Relation between cell size, Y+ and turbulence models

    • CDF RT

      Hello everyone,

      I have studied about the law of the wall and turbulence models, anyways I still have problems to understand how Y+ values and turbulence models affect the simulation.

      As I understand it, the law of the wall states that fluid velocity follows a different function in the viscous sublayer Y+<5 and in the log layer 30

      In example, if the center of the cell closer to the wall has an Y+ of 23 and I am simulating with standard k-?, the model is calculating as if  this cell was is in the log layer but it really is in the buffer layer, so the simulated flow velocity is incorrect. Am i right in assuming this ideas or I did not understand it?

      Mi doubt in first place is, what happens with models like k-ω SST, that have an "Y+ Insensitive Wall Treatment Equation"? Does this model work correctly regardless of what Y+ I have? I have read somewhere that k-ω SST needs an Y+<1 to work properly, why less than 1 if we are inside the viscous sublayer when Y+ is less than 5?

      My doubt in second place is, if the stated above is correct, how is it supposed to be used this information? When there is a complex flow, the velocity in one wall is totally different to the rest of the walls, would each wall require a different cell size to match the Y+ value? Do I need to simulate, then check the Y+ graph and resize the mesh until I achieve a simulation with the correct Y+ values for all the walls in the domain?

      I am currently carrying a simulation usong the SST model, as you can see in the image attached the Y+ varies a lot in the walls, should I assume the results are incorrect? Do I need to resize the mesh until I get the correct Y+ in the whole domain?

      In the other hand I dont know if there is any better way of checking Y+ values than graphing it the way I did.

      And lastly, where can I get proper information about the Y+ values for each turbulence model? The only information I have found is in forums but did not find any ANSYS recommendation for Y+ values or something like that.

      Thank you in advance.


      PS. Is there any turbulence model best suited for ANSYS academic? I dont have the posibilty of creating very fine meshes and I would like a model that works properly whitouth having to check the Y+ values.



Viewing 0 reply threads
  • You must be logged in to reply to this topic.