[immortality]

OK.
thanks for good clarification!
then it splits only the first cell.
how does it calculates the number of splitts?
and if I change the mesh by blockMesh I have to use refine tool again?

Note: add the option "-overwrite" is you want it to override the existing mesh.

do you mean writing in boundary,faces,cells,...files in blockMeshDict?

[wyldckat]

Originally Posted by immortality

how does it calculates the number of splitts?

With this:

refineLevelValue = ceil
            (
                std::log(YpMax/targetYPlusValue)/std::log(scalar(2.0))
            );

Note: "targetYPlusValue" is 1.0 by default, unless specified from the command line.

Originally Posted by immortality

and if I change the mesh by blockMesh I have to use refine tool again?

Yes.

Originally Posted by immortality

do you mean writing in boundary,faces,cells,...files in blockMeshDict?

Only affects the actual mesh files.
It does not affect the "blockMeshDict" file, because it's used for creating a mesh, it's not "the mesh" .

[immortality]

and another question.if y+ be 1 for example and we want to increase it to 30 this tool can't be used,can?

[wyldckat]

Originally Posted by immortality

and another question.if y+ be 1 for example and we want to increase it to 30 this tool can't be used,can?

Nope. That goes in the wrong direction You have to create the mesh first, then refine it.

You can check post #14 for more ideas: http://www.cfd-online.com/Forums/ope...tml#post397239

[immortality]

Hi
why the first cell height after refining is bigger than second one?
I used

refineWallLayerSmart -basedOnYPlus -targetYPlus .7 -overwrite walls 1.5

[immortality]

in this version
1) it works on yPMax instead of yPMean.so the maximum y+ in the domain will be lower than specified -targetYPlus that leads more accurate results than the average.
2)also the modification of first cell that was bigger than second cell has done according to Bernhard proposed .C file.
3)also it now accepts float numbers for -targetYPlus like -targetYPlus .5
Hope it be useful.

[immortality]

in what it used to be till now we wanted to maximum yPlus be lower
than targetYPlus (as above post)
now I want minimum yPlus be bigger than
targetYplus.it refines mesh still(when for example minimum of y+ is 200) but min(yPlus)>=30(that uses in high-Re mesh's)
it seems an equation should be changed from max(yPlus)<=target to
min(yPlus)>=target.
I'm using average version now but if anyone could have a look into it too
it will be more accurate.
its nice to modify above post file.
thanks.

[mezomatic]

Originally Posted by immortality

in this version
1) it works on yPMax instead of yPMean.so the maximum y+ in the domain will be lower than specified -targetYPlus that leads more accurate results than the average.
2)also the modification of first cell that was bigger than second cell has done according to Bernhard proposed .C file.
3)also it now accepts float numbers for -targetYPlus like -targetYPlus .5
Hope it be useful.

Hello there!
I used your Version of the tool on my mesh but the first cell height ist still larger than the second, just like in your post.

I used

refineWallLayerSmart -basedOnYPlus -targetYPlus 1 Wall 1.15 -overwrite

Did you get different results?

[immortality]

Hi
in my case it works right.
send your mesh to me.

[mezomatic]

Hi!
This is my case, it's just a simple 2D BlockMesh...

[lakeat]

Originally Posted by wyldckat

Greetings to all!

Attached is a modification as requested by Ehsan. It's based on Daniel's original modifications and I changed the following

Thanks Bruno!

Using YpMax seems to me too cruel.

[immortality]

Hi!
This is my case, it's just a simple 2D BlockMesh...

this is somewhat strange!why you have set number of cells in y-direction to 1!
then you expect what to be refined in y-direction?!

[immortality]

Thanks Bruno!

Using YpMax seems to me too cruel.

thanks for original utility,
in cases we want for example y+<1 its more accurate that maximum value of y+ on cells be lower than 1 too at least I think so if you want to correct my understanding I welcome!

[mezomatic]

Originally Posted by immortality

it works in y-direction.

What do you mean by that? Why would it work in only one direction?
I expect nothing to be refined in the y-direction, it's a 2D mesh and it lies in the x-z-plane...
I'm sorry but I didn't understand your concerns about that, the refinement works just fine (see attached screenshot). Only "problem" is the first cell being larger than the second...

[immortality]

I had forgotten how it was working.I don't know whats wrong with your case.it ends with a crash.
have you set faces correct?I think maybe there is an issue in blockMeshDict.

[wyldckat]

Greetings to all!

@mezomatic:

Originally Posted by mezomatic

Hi!
This is my case, it's just a simple 2D BlockMesh...

I was not able to reproduce the same error

I ran blockMesh, then yPlusRAS and then the command indicated above:

Originally Posted by mezomatic

I used

refineWallLayerSmart -basedOnYPlus -targetYPlus 1 Wall 1.15 -overwrite

Did you get different results?

But I got this message instead:

--> FOAM FATAL ERROR: 
targetYPlusValue is larger than max yPlus (0), no need to refine the wall layer mesh.
Or you can use "refineLevel" option to force refine.

Best regards,
Bruno

[mezomatic]

Hey Bruno,

try running simpleFoam on it first for a few timesteps, so the yPlusRAS tool can compute the y+ values properly. Otherwise it's just zero everywhere (see error message) and as that is already below the desired y+ of 1, there won't be anything refined...

[wyldckat]

Hi mezomatic,

Originally Posted by mezomatic

try running simpleFoam on it first for a few timesteps, so the yPlusRAS tool can compute the y+ values properly. Otherwise it's just zero everywhere (see error message) and as that is already below the desired y+ of 1, there won't be anything refined...

Uhm... simpleFoam crashed at iteration 538 due to massive continuity problems.
Nonetheless, I checked the y+ values on the wall at the iteration 500... and the maximum value of "y+ = 1e9" is waaaaaaaaaaaay too much! Which means that refining the mesh to "y+ = 1" would require the cells to be so thin near the wall, that... uhm... maybe as thin as an atom? I didn't even do the math... all I know is that checkMesh seriously complains about the resulting refined mesh.

Therefore, I strongly advise you to first make your case simulate without crashing, and only then you should try refining the mesh near the wall

Best regards,
Bruno

[immortality]

as Bruno says I received an error too and it seems your BC's and maybe other settings have a big problem in them!find that first.

[mezomatic]

Hmm that's strange... I've got no problems at all running the the case without the refinement. My max y+ is 59.

Maybe there's someting wrong with the files. Here's the case again.
I'm using Version 2.2.0