BC -> k e simpleFoam

soonic · April 13, 2012, 04:43

yes, I used the same approach to initialize.
If I wanted to drive the flow with 0.5Pa, I computed the velocity. The velocity would be ~0.9m/s. Check the Re, to identify the flow character. Then set the pressure BC value to p/rho = 0.5 / 1.23 to get velocity ~ 0.9m/s in pipe. I can be wrong. I am not an expert.

If you want to compare the solvers for the same k and € for OF and fluent, you should probably have a look at what values are coming to equations. That approach is the only one comparable. Who knows what correction is done before the solver gets value written as BC. As I saw, you thought about it already, and I have the same idea.

When you want to compare, you should give the boundaries correct values atleast in scale (as far you need to have physics correct on the BC). The value at BC will not change over the solution. It drives the solution to that value, so It have to be set correctly. Like the solution will go the way to have 0.5Pa on inlet faceCell and zero velocity on the wall. When you will have 1 for € and k, it will do everything to keep that values. When they are wrong, your solution diverges becouse of non-solvable case. For k and € there are several boundary types computing the value from length scale and intensity of turbulence. I do not know how to use them.

The initial value does not matter too much, it should be computed over domain, but strange value can push the solution to diverge too. Why not. Help the solution as much as you know

klio · April 13, 2012, 06:26

yes you are right !!

I GOT IT!!!!!!

@ lobstar: What are your k and eps in fluent and what is your hydraulic diameter D?

in my fluent simulation k and eps were both 1! Because of the formula e = c_nu ^(3/4)* (k^(3/2)/L) -> L = 0.07*D the values are only in right for a D of about 2.3 m. but if you have a D of 6mm like I do Open Foam means that this can not be possible and crashes or calculates wrong velocities. So it assumes that the user knows what he does

Fluent is that much intelligent that it can correct wrong initial contritions set by the user! You will get gradients were you can see the correction from input value to realistic value but then it works!

You can also set wrong k and e in OpenFoam and you will also get correcting gradients but the ratio of them has to fit to your geometry

HTH

klio · April 16, 2012, 04:45

@soonic: Do you have experience with heat transfer sim?

soonic · April 16, 2012, 14:13

A little. What about new thread ?

Describe what is your aim but first check some heat transfer tutorials and a solver suitable for your problem

klio · April 17, 2012, 02:25

Hi!

This is the link to the new thread!

http://www.cfd-online.com/Forums/ope...implefoam.html

But now just want to do a compressible flow but i really don't understand what p_rho does. But you will see in the new thread!

thanks a lot for your help!

lobstar · April 17, 2012, 06:33

Congrats Kilo! Glad you managed to work it out! I just wanted to ask what bc you used for k and e? I'm using the values calculated by fluent with fixedValue inlet, zeroGradient outlet and respective wall function for walls but still getting very large velocities! Is that what you used?

klio · April 18, 2012, 08:04

Quote:

Originally Posted by lobstar

Congrats Kilo! Glad you managed to work it out! I just wanted to ask what bc you used for k and e? I'm using the values calculated by fluent with fixedValue inlet, zeroGradient outlet and respective wall function for walls but still getting very large velocities! Is that what you used?

How did you get the values?

Because its not very clear how fluent calculates k and epsilon!

It is better you calculate them on your own!
http://en.wikipedia.org/wiki/Turbule...ary_conditions

my BC:

inlet --> fixed Value
outlet --> zeroGradient
wall --> WallFunction

klio · April 19, 2012, 02:59

Did you use totalPressure at Inlet?

lobstar · April 19, 2012, 03:59

Thanks for the tip on calculating k and eps!!! I had no idea before. I was using fixedValue for the inlet at the pressure. did you use/ would you recommend totalPressure?

klio · April 19, 2012, 09:58

Yes you have to take totalPressure because if you just fix the Value in field p you fix the static pressure. you can calculate the total pressure in the visualization tool with p0 = p*rho + rho/2 * abs(u_magnitude)²!

And because of that same k and eps have to be set as in fluent! so forget what i said before. it realy doesent matter what k an e you take when the ratio is ok for your geometry! It will just change the number of iterations you need for convergence and your sim time.

so just set inlet :
p0 *your totalPressure value*/rho;
gamma 1;

and you should also take the same converge settings as in fluent maybe your simulation stops to early (e & k : minimum 1e-4 )

lobstar · April 19, 2012, 10:16

Thanks Kilo! I really appreciate the help. I'll try your recommendations and let you know how it goes! Hope you're having some luck with your heat transfer case!

klio · April 19, 2012, 15:59

hmm works not that bad! It took me a little time to find out how to handle with atmospheric pressure but soonic gave me the input I needed to solve a simple case (flow in a pipe again!!!)

