CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   How to abtain the correct y plus value. (http://www.cfd-online.com/Forums/fluent/50751-how-abtain-correct-y-plus-value.html)

sophie March 3, 2009 12:04

How to abtain the correct y plus value.
 
Hi all,

I am currentley doung a cfd project at the university of nottingham. I am trying to get a y+ value < 1 so that the boundary layer can be properly resolved using the SA solver.

I have created several fine meshes using the boundary layer function in gambit, however I cannot seem to get a y+ value below 10. Can anyone suggest anything other than adapting the gradient in fluent as this does not seem to work.

Thanks

sophie

Micael March 3, 2009 15:10

Re: How to abtain the correct y plus value.
 
y+ is related to mesh size, so your mesh is not enough fine if you got y+ > 10. Otherwise, reducing fluid speed might reduce y+. Be aware that y+ < 1 may require extremely fine mesh.

The boundary layer function is a good tool to use. If your surface is planar and fluid flow is parallel to it, then you can try to refine your mesh only in the normal direction to that surface.

sophie March 4, 2009 09:50

Re: How to abtain the correct y plus value.
 
Can you tell me how to achieve small y+ values(y<3)without reducing the speed? I want to create a realistic model at real flow speeds so that i can validate my model against published NASA data.

The meshes that i have created have extremley fine boundary layers, and making them finer any finer has little effect. Is there anyway of getting the y+ values I want without reducing the speed?

Thanks,

sophie

Micael Boulet March 4, 2009 12:20

Re: How to abtain the correct y plus value.
 
First of all, did you mean Spalart-Allmaras with "SA"? I never use this model, so I cannot really help. I use k-epsilon. Anyway, I put below a consideration about mesh size and y+. That consideration is true for k-epsilon family model, but I don't know about Spalart-Allmaras.

Here is the definition of y+: (density)*(friction velocity)*(wall distance)/(viscosity)

Wall distance is the mesh size (in direction normal to boundary).

friction velocity is [(shear stress at the wall)/(density)]^0.5

It may look that if you halve the wall distance, y+ does become half of the y+ for the original mesh. This is not necessarily true because friction velocity is solution-dependent and solution is grid-dependent, especially when you are out of the valid range for y+.

Is your model 3D? Can you build a similar 2D case with extremely fine mesh to investigate it further?

sophie March 4, 2009 13:13

Re: How to abtain the correct y plus value.
 
Yes i am working with the Spalart-Allmaras solver. I am using this as it is a relatively simple one-equation model that works well for aerospace applications involving wall bounded flows. However i have run models using k-epsilon aswell but this does not make much difference on the y+ values.

I am modelling flow over an unusual shaped aircraft developed by NASA called a lifting body (M2-F2). I am currentley running preliminary models in 2d but hope to move onto 3d soon.My meshes have very thin boundary layers ranging from starting values from 1e-3 to 1e-7, so i dont think i can create even finer meshes than this.

Micael Boulet March 4, 2009 13:48

Re: How to abtain the correct y plus value.
 
I think I understand your difficulty. You are modeling an aircraft and the viscous layer is extremely thin because of high speed. That seem normal that you need an extremely fine mesh to resolve that layer.

Can you estimate the viscous layer thickness from litterature?

Did you try y+ > 30 (but near to 30)? There are comment about it in the FLUENT 6.3 user's guide (12.11.1 Near-Wall Mesh Guidelines).

That is the best I can do, your subject is out of my expertise. I'm mostly working with slow flow, like 1 m/s.

Ralf Schmidt March 5, 2009 06:20

Re: How to abtain the correct y plus value.
 
Hi!

how many rows has your boundary layer? And what is about the grow ratio b/a?

I would set the following: number of rows = 25, b/a = 1.05

Hope, that helps..

Ralf

Friend March 5, 2009 09:38

Re: How to abtain the correct y plus value.
 
Dear

In the definition of the y plus given by Micael, it is clear that you will need to reduce the near wall distance by a factor of 1/10 in order to reduce the y plus from 10 to near 1. Otherwise if you have any limitations in constructing a smaller size then you have no more choice than to change the fluid physical properties or to choose a higher operating velocity.

Also you have the option of adapting by y plus vale of course after obtaining an initial converged solution.

Managing y plus involves an art work.... Good luck

Friend March 5, 2009 09:57

Re: How to abtain the correct y plus value.
 
y plus estimation tool: http://geolab.larc.nasa.gov/APPS/YPlus/

Source: http://www.cfd-online.com/Links/tools.html#yplus

Hope it can help

johnstewart00 April 2, 2009 09:06

project-airfoil
 
Hi,
I am having trouble validating a mesh for an airfoil created in gambit.

