I have been working several months trying to teach myself CFX and working through the tutorials. I have been searching this site and I cannot find what I am doing wrong. When I walk through the tutorial for the blunt body, I see recirculation behind the body. However, I am unable to recreate this with an airfoil at different AOA and I always have a laminar flow. What am I doing wrong. