CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   OpenFOAM (http://www.cfd-online.com/Forums/openfoam/)
-   -   p & U outlet bcs for realistic wind inlet profile (http://www.cfd-online.com/Forums/openfoam/78261-p-u-outlet-bcs-realistic-wind-inlet-profile.html)

stevek July 16, 2010 08:51

p & U outlet bcs for realistic wind inlet profile
 
2 Attachment(s)
Hello,

I am running simpleFoam with realistic wind profile inlet for flow over terrain and I am having problems with outlet BCs. As can be seen from the attached images, the flow is not being allowed to exit the domain and there seems to be an inconsistency with how I'm dealing with the pressure.

The simulation converges to 10e-4, I have turbulence turned off for now and the domain is a simple box. Inlet is on the left, outlet on the right, top and other sides are symmetry planes and the bottom is rough terrain (no turbulence right now). Also, I am using first order convection scheme.

My bcs are:
Inlet:
U: specified wind profile
p: zeroGradient

Outlet:
U: inletOutlet
inletValue (0 0 0)
value (0 0 0)
p: fixedMeanValue
meanValue 0
(on another post Hrvoje Jasak mentions this is a nice option for stratified outlet profiles ... am I using it correctly in this sense)

So my question is: for this case what would be appropriate outlet conditions for velocity and pressure to remove this problem? I have tried several combinations but no luck. (I am running 1.7.0 with same extra bcs picked up along the way)

Thanks,
Steve

pmarek July 16, 2010 12:29

Hi,

I had a similar issue but it seemed to be caused by too high convergence setting. I basically lowered it so the simulation ran for longer and the flow 'went through' the whole domain. The pressure build up seems to be only at the bottom of it, so I'm not too sure whether the flow is not able to leave it (I think it is actually able to leave it).

BTW, how did you get a nonlinear inlet velocity profile in OF 1.7? I used groovyBC before but OF 1.7 doesn't seem to like it too much.

Przemek

stevek July 16, 2010 12:36

Hi przemek,

I am not too sure what you mean about the convergence settings. This should not make a difference. What I meant is that the flow is leaving the domain but the profiles are not maintained throughout the domain (more or less) and get completely smeared at the outlet.

It is not too difficult to specify any inlet profile you want. You can write a utility for this where prior to your simulation you look over inlet boundary faces and assign values of your choice. In this case they are interpolated from profiles obtained from a numerical weather prediction model.

Cheers,
Steve

chrisb July 19, 2010 01:43

Just an idea
 
Hi Steve,
I have not been using OpenFOAM for too long but the values you have assigned to the inletOutlet condition seem strange to me. Perhaps someone who has used the boundary condition before could comment.

My suggestion would be to try a zero gradient condition for velocity at the outlet. There is are some assumptions made about the flow (that it is essentially not changing moving upstream) near to the boundary but it should work.

Cheers
CB

stevek July 19, 2010 02:29

Hi Chris:

The inletOutlet condition sets zeroGradient when the boundary patch acts as an outlet and then kills any reversed flow by assigning (0 0 0) to inward flow.


All times are GMT -4. The time now is 20:07.