[Moreza7]

Hello,

I have simulated the flow around an airfoil using LES, but I have a problem.
In some articles the skin friction coefficient becomes negative in the laminar separation region, but in my case it just decreases to near zero, but does not become negative. Just like the attached image. Suppose that the red line is mine and black line is for the article which I compare my results to.
Does this mean that the separation is not occured or something else?

Regards.

[FMDenaro]

Originally Posted by Moreza7

Hello,

I have simulated the flow around an airfoil using LES, but I have a problem.
In some articles the skin friction coefficient becomes negative in the laminar separation region, but in my case it just decreases to near zero, but does not become negative. Just like the attached image. Suppose that the red line is mine and black line is for the article which I compare my results to.
Does this mean that the separation is not occured or something else?

Regards.

Are you sure to compare the curves at the same z/c locations or the curves are the mean in spanwise direction? Is your grid enough fine?

[Moreza7]

Originally Posted by FMDenaro

Are you sure to compare the curves at the same z/c locations or the curves are the mean in spanwise direction? Is your grid enough fine?

Thanks for your reply.
The grids are the same and the curves are spanwise averaged.
My main questions are:
1- If the skin friction doesn't become negative, does it mean that the separation is not occured?
If yes, then why it does not become negative in turbulent separation region? How can we tell that the separation is occured by observing Cf?
2- Should I calculate the wall stress by du/dy or du/dn?
I used du/dn in this case.

[FMDenaro]

Originally Posted by Moreza7

Thanks for your reply.
The grids are the same and the curves are spanwise averaged.
My main questions are:
1- If the skin friction doesn't become negative, does it mean that the separation is not occured?
If yes, then why it does not become negative in turbulent separation region? How can we tell that the separation is occured by observing Cf?
2- Should I calculate the wall stress by du/dy or du/dn?
I used du/dn in this case.

Evaluating the skin friction or du/dn is the same, you have the tau_wall stress to compute. In your case there is no appearence of a separation point. Are you averaging along z?

And yes, you should use the local reference system (t,n) along the wall so that you compute dut/dn.

[Moreza7]

Thank you.

Yes, as I said, the results are spanwise-averaged along z.

I just noticed that it happened because of the premature numerical trip of boundary layer in CDS. As the mesh is coarse in the upstream of the airfoil.

The problem solved by using blending between CDS and 2nd order upwind in the laminar region and upstream of the airfoil.