Tensian · July 2, 2013, 14:13

Hi everyone,

I was Reading your post and I think maybe you can help me, because I am trying to simulate a pressure driven flow across a complicated cannel: I don´t have any information about values over inlet and outlet, I only know that pressure inlet value is 0 Pa (due to atmospheric pressure) and pressure outlet is -20 Pa. I am running simpleFoam in laminar regime, and in a few iterations velocity and pressure explodes, with no physical meaning. It looks similar like the problema you are discussing right? (I don´t speak English very well so I don´t understand some parts of the conversation).
I tried changing fixedValue for pressureInletOutletVelocity, using pyFoamPotentialFlow for initialize (with errors)...and nothing makes it runs.
Any idea will be really apreciated, I am getting crazy! Fluent just do it inmediatly!
What do you think? (it looks easy but I can´t do it alone!)
Best regards.

izna · July 2, 2013, 15:10

hi
you need to insert a value of velocity in your model..

also you can read about how to manipulate the tutorials first.
i advice you to first start with the cavity tutorials and then the elbow tutorials. This would help you in understanding better how to edit Your project accordingly..

best of luck
izna

Tensian · July 3, 2013, 10:45

Hi Izna,

Thanks for the reply. I can´t give any value for velocity, because it´s not known!
Is really impossible to make the setup in terms of pressure difference between inlet and outlet? In FLUENT there´s no problem with these...
I can´t believe it. Maybe I have to implement something to make this possible...
Anymore ideas?

Cheers.

April 13, 2012, 06:26		#42
klio Member Klio Join Date: Mar 2012 Location: Austria Posts: 45 Rep Power: 14	yes you are right !! I GOT IT!!!!!! @ lobstar: What are your k and eps in fluent and what is your hydraulic diameter D? in my fluent simulation k and eps were both 1! Because of the formula e = c_nu ^(3/4)* (k^(3/2)/L) -> L = 0.07D the values are only in right for a D of about 2.3 m. but if you have a D of 6mm like I do Open Foam means that this can not be possible and crashes or calculates wrong velocities. So it assumes that the user knows what he does Fluent is that much intelligent that it can correct wrong initial contritions set by the user! You will get gradients were you can see the correction from input value to realistic value but then it works! You can also set wrong k and e in OpenFoam and you will also get correcting gradients but the ratio of them has to fit to your geometry HTH Last edited by klio; April 13, 2012 at 06:54.*

July 2, 2013, 14:13	bc for pressure driven flow	#53
Tensian Senior Member James Join Date: May 2013 Posts: 116 Rep Power: 12	Hi everyone, I was Reading your post and I think maybe you can help me, because I am trying to simulate a pressure driven flow across a complicated cannel: I don´t have any information about values over inlet and outlet, I only know that pressure inlet value is 0 Pa (due to atmospheric pressure) and pressure outlet is -20 Pa. I am running simpleFoam in laminar regime, and in a few iterations velocity and pressure explodes, with no physical meaning. It looks similar like the problema you are discussing right? (I don´t speak English very well so I don´t understand some parts of the conversation). I tried changing fixedValue for pressureInletOutletVelocity, using pyFoamPotentialFlow for initialize (with errors)...and nothing makes it runs. Any idea will be really apreciated, I am getting crazy! Fluent just do it inmediatly! What do you think? (it looks easy but I can´t do it alone!) Best regards.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Laminar simpleFoam and inviscid simpleFoam	herenger	OpenFOAM Running, Solving & CFD	7	July 11, 2013 06:27
Trying to run a benchmark case with simpleFoam	spsb	OpenFOAM	3	February 24, 2012 09:07
simpleFoam crash -> How to solve	tH3f0rC3	OpenFOAM	4	May 12, 2011 07:07
1.7.x -> buoyantPimpleFoam -> hRhoThermo -> incompressible and icoPoly3ThermoPhysics?	will.logie	OpenFOAM	0	December 16, 2010 07:08
Naca0012 k-e mpirun gives fpe whereas simpleFoam not	Pierpaolo	OpenFOAM	1	May 8, 2010 03:08

