[Leifheit]

Hello,

I am trying to run a supposedly very simple 2D simulation of a propeller blade.
Basically it is a 2D airfoil simulation using water instead of air.
Its a steady simulation in CFX.

Here is some relevant data:

RE:3*10^6
first cell size: 0.0005 --- corresponding Y+ values vary on the airfoil surface from 10 to 80
SST turb. Model used

I am trying to verify my simulation by comparing results for a NACA1408 shape to experimental data.

At 8 deg angle of attack experimental data gives a lift coefficient of 0.95, my simulation says 0.86.

Does anyone have ideas how to approach this problem ?
I tried:
-changing mesh resolution; general resolution
-changing first cell size
-changing "time step"

I do see a disruption of e.g. velocity profiles at a GGI interface i am using but i dont think this should affect the lift coefficient too much. Mesh resolution at both sides of the interace is quite similar too.

Furthermore I have some questions about CFX which are secondary but if anyone is willing to it would be nice to have them answered

- ANSYS documentation talks about a "pseudo timestep" when doing steady simulations; I dont really get the explanation though, so if anyone could explain this "pseudo timestep" it would be really nice
####ANSYS doc: CFX applies a false timestep as a means of underrelaxing the equations
-> The timestep determines to what extend under relaxation is done ?
-> Why not just set a under relaxation factor then ?

Thanks a lot for all answers in advance

Regards,

Tobias

[ghorrocks]

This is an FAQ: http://www.cfd-online.com/Wiki/Ansys..._inaccurate.3F