CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   OpenFOAM Running, Solving & CFD (http://www.cfd-online.com/Forums/openfoam-solving/)
-   -   Flat Terrain ABL problem (http://www.cfd-online.com/Forums/openfoam-solving/116690-flat-terrain-abl-problem.html)

Davide_sd April 23, 2013 07:47

Flat Terrain ABL problem
 
1 Attachment(s)
Hi all,
i'm new to openFoam; i have read the userguide and i've done the tutorials.

I have started with a small simulation of a 2D flat terrain; i've managed to setup the Atmospheric Boundary Layer inlet condition. Basically i have modified the simpleFoam/turbineSiting tutorial, and i've reduced the turbulentKE because my simulation domain is small (1.65 m long, 0.24 m height).

I'm puzzled about the results; as you can see in the picture, i've plotted the velocity profile at the inlet (dark red), mid section (green), outlet (cyan). In the horizontal axis there is velocity magnitude [m/s], in the vertical axis there is Height [m].
The velocity profile is somehow modified along the terrain, and i'm pretty sure it is wrong. What can i do to correct this issue?

PS: In this link you can find the simulation... it is too big to upload it in the attachment.

Davide_sd April 25, 2013 03:08

please, anyone can help me?

studenric April 30, 2013 11:29

hi

Iíve just seen your thread and I've checked your case. I'm not sure, because i'm also new to openFoam, but i think your top conditions shouldn't be those of a symmetry plane...they should be equal to the inlet conditions...but that's my opinion and like Iíve said, Iím also new to openFoam.

best regards

Lieven April 30, 2013 16:10

Dear Davide,

What you see in the graph is the result of two things:
1. The log-profile is only the solution to the equations for a particular set of BC. E.g. at the top of the BC you should set a fixed shear stress and not a symmetry plane.
2. (the main reason) At ground level, the gradient of the velocity behaves strongly non-linear. When you apply e.g. linear interpolation to the velocity field, you will make a large numerical error. At higher altitudes, the gradient is smaller and this is less of a problem. Hence, the profile is mainly distorted at ground level.

I would recommend you to read the following papers. The first one discusses (1) and the second one (2). If you implement the content of both papers in OF and rerun your simulation, you will get a perfectly logarithmic velocity profile.

http://www.sciencedirect.com/science...6761050600136X

http://onlinelibrary.wiley.com/doi/1...nticated=false

Enjoy ;-)


Lieven

Davide_sd May 6, 2013 08:52

Thank you guys.
@Lieven, those are really interesting papers, thank you!!!

Eloise August 6, 2013 09:11

Quote:

Originally Posted by Lieven (Post 424269)
1. The log-profile is only the solution to the equations for a particular set of BC. E.g. at the top of the BC you should set a fixed shear stress and not a symmetry plane.

Hello Lieven,

Could you explain how do you implement a fixed shear stress on the top boundary condition in OpenFoam? Is is sufficient to impose a constant velocity (Ux equal to the inlet velocity at corresponding height)?

Regards,
EloÔse

Lieven August 6, 2013 09:15

Hi EloÔse,

The fixedShearStress boundary condition is available in OF ;-).

Imposing the velocity might work, and is physically "correct" in case of an open field with perfect numerics. But as soon as you have objects in your domain and numerical errors (due to discretization), it is not the right way to go.

Cheers,

Lieven


All times are GMT -4. The time now is 16:20.