Strange results for rough surface
 

Hello Everybody,

I´ve done a simulation concerning boundary layers first with a smooth wall and then with a rough wall.
Booth simulations with the k-epsilon model and the roughness height is smaller than the first cell.

I have attached a plot of the velocity vs height. As you can see the velocity for the rough wall is to large.
I would have expected the lines not to cross and the velocity for the rough wall smaller than the velocity for the smooth wall.

Is this a Problem of my Simulation, the keps Model or the rough wall?

Any help is appreciated!

Thanks

Camoesas

What y+ are you using? When you use wall functions you should not go below a y+ of about 15.

HI Mr. Horrocks,

I have an y+ of about 1.1-1.8 (on the wall, plotted in Post)

What I dont understand is: y+ is defined as: something * wall Distance
Thus y+ should be 0 at the wall, because the Distance is zero.

But in CFX Post y+ is everywhere 0 beside from my wall...

Are there different y+ or where is my fault?

Thanks

y+ is calculated in the first control volume, but just shown on the wall. The distance from the wall is therefore the distance of the centroid of the first cell from the wall.

Your y+ is too low for a wall function approach. You need wall functions to use the rough wall model. You need to coarsen the model to y+>15 (about).