[anita]

Hi,

my problem is the 2D simulation of a flow around a bluff body in a pipe. The inlet velocity should be a fully developed turbulent velocity profile.

Is it possible to map the profile generate in an other case (empty pipe with uniform inlet velocity) to the inlet boundary?
Or is there a analytic solution describing the profile and a way to set the inlet velocity by this equation?

Anita

[cedric_duprat]

Hi Anita,

Why don't you use the directMappedPatch for your inlet boundary condition ?
The idea is to designate as inlet region a "region" of your main computational domain. At the end of this "region", the flow are sampled and fed back into the start computational domain. This cyclic region will become fully turbulent quite faster in 2D.
So, it should be faster than read a file from previous calculation, and interpolate it as inlet, shouldn't it ?

Hope it helps,

Cedric

Cedric,

Where did you find the information about directMappedPatch? I'm trying to build up a simple airfoil simulation and I'm having a devil of a time getting the initial conditions set up correctly.

How do you find out what values a patch will accept? I've looked through the programmers guide and the users guide and I can't find good specifications for what to put in the U, p, k and epsilon files for turbFoam.

Joe

[cedric_duprat]

Hi Joe,

You can find a tutorial on the forum there:
http://www.cfd-online.com/OpenFOAM_D...html#POST13069

and then there are some comments there from Eugene:
http://www.cfd-online.com/OpenFOAM_D...html#POST17174

Then, if you need more documents, still from Eugene, you can read his Ph'D thesis where the basic idea is explained.

In fact, according to me, there is no link between the directMappedPatch and your case. The reason is that it is a patch to create turbulent structure in a turbulent flow. Usualy, in an airfoil simulation, there are no such structure because the rate of kinetic energy should be low. But, maybe I'm wrong.

Then, to fixed k and epsilon in a k-epsilon turbulent calculation, there is one tool to get epsilon from k on the wiki and I'm sure that there is an example in the user guide somewhere.

Hope it will help you.

Cedric

Cedric,

Thank you. Your hints have proved invaluable. Thank you.

The problem I was having was that people would post meshes to the Internet for which I am grateful. However, no one ever seems to post the contents of the 0/ directory which are just as important as the mesh itself. I found an airfoil mesh that would do what I need but provided no initial conditions. My attempts to build up initial conditions has met with failure because I couldn't find out which wall types took what kind of conditions/types/values/etc etc.

The forum links you posted gave me some hints as to where I could find the configuration options. In the source directory for turbFoam I found a turbFoam.cfg file. Horray! The turbFoam.cfg file listed out all the configuration options (or where to find them) I need to build up a valid set of initial conditions.

I wish I didn't have to dig through source to find this info but oh well.

The configuration options for each solver can be found in the <<solver>>.cfg file. For turbfoam it was U, p, epsilon, k, nutilda and Rkinematic.

Figured out what was wrong. The problem was in 'U', not 'p'. Sorry for the trouble. My simulation is finally working!