CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   OpenFOAM Pre-Processing (https://www.cfd-online.com/Forums/openfoam-pre-processing/)
-   -   boxTurb utility (https://www.cfd-online.com/Forums/openfoam-pre-processing/134217-boxturb-utility.html)

aylalisa April 26, 2014 03:12

boxTurb utility
 
Dear Foamers,


I've used boxTurb utility to initialize the flow field with correlated noise. The mesh is created with blockMesh (no grading) and consists of two blocks.

Using contour plot shows the U_mag (boxTurb) of initialized field U:
http://up.picr.de/18085510mk.png
and
http://up.picr.de/18085511ko.png

It seems that the computed field U is split according to the two blocks of the mesh.
Before I've started boxTurb I've mapped a parabolic velocity profile from another case with a consistent mesh to that case (consisting of two blocks, too).

Is that field for further computations o.k.? Why can I see the two blocks in the field U?

To be able to use contour plot I had to calculate U_mag manually before. I understand that contour plot needs scalars but I thought, since the field U magnitude is already available, I could directly choose it in the 'Properties' tab of Contour filter in Paraview?


Best regards,
Aylalisa

Z.Q. Niu May 5, 2014 23:29

Hello, Elisabeth!
I has trouble in boxTurb too, when I boxTurb my box case, the message followed appears:
Create time

Create mesh for time = 0

Reading field U

Reading boxTurbDict

k(0) = 0
end
Did you meets this trouble?

aylalisa May 6, 2014 08:50

Hello Z.Q. Niu,

no, boxTurb utility (fortunately) worked.

Did you specify values for Ea and k0 in boxTurbDict?
e.g.
Ea 10;
k0 5;

Did you initiate your velocity field with a parabolic velocity field before starting boxTurb?
I've mapped an internalField U (with parabolic profile from last time folder 580.7) from another case and then start boxTurb.

I've created a TestCase and repeated what I've done before. Here is the feedback from boxTurb utility:

Quote:

Create mesh for time = 580.7

Reading field U

Reading boxTurbDict

k(580.7) = 0.000400794
end
kind regards,
Elisabeth

Z.Q. Niu May 6, 2014 23:50

Hello Elisabeth,
I have found my problem using boxTurb, the boxTurb requires the mesh number must be power of 2. But I still has some confusion that why a parabolic velocity field should be intiated before starting boxTurb? this parabolic velocity file maybe will be covered by boxTurb. I has read a tutorial about dnsFoam at http://www.tfd.chalmers.se/~hani/kur...de_Mare_v3.pdf. in this PDF, it illustrates that boxTurb creates a random initial 3 dimensional flow field consistent with the current
understanding of isotropic turbulence.

aylalisa May 7, 2014 17:36

Hello Z.Q. Niu,

I've mixed it up, sorry for that! I've mapped a parabolic profile before I've used perturbU, so I did the same before I've used boxTurb :o !
(But I think boxTurb needs a velocity field? To be sure I have to go back to the code)

Z.Q. Niu May 7, 2014 23:30

In my opinion, and contents in http://www.tfd.chalmers.se/~han/kur...de_Mare_v3.pdf, the initial velocity field should be generated by boxTurb completely. But I can't be sure of it .

aylalisa May 8, 2014 04:11

You are right! --> turbGen.C, sorry that I've caused confusion.

Z.Q. Niu May 8, 2014 23:29

Dear aylalisa,
Do you know how to plot the curves of u+ versus y+(i.e. yplus) near the wall, I have not found the utility to achieve it. Thanks!

aylalisa May 9, 2014 06:28

Dear Z.Q. Niu,

please have a look at the real helpful answer from Mr. B. Gschaider:
http://www.cfd-online.com/Forums/ope...tml#post490196

I still stick to that idea (without solution) . My problem is two-minded.

1) I would like to principally learn how to get/work with internal fields (to be able to do each kind of postprocessing I need to do) , how can I create/delete OF objects (including swak4Foam) and actively work with the code

2) If I would like to show u+/y+ ,let's say from the wall viscous layer to transition layer finally to layer where wall turbulence develops, I need more than only one value for y+ . Therefore I try to extract the complete field. If I have the internal volume fields of y+ and u+ I could analyze the flow characteristics at each possible position.

Please have a look at this post and the very helpful answer of cosimo bianchini:
http://www.cfd-online.com/Forums/ope...tml#post490777


Is that idea about y+ wrong???


Best regards,
aylalisa

ZJUXJK October 26, 2018 06:32

Dear Z. Q. Niu,
I got the same trouble as you mentioned. Ihave noticed that in the tutorial dnsfoam case, the intial velocity field is presetted. I deleted the interfiled, and run the boxTurb. Then the tutorial case does not work. Have you fixed this problem?


best,
Jiangkuan
Zhejiang Unviersity


All times are GMT -4. The time now is 22:28.