Hi, I am doing a problem using k-epsilon realizable model with Re 1.6E6. My y+ values are above 2000. I have a grid of 2.1 Million. What should I do?

1 Coarse the mesh? 2 Refine the mesh near wall? 3 adapt the grid? and if I do should I initialize the solution again? 4 waht must be my y+ value 5 Should I change the model?

The answer depends on what you are trying to determine in the analysis. If you are going after heat transfer characteristics, you will need to refine the mesh near the walls. Herein, you have two options: 1) if a "law-of-the-wall" turbulent boundary layer heat transfer characteristic is appropriate for your problem, you should prizm/hex refine the wall adjacent mesh to 20<Y+<~300... or... 2) if full boundary layer resolved heat transfer is required, you should prizm/hex refine the wall adjacent mesh to Y+<1, with probably at least 10-15 near wall prism/hex layers spanning with a ~1.05-1.15 expansion ratio. (The above are reasonable guidelines...for most problems).

If you have a reasonable flow field solution available from your current analysis, your can write out an interpolation file and a boundary condition file and re-initialize the new mesh with these. If not, start again as a new analysis and initialize mesh, as appropriate.

Happy Computing!

Hey,

I am performing a turbulent simulation of a flow past a simplified sedan using the k-epsilon model. I am just simulating half the car geometry with a symmetry plane. Any idea on what my y+ values should be ? and any idea on what are values of k-epsilon to input and how to calculate it ?

Looking forward to hearing from you guys. Thanks, Arjun

hi . I would like to take some suggestion regarding wall y+.. I am simulation supersonic case and using SAmodel .. presently the wall y+ value is coming around 20 but it doesnot become steady .. after may iterations it still fluctuates.. what is the reason fro that .. adn is Y+=20a;; right for SAM in fluent Thanks in advance