# Wall distance experience, do these calculations look right? (y+ yPlus)

#1

Hey,
sorry, i dont have any experince in calculating the “right” wall distance for a turbulence modell.
I would like to ask, if the results could be correct.
Initital values:

rh = L = 0,12\:m \\ U = 1,025\:m/s \\ \rho = 999,1\:kg/m³\\ \mu = 0,00114\:kg/m\cdot s \\ y+ = 30

Reynoldsnumber:

Re =(\rho \cdot U \cdot L) / \mu = ( 999,1 \cdot 1,025 \cdot 0,12)/0,00114 = 104.324

Cf:

Cf = (2\cdot \log_{10}(Re)-0,65)^{-2,3} = 0,0058

tau wall:

\tau_{wall}= (1/2) \cdot C_{f} * \rho \cdot U^{2} = (1/2) * 0,0058 \cdot 999,1 \cdot 1,025^{2} = 3,043 kg/(m\cdots^{2})

Ufric

U_{fric} = \sqrt{\tau_{wall} / \rho} = \sqrt{ 3,043 / 999,1 } = 0,0552

y:

y= (y+ \cdot \mu) /( U_{fric} \cdot \rho) = (30 * 0,00114)/(0,00552 \cdot 999,1) = 0,00062\:m\\ 1st\:Cellheight\: 2\cdot y = 0,00124\:m

So the first cell has to be 1,24 mm height, would you say ok, that could be, or did it feel way to small ?
I’d like to settle this before I create the mesh and consume many core hours in simulations .

Best regards
Kai

#2

Hi @Kai_himself,

PointWise yPlus calculator

Cheers,

Retsam

#3

your calculated value doesn’t match with an online calculator value. I think you made a mistake in the hand calculation. OR my input value was wrong?

Rohit.

#4

Make sure that you consider this. (EDIT: I did see that you did consider that 2x multiplier (perhaps @ROHIT_SR forgot to multiply his output ) , I was just trying to make sure that people who use this topic for future reference will not miss it )

And then you need to confirm that you actually achieved the y+ which you seek, by viewing a y+ surface map.

BUT, remember that a y+ surface mapping is most valid on results from a converged Sim Run WITH stable results values.

I look for forces to be stable to less than 1% range over the last 500 iterations. Range is WRT each forces average value over the last 500 iterations, or the end value at the last iteration that they appear to be converging to. I have proposed a new way to specify results stability with an ORSI value which you can have a look at here.

More accurately than visually guessing at an average achieved y+ mapping value in SimScale Post Processor, you can download the results of your Sim Run and fairly easily have ParaView give you an average y+ value OR with a little more effort you can use my yPlusHistogram program on a csv exported from ParaView for a detailed analysis of your y+ surface mapping.

@Kai_himself , since that is your 1st try at y+, CONGRATULATIONS, well researched, as I believe everything that you did correctly, likely did not come from one source, did it? …

#5

Damn you PowerUsers are fast in helping Kai Keep up the good work and communication guys, appreciate it! And props to @Kai_himself for being such an active (and of course cool) member

Best,

Jousef

#6

Thankys for this many answers and to check my results @ROHIT_SR,
yes i double the result to get the cell height.
@DaleKramer,
thank you for your explanation, i will try your programm tomorrow. I hope i works with my simulation type.

The most used source is this good post from @jousefm Post, but i use the a different fomrular for the skin friction coefficent, its the equastion from Schlichting in that source Link.

I will get back to you tomorrow

Best regards
Kai

#7

I am pretty sure that you did not learn about the doubling from Jousef’s fine post, or the second link…
Where did you learn to double?

#8

@DaleKramer,
Hey, and sorry for my late replay.

I think its the same in Jousef’s post, there you are calculation y, and this is the cell center, or get i wrong ?

Best regards
Kai

#9

WOW, excellent video, great find…

It should be required viewing for anyone that needs to work with y+…

Jousef’s post does not seem to talk about cell centroid vs cell height issue…

I think Jousef stops at just calculating y, which is the cell centroid distance from the surface (here is excerpt from his post):

However our meshes need to use cell height, which is 2*y

I think that your topic has turned out to be the best topic so far for people to read if the want to use y+ correctly…