorrocks

Do you need the expert parameter? Have you checked you need it?

This part I am not sure. your advice is highly needed.

Thank you

ghorrocks · July 29, 2013, 22:57

Time Step: Do not guess, invariably you will get it wrong. Use adaptive time stepping, with 3-5 coeff loops per iteration. Then the solver will find the correct time step size.

Free slip: sure, you can use free slip but is that what you want? Then you will not get any realistic flow profile in the artery.

Do not put expert parameters in unless you know you need them and you know what they are doing. They are not called expert parameters for nothing.

shaswat · August 1, 2013, 02:39

Quote:

Originally Posted by ghorrocks

Time Step: Do not guess, invariably you will get it wrong. Use adaptive time stepping, with 3-5 coeff loops per iteration. Then the solver will find the correct time step size.

Free slip: sure, you can use free slip but is that what you want? Then you will not get any realistic flow profile in the artery.

Do not put expert parameters in unless you know you need them and you know what they are doing. They are not called expert parameters for nothing.

I am facing the same problem . One cycle completed but there is no fluid flow inside the porous domain.

I changed from transient to steady state analysis, still I did't get.

can I assume porous domain initialization with free slip.

Thank you

ghorrocks · August 1, 2013, 03:28

Slip or no slip should not matter on the porous interface. If you are getting no flow in the porous region you have a more fundamental problem with your simulation.

I note your permeability is 2e-18[m^2]. I am no expert in porous flows but this sounds pretty low. Wouldn't that pretty much stop flow in the porous region?

shaswat · August 1, 2013, 09:32

Quote:

Originally Posted by ghorrocks

Slip or no slip should not matter on the porous interface. If you are getting no flow in the porous region you have a more fundamental problem with your simulation.

I note your permeability is 2e-18[m^2]. I am no expert in porous flows but this sounds pretty low. Wouldn't that pretty much stop flow in the porous region?

I am also thinking . I tried with permeability 1[m^2] . I could not get a flow in the porous domain.

Since I am using turbulent model in the main flow. when the flow enters into the porous region it would be laminar flow . How to handle this ?

Thank you

ghorrocks · August 1, 2013, 19:04

Rather than randomly trying permabilities, how about working out what the permability actually is?

shaswat · August 1, 2013, 19:27

Quote:

Originally Posted by ghorrocks

Rather than randomly trying permabilities, how about working out what the permability actually is?

IT is actually 2e-18[m^2]

ghorrocks · August 1, 2013, 21:42

OK - so with a permability so low, will you get any flow?

In post #27 I recommended you simplify the model to get the components working. Have you done this?

shaswat · August 2, 2013, 03:30

Quote:

Originally Posted by ghorrocks

OK - so with a permability so low, will you get any flow?

In post #27 I recommended you simplify the model to get the components working. Have you done this?

Yes Dr.
It works fine . When I add a porous layer the momentum and mass for porous region not at all showing any response.
I now changed to transient to steady flow . I set 200 iteration. I could not not see any flow in the porous region.
Now , I am thinking to initialize the porous domain with Cartesian velocity components . But Don't know how to implement with out knowing velocity . Any suggestion

ghorrocks · August 2, 2013, 09:25

I have already said my suggestion two times - so here it is for a third time. Have you simplified your model (ie just use a newtonian fluid, and a simple geometry) to test that the porous material works as expected for a simple case?

shaswat · August 3, 2013, 07:31

Quote:

Originally Posted by ghorrocks

I have already said my suggestion two times - so here it is for a third time. Have you simplified your model (ie just use a newtonian fluid, and a simple geometry) to test that the porous material works as expected for a simple case?

I tested with straight pipe . the wall I consider porous . I simulate with Newtonian flow . I see the same result when I use non Newtonian fluid. Inside the porous domain momentum and mass equation is not solving. I applied free slip boundary condition.

I have a question . at the domain interface is it necessary to introduce source terms

Thank you

ghorrocks · August 3, 2013, 09:48

No, you should not source terms to model what I understand you are trying to model.

As the porous model is not working on a simpler model I would concentrate on getting it working on the simple model before going back to the full model. I do not know why it is not working for you, but try these things:
1) Try the porous region as a subdomain of the fluid domain rather than a separate domain. This should not need a GGI so you will have to remove that.
2) Change the porosity model options, like loss velocity type, the expert parameter and any others which look interesting. No harm in trying everything.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Outlet boundary condition for wave flume with interFoam solver	Arnoldinho	OpenFOAM	9	July 10, 2018 06:15
Setting outlet Pressure boundary condition using CAFFA code	Mukund Pondkule	Main CFD Forum	0	March 16, 2011 04:23
Outlet boundary condition	colen	CFX	6	March 8, 2010 23:49
VOF Outlet boundary condition in cfd - ace	JM	Main CFD Forum	0	December 15, 2006 09:07
Convective Heat Transfer - Heat Exchanger	Mark	CFX	6	November 15, 2004 16:55

July 28, 2013, 20:43		#23
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	What is momentum and mass -2? Is that fluid flow in the porous domain?

July 29, 2013, 01:13		#25
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	So it looks like it is not solving the fluid equations in the porous region. Can you post your CCL?

July 29, 2013, 09:04		#27
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	Where do you get the time step size from? Did you actually do something to show that time step was required or did you just guess? You have your artery wall set to solid morphology. You probably want this porous (I am not sure about that). You have the mass momentum model as free slip on the interface. You will want this to be no slip. Why have you set a max coeff loops of 3? And why a minimum of 1? Remove the min loops and make the max loops something like 10. Do you need the expert parameter? Have you checked you need it? I would simplify this model to get the components working. I would model the arterey only (fluid flow only, and the fluid is a newtonian fluid) to make sure the time step and boundaries are working. Then add the porous wall. When that works add the non-newtonian fluid model.

July 29, 2013, 22:57		#29
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	Time Step: Do not guess, invariably you will get it wrong. Use adaptive time stepping, with 3-5 coeff loops per iteration. Then the solver will find the correct time step size. Free slip: sure, you can use free slip but is that what you want? Then you will not get any realistic flow profile in the artery. Do not put expert parameters in unless you know you need them and you know what they are doing. They are not called expert parameters for nothing.

August 1, 2013, 03:28		#31
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	Slip or no slip should not matter on the porous interface. If you are getting no flow in the porous region you have a more fundamental problem with your simulation. I note your permeability is 2e-18[m^2]. I am no expert in porous flows but this sounds pretty low. Wouldn't that pretty much stop flow in the porous region?

August 1, 2013, 19:04		#33
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	Rather than randomly trying permabilities, how about working out what the permability actually is?

August 1, 2013, 21:42		#35
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	OK - so with a permability so low, will you get any flow? In post #27 I recommended you simplify the model to get the components working. Have you done this?

August 2, 2013, 09:25		#37
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	I have already said my suggestion two times - so here it is for a third time. Have you simplified your model (ie just use a newtonian fluid, and a simple geometry) to test that the porous material works as expected for a simple case?

August 3, 2013, 09:48		#39
ghorrocks Super Moderator Glenn Horrocks Join Date: Mar 2009 Location: Sydney, Australia Posts: 17,870 Rep Power: 144	No, you should not source terms to model what I understand you are trying to model. As the porous model is not working on a simpler model I would concentrate on getting it working on the simple model before going back to the full model. I do not know why it is not working for you, but try these things: 1) Try the porous region as a subdomain of the fluid domain rather than a separate domain. This should not need a GGI so you will have to remove that. 2) Change the porosity model options, like loss velocity type, the expert parameter and any others which look interesting. No harm in trying everything.