CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   FLUENT (https://www.cfd-online.com/Forums/fluent/)
-   -   y+ and y* issue (https://www.cfd-online.com/Forums/fluent/38341-y-y-issue.html)

 Freeman October 23, 2005 11:20

y+ and y* issue

Hi all!

I am just starting using fluent for my simulations and in many bibliography the y+ and y* appears to be very important to make a right mesh, but the fact is that i still don't understant the meaning of these dimensionless parameters and their implications/utility for meshing?

Can someone explain them to me in plain english please?

 Ahmed October 23, 2005 22:08

Re: y+ and y* issue

Panton book "Incompressible Flow" develops the Law of the wall. In short, we cannot generate a mesh to accuratley capture the flow details inside a boundary layer, instead all CFD programmes represent the boundary layer by a single wall layer. The thickness of this layer is selected so that the law of the wall is valid. This is important because the pressure drop along the flow path depends on the skin friction which is calculaed from the Law of the wall.

 Freeman October 24, 2005 15:07

Re: y+ and y* issue

Thanxs Ahmed. That was really helpfull.

So, from your reply I understand that near the wall CFD soft calculates flow fields following the law of the wall and not only the RANS, don't them?

The only think that I don't know is which is the interval of y+ in what the Yplus Plot must be limited by, depending on the model I use to solve the problem. Can you or anyone highlight me in this? For example, I've read that for the Spallart-Allmaras the Yplus must be grater than 30 or near 1 but not in the 1-30 region? Is that right? And for the other RANS models?

Many thanxs!

 pUl| October 24, 2005 17:27

Re: y+ and y* issue

Once more:

The fourth comment in the above link to be precise.

 Ahmed October 25, 2005 06:23

Re: y+ and y* issue

 Freeman October 25, 2005 13:17

Re: y+ and y* issue

Ahmed: in point 11.9.1 of Fluent's users guide there's one plot of the law of the wall. And I've also found it in "Prandtl's Essentials of Fluid Mechanics" book. Know in this post I've realized about the importance of Y+ plot. With this, I conclude that I have to avoid the y+ located in the buffer layer, but new questions come to me:

1. In Fluent's users guide, it is said that Fluent makes calculations with the law of the wall for mean velocity when mesh is such that y*>11.225 and when it's lower than this value, it aplies the laminar stress-strain relationship U*=y*. So, this means that although Fluent makes this calculations, they would be not very reliable if I get a mesh with y+ values less than 30?

2. Which is the difference between y* and y+? Does Fluent always use y* for its calculations? In which manner does this affect in the recommended interval of [30,300] of y+?

3. When using Near-wall models (instead of wall functions), is still valid all the methodology done with y+ as we do when using wall functions? I mean, if plotting y+ makes still sense or I have to revise my meshes with other techniques?

I know the above was quite long ;) Thanxs a lot!

 pUl| October 25, 2005 17:09

Re: y+ and y* issue

I can answer the thrid question quickly and so here it is:

When you use Enhanced Wall Treatment (EWT), you're really resolving the flow all the way down to the wall (actually the turbulence models are modified to work that way when you use EWT). Ergo, you need to maintain a very low Y+ value near the wall. The recommended range is (1 to 5) although a value around 1 is usually recommended. Creating a mesh with such resolutions is often tricky as one also needs to ensure that cell aspect ratios (ratio of one edge of a cell to the other) greater than 1:5 are best avoided. So plotting the variation of Y+ along the wall is still useful as you can check if your resolution is within prescribed limits. However, bear in mind that the turbulent reynolds number has to be below 200 or some such for the first 10 odd cells. This is mentioned in the Fluent Users Guide.

Good Luck!

 Freeman October 26, 2005 10:19

Re: y+ and y* issue

Ok pUl|, many thanks. These days I am reading point 11 of fluent's users guide intensively to take care about all these things about the mesh size near walls. I hope I could answer the othr questions for myself in the following days, but those have came to me reading the user's manual xD. Specially the fact that fluent does not work directlly with y+, as he makes calculations with y* and that in some cases,it uses the linear equation U*=y* (I imagine that it is to have a solution "at any cost", knowing that the mesh is not fine enough to use standard wall-equations and the result may be not accurate)

I will keep working on that (and reading/searching the forum for new answers). Thanxs again!

 Anna Tian December 7, 2013 16:23

Quote:
 Originally Posted by Freeman ;125537 Ok pUl|, many thanks. These days I am reading point 11 of fluent's users guide intensively to take care about all these things about the mesh size near walls. I hope I could answer the othr questions for myself in the following days, but those have came to me reading the user's manual xD. Specially the fact that fluent does not work directlly with y+, as he makes calculations with y* and that in some cases,it uses the linear equation U*=y* (I imagine that it is to have a solution "at any cost", knowing that the mesh is not fine enough to use standard wall-equations and the result may be not accurate) I will keep working on that (and reading/searching the forum for new answers). Thanxs again!

If that is true, why do we still need to know Y+? Y* is already enough to make sure that the first line is inside the subviscous layer.

 elmo555 December 6, 2017 12:46

Quote:
 Originally Posted by Freeman ;125487 2. Which is the difference between y* and y+? Does Fluent always use y* for its calculations? In which manner does this affect in the recommended interval of [30,300] of y+?
Quote:
 Originally Posted by Anna Tian (Post 465283) If that is true, why do we still need to know Y+? Y* is already enough to make sure that the first line is inside the subviscous layer.
I know this is an old topic, but in the case anyone has a similar question in the future: The Fluent guide (for version 12.0) says on page 12-6:
Quote:
 The log-law, which is valid for equilibrium boundary layers and fully developed flows, provides upper and lower limits on the acceptable distance between the near-wall cell centroid and the wall. The distance is usually measured in the dimensionless wall units, y+ (≡ ρuτy/µ), or y*. Note that y+ and y* have comparable values when the first cell is placed in the log-layer but are different by C_µ^(1/4), i.e. ≈ 0.5.

 All times are GMT -4. The time now is 11:29.