April 13, 2012, 04:43		#41
soonic New Member Peter Join Date: Apr 2010 Location: Brno, Czech Republic Posts: 24 Rep Power: 16	yes, I used the same approach to initialize. If I wanted to drive the flow with 0.5Pa, I computed the velocity. The velocity would be ~0.9m/s. Check the Re, to identify the flow character. Then set the pressure BC value to p/rho = 0.5 / 1.23 to get velocity ~ 0.9m/s in pipe. I can be wrong. I am not an expert. If you want to compare the solvers for the same k and € for OF and fluent, you should probably have a look at what values are coming to equations. That approach is the only one comparable. Who knows what correction is done before the solver gets value written as BC. As I saw, you thought about it already, and I have the same idea. When you want to compare, you should give the boundaries correct values atleast in scale (as far you need to have physics correct on the BC). The value at BC will not change over the solution. It drives the solution to that value, so It have to be set correctly. Like the solution will go the way to have 0.5Pa on inlet faceCell and zero velocity on the wall. When you will have 1 for € and k, it will do everything to keep that values. When they are wrong, your solution diverges becouse of non-solvable case. For k and € there are several boundary types computing the value from length scale and intensity of turbulence. I do not know how to use them. The initial value does not matter too much, it should be computed over domain, but strange value can push the solution to diverge too. Why not. Help the solution as much as you know

April 16, 2012, 04:45		#43
klio Member Klio Join Date: Mar 2012 Location: Austria Posts: 45 Rep Power: 14	@soonic: Do you have experience with heat transfer sim?

April 16, 2012, 14:13		#44
soonic New Member Peter Join Date: Apr 2010 Location: Brno, Czech Republic Posts: 24 Rep Power: 16	A little. What about new thread ? Describe what is your aim but first check some heat transfer tutorials and a solver suitable for your problem

April 17, 2012, 02:25		#45
klio Member Klio Join Date: Mar 2012 Location: Austria Posts: 45 Rep Power: 14	Hi! This is the link to the new thread! http://www.cfd-online.com/Forums/ope...implefoam.html But now just want to do a compressible flow but i really don't understand what p_rho does. But you will see in the new thread! thanks a lot for your help!

April 17, 2012, 06:33		#46
lobstar New Member Oluwalogbon Akinnola Join Date: Mar 2012 Posts: 23 Rep Power: 14	Congrats Kilo! Glad you managed to work it out! I just wanted to ask what bc you used for k and e? I'm using the values calculated by fluent with fixedValue inlet, zeroGradient outlet and respective wall function for walls but still getting very large velocities! Is that what you used?

April 19, 2012, 02:59		#48
klio Member Klio Join Date: Mar 2012 Location: Austria Posts: 45 Rep Power: 14	Did you use totalPressure at Inlet?

April 19, 2012, 03:59		#49
lobstar New Member Oluwalogbon Akinnola Join Date: Mar 2012 Posts: 23 Rep Power: 14	Thanks for the tip on calculating k and eps!!! I had no idea before. I was using fixedValue for the inlet at the pressure. did you use/ would you recommend totalPressure?

April 19, 2012, 09:58		#50
klio Member Klio Join Date: Mar 2012 Location: Austria Posts: 45 Rep Power: 14	Yes you have to take totalPressure because if you just fix the Value in field p you fix the static pressure. you can calculate the total pressure in the visualization tool with p0 = prho + rho/2 abs(u_magnitude)²! And because of that same k and eps have to be set as in fluent! so forget what i said before. it realy doesent matter what k an e you take when the ratio is ok for your geometry! It will just change the number of iterations you need for convergence and your sim time. so just set inlet : p0 your totalPressure value/rho; gamma 1; and you should also take the same converge settings as in fluent maybe your simulation stops to early (e & k : minimum 1e-4 )

April 19, 2012, 10:16		#51
lobstar New Member Oluwalogbon Akinnola Join Date: Mar 2012 Posts: 23 Rep Power: 14	Thanks Kilo! I really appreciate the help. I'll try your recommendations and let you know how it goes! Hope you're having some luck with your heat transfer case!

April 19, 2012, 15:59		#52
klio Member Klio Join Date: Mar 2012 Location: Austria Posts: 45 Rep Power: 14	hmm works not that bad! It took me a little time to find out how to handle with atmospheric pressure but soonic gave me the input I needed to solve a simple case (flow in a pipe again!!!)

July 2, 2013, 15:10		#54
izna Senior Member izna O'connor Join Date: Jun 2013 Posts: 143 Rep Power: 12	hi you need to insert a value of velocity in your model.. also you can read about how to manipulate the tutorials first. i advice you to first start with the cavity tutorials and then the elbow tutorials. This would help you in understanding better how to edit Your project accordingly.. best of luck izna

July 3, 2013, 10:45		#55
Tensian Senior Member James Join Date: May 2013 Posts: 116 Rep Power: 12	Hi Izna, Thanks for the reply. I can´t give any value for velocity, because it´s not known! Is really impossible to make the setup in terms of pressure difference between inlet and outlet? In FLUENT there´s no problem with these... I can´t believe it. Maybe I have to implement something to make this possible... Anymore ideas? Cheers.