[mtheory]

Hi,

I'm currently performing a mesh sensitivity study on a NACA 63-415 airfoil (chord=0.6m, Re=1.6e6) with StarCCM+ and I decided to use an unstructured mesh with prism layers.

How many prism layers should I try to simulate? What is the direct influence of the number of prism layers on the accuracy of the results? In other threads I read that the number of prism layers should be adeguate to:

- Contain the boundary layer;
- Ensure a smooth transition to the coarse mesh outside the boundary layer.
How can I be sure that the requirements are satisfied?

Thanks in advice

[fluid23]

respect the wall y+ requirements of your turbulence wall model

https://www.cfd-online.com/Wiki/Law_of_the_wall

[Aeronautics El. K.]

In addition to what fluid23 said, you could estimate the maximum thickness of the boundary layer with some method and then, assuming a geometric (or other) progression with a given 1st cell height and growth factor, calculate the number of cells you will need for the estimated maximum BL thickness.