SST model in Fluent and CFX applies so-called "automatic wall functions". The wall treatment is automatically adjusted to the near-wall y+. If y+ is small (e.g. y+

Hi,

I wonder, why are you using fine mesh and you want to turn off the wall function.

Knowing that wall function or wall treatement play an important role when using turbulence modelling and the flow is always laminar near the wall and depend on the dimensionless wall distance y+، so, that is the reason to use wall function.

For more detailed information, check the theory section in Fluent help.

Hi

I did not get you exactly. Maybe your flow is laminar and you'd better don't use of dense boundary layer to reduce your aspect ratio, etc that. But your mean is that there is an automatic order in fluent and it reduces your size mesh around your solid and you want to off it ???

Hi

Sir/Madam please ask your problem clearly.I think wall function is needed in CFD why you turn off that function please explain clearly.

Good topic

SST model in Fluent and CFX applies so-called "automatic wall functions". The wall treatment is automatically adjusted to the near-wall y+. If y+ is small (e.g. y+

Dr. Thomas Frank ,

Thank you for your answer.

As per the contents of many CFD forums I have read, I thought fluent turns off wall function for Y+= 1 or less.

1) I am content if the value is Y+

For others who kindly took time to reply to my question, here is the reason why I am looking into removing the wall function and why I asked this question:

Simulation Details: I am simulating flow around a bluff body, Boundary layer height for the first layer is (depth/1000 = ~ 1 x 10^-5 m)

1) As per what I read in the CFD forums, I believed that Ansys Fluent would turn the wall function off automatically if the Y+ was less than or equal to 1. I thought that there had to be a way to turn off the wall function if the Y+ value was close to 1 and not 1 or less.

2) I believe that my mesh can resolve the boundary layer thus I wanted to be sure that the program wasn't using the wall function.

3) There is literature that suggests that for bluff body simulations it's recommended to make the mesh sufficiently fine and remove the wall function for proper flow separation ( I know that k-w SST is designed to model proper flow separation but I just wanted to be sure).

4) I already modeled the same simulation with a coarse mesh and got a good result. I just want to ensure mesh independence and also be able to obtain satisfactory results with a fine mesh. When running the fine mesh I got lower values for Cd than expected. I just wanted to rule out if the wall function was affecting the results since the mesh is already quite fine.

Please read the Fluent documentation.

If y+

Dr. Thomas Frank , thank you for your answer. That really helps. I will look into the documentation. Thank you.

Best wishes for the New Year!

Manoj

Dear Thomas Frank,

I'd appreciate it so much if you clear up a couple of things for me.

- Is the "automatic wall treatment" offered by CFX doing the same thing as "enhanced wall treatment" in Fluent?

- You pointed out that the "automatic wall treatment" switches from wall-function mode to mesh-resolved mode depending on y+. I'm looking forward to more details on how such a transition come along. Apparently, it's not abrupt as the manual has mentioned "blend". The CFX manual explanations on this topic seems so confusing to me.

Regards,

Armin

Thank you for posting this. In fact I have never thought about turning off wall functions or its possibility in ANSYS Fluent. I have also noticed that while using boiling models (RPI model) the turbulence is automatically switched on even if it is a low Re simulation. Is this some kind of inbuilt feature of the code and does it effect the final solution?

Thank you,

Kiran

Armin Hajighasem Kashani :

CFX automatic wall treatment = enhanced wall treatment?

It should realize essentially the same thing, i.e. a smooth blending between using logarithmic law of the wall for meshes with large y+ to mesh-resolved for meshes with y+

Kiran Saikia :

I cannot really comment on this, since I used wall boiling models only in ANSYS CFX, never in ANSYS Fluent. But as far as I'm familiar with the model formulations and its dozens submodels, they have never been developed for low-Re flows. In particular the turbulent dispersion force acting on vapor bubbles plays an essential role in removing the vapor layer from the heated wall. Probably this RPI wall boiling model was never really applied to low-Re flow conditions and is not validated for such conditions. I could imagine, that this is enough of a reason to activate a turbulence model by default, once the RPI wall boiling model is chosen. But I have never used it by myself in Fluent. And in ANSYS CFX such an automatic model pre-selection is not existing to my knowledge.

Regards,

Dr. Th. Frank.

Armin Hajighasem Kashani :

Enhanced wall treatment is used in FLUENT only applicable to epsilon-based models. It is a two-layer approach, where a one equation model (Wolfstein) is used in the viscosity affected near-wall layer (if the mesh is able to resolve this, on a high y+ mesh its pure wall function) to circumvent problems when integrating the epsilon equation towards the wall. It also involves a blending procedure between the mu_tau values of the two layers. Therefore it is essentially different from the "automatic wall treatment" for omega based models (see Fluent Theory Guide). Fluent documentation is not giving much detail about the latter but I would assume it is pretty much the same as the implementation in CFX.

Which part of the CFX documentation on this topic is confusing you? Basically it blends between the log layer and the analytical sublayer expression for omega, depending on the local y+. And it does the same for u_tau. The blending function itself is purely empirical (2nd power and 4th power), but shows a reasonable behavior through the buffer layer.

Best regards,

Oliver