Fluid is air

I am having trouble getting the wall shear stress.Is it correct that i complete analysis in fluent and then use the value i get from fluent, in wall fluxes, as tau(wall shear stress)

The height of the first cell is 0.00038. not sure of the units
Any help would be very much appreciated because I am at a stand still with my project at the moment

Thanks,
John

panda April 2, 2009 19:36

If you have completed the analysis, you may find the y plus distribution anywhere around the wall directly from the Fluent postprocessing

hassan79 August 29, 2010 01:20

Dear
 
I want simulate heat transfer in turbulent air flow with sudden expansion pipe with constant heat flux, i need to calculate surface Nusselt number for down stream......
I want to ask you i get y+37 I tried to change first row no thing happen how can calculate first row and growth factor and no of row to get y+ near 1 ...........

Here is the definition of y+: (density)*(friction velocity)*(wall distance)/ (viscosity)

How can I know wall distance from equation above?

Wall distance is the mesh size (in direction normal to boundary).

Friction velocity is [(shear stress at the wall)/ (density)] ^0.5

How can I know shear stress at the wall for turbulent air flow from equation above?

My model 3D? Re 5000-90000 and d=0.0254 and D= 0.09525

George Chen 28 August 30, 2010 01:20

Quote:

Originally Posted by hassan79 (Post 273221)
I want simulate heat transfer in turbulent air flow with sudden expansion pipe with constant heat flux, i need to calculate surface Nusselt number for down stream......
I want to ask you i get y+37 I tried to change first row no thing happen how can calculate first row and growth factor and no of row to get y+ near 1 ...........

Here is the definition of y+: (density)*(friction velocity)*(wall distance)/ (viscosity)

How can I know wall distance from equation above?

Wall distance is the mesh size (in direction normal to boundary).

Friction velocity is [(shear stress at the wall)/ (density)] ^0.5

How can I know shear stress at the wall for turbulent air flow from equation above?

My model 3D? Re 5000-90000 and d=0.0254 and D= 0.09525

you can decide "wall distance"; "density" and "viscosity" are fluid property. someone had said "shear stress at the wall" was solution-depended and solution is grid-depended. it means that at first you should obtain a convergenced solution and then you can calculate y+ value, finally grid-adaption by y+ value.

hassan79 September 5, 2010 00:04

HI
 
aske about any relation ship between first row and groth factor and now ofv row and depth

my case it is expansion pipe i fond formala from gird spacing calculator to find first row ? iam not sure about this formal it is correct or no ?


could u help me
Re 8112
d= 0.0254

D = 0.09525
fluid = air
temprature inlet = 20
constant heat flux = 720
i want to calculate surface nusel no



thank u
Hassanhayawi@hotmail.com

George Chen 28 September 5, 2010 06:14

why not write complete sentences?

FabioT September 17, 2010 16:20

Quote:

Originally Posted by Friend
;157166

I found this same tool on the web but I don't understand what the Ref. Length is.

Does anybody know it?

Thanks,

Fabio

FabioT September 21, 2010 16:03

Quote:

Originally Posted by FabioT (Post 275643)
I found this same tool on the web but I don't understand what the Ref. Length is.

Does anybody know it?

Thanks,

Fabio

Check here http://www.cfd-online.com/Forums/ans...nel-model.html .

the ref. length is the average length of the edge of the considered object.

srr September 21, 2010 23:23

Try to have a y+>30 it is also ok for Spalart-Allmaras.

http://www.cfd-online.com/Forums/flu...ras-model.html

sgrshukla February 11, 2011 00:33

similar problem..
 
Hi, i have a linear jet throwing air at high speed at some angle with wall, using enhanced wall treatment, Realizable, k-epsilon model, i have pretty fine mesh but still facing problem with wall yplus >5 near jet region.
the flow is 15C and operating temp is 25C.

I would appreciate your help.

hooman.4028 April 3, 2011 20:59

Strange Problem
 
Dear all,

I am modelling flow over an ellipse using k=e model. I have a problem with my yplus magnitude.....my yplus magnitude near my wall is around 500 and it decreases by going far from the wall and reaches to zero in the first node. first i don't know why it doesn't start from zero and then goes to higher magnitudes? second, why i have such a high magnitude of yplus?my speed is 40 m/s and my meshes are really fined. also i defined boundary layer meshes in gambit.

Thanks,


All times are GMT -4. The time now is 04:27.