[dunapkt]

Hi everyone,

I'm currently working on modeling User-defined wall function for ANSYS fluent. I'm using standard k-epsilon model for my simulation, and fluid domain has a width of 3H, a height of 2H, and a length of 6H, where H is a variable. The domain is bounded by lower and upper smooth walls, with a Friction Reynolds number(Re_tau) of 180 and a Bulk Reynolds number(Re_b) of 5600.

Now, I'm confused about y+, y* for meshing and how wall function works.

Here's my current understanding:

1) If y+ <11.225, u+ = y+ is applied; otherwise, u+ = 1/kappa*ln(Ey+) is used

2) It's recommended to avoid putting cells in the buffer layer, i.e. 5<y+<30.

3) According to the Ansys Fluent theory guide, Fluent uses y* and U* rather than y+ and U+ for Standard wall functions. The lower limit always lies in the order of y*~15.

4) In most case, y+~y* anyway.


My question is,

If we set lower limit of y*~15, as y+~y*, then y+ is also ~15. It means cells are on the buffer layer.

So, to meet both y+ and y* conditions, should I make the first cell's y+ and y* values at least 30, or just need to satisfy y* ~ 15? If it's necessary to meet both conditions, it seems that the cells are too large from my perspective. I have around 6 cells between the upper and lower walls.

Also if it's wrong, y*~15 gives too large cells either, around 10 between walls.

I'd appreciate any insights or guidance on this matter.

Thanks.