CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   OpenFOAM (http://www.cfd-online.com/Forums/openfoam/)
-   -   logarithmic profile velocity - inlet BC (http://www.cfd-online.com/Forums/openfoam/119328-logarithmic-profile-velocity-inlet-bc.html)

pedroxramos June 14, 2013 09:58

logarithmic profile velocity - inlet BC
 
Hello Openfoamers,

I want to define a logarithmic velocity profileto my open channel flow (like this: http://www.fhwa.dot.gov/engineering/...ges/fig4_3.gif
What do you suggest? Groovy BC? I have some problems to install it with OF 2.0.

There is a simpler way?

Best regards,

Lieven June 14, 2013 15:06

If you are working with a RANS turbulence model, you should be able to use the atmBoundaryLayerInletVelocity which simply imposes the logarithmic velocity profile.

Cheers,

L

shash June 15, 2013 05:09

hi pedro,

What about expanding the log and using codedFixedValue (http://www.cfd-online.com/Forums/ope...condition.html) , as i am a newbie my knowhow is limited too. Expect this may help :D

wyldckat June 15, 2013 11:25

Greetings to all!

@Pedro: OpenFOAM 2.0 or 2.2.0?
Quote:

Originally Posted by pedroxramos (Post 434024)
Groovy BC? I have some problems to install it with OF 2.0.

If it's OpenFOAM 2.2.0, read the instructions on this post: http://www.cfd-online.com/Forums/ope...tml#post415028 post #16

Best regards,
Bruno

pedroxramos June 17, 2013 05:08

Hello.

Thanks. Yes, It's OF 2.2.0. I'll try install it and input the log profile with swak4foam.

Regards,

pedroxramos June 17, 2013 05:30

Quote:

Originally Posted by Lieven (Post 434053)
If you are working with a RANS turbulence model, you should be able to use the atmBoundaryLayerInletVelocity which simply imposes the logarithmic velocity profile.

Cheers,

L

Thanks, but I'm using InterFoam with LES model.

Lieven June 17, 2013 10:49

Ok, completely different situation.

The logarithmic velocity profile is the profile a time averaged profile. This profile does not exist as instantaneous profile so it is physically not correct to impose the log-profile directly at the inlet.

the most correct thing to do is use an 'infinite inlet' to generate profiles which result in the log-profile when they are time averaged. Two ways to do this:
1. Do a separate simulation of a cyclic channel and store every inlet
2. Increase the length of the inlet and use the mapped BC at the inlet.

Cheers,

Lieven

pedroxramos June 19, 2013 06:06

1 Attachment(s)
Quote:

Originally Posted by Lieven (Post 434467)
Ok, completely different situation.

The logarithmic velocity profile is the profile a time averaged profile. This profile does not exist as instantaneous profile so it is physically not correct to impose the log-profile directly at the inlet.

the most correct thing to do is use an 'infinite inlet' to generate profiles which result in the log-profile when they are time averaged. Two ways to do this:
1. Do a separate simulation of a cyclic channel and store every inlet
2. Increase the length of the inlet and use the mapped BC at the inlet.

Cheers,

Lieven

Hi Lieven, thanks for your help.

Which boundary condition should I put in the bed to to the second suggestion that you gave me? I put the no slip condition, like this:

bed - U
{
type fixedValue;
value uniform (0 0 0);
}

Attachment 22815

And the contour of the velocity in the midplane of the channel is: http://d.pr/i/ZWvo

http://d.pr/i/ZWvohttp://d.pr/i/ZWvo
The distance between the inlet and the pille is 1 meter. Should I keep the bed boundary condition and increase that distance?

Best regards.

Lieven June 19, 2013 08:22

Hi Pedro,


You should use the 'mapped' boundary condition for that. Have for example a look at
http://www.cfd-online.com/Forums/ope...condition.html
If you google a bit or search this forum, I'm sure you will find more information on it if the link about is not enough.

Cheers,


Lieven


All times are GMT -4. The time now is 07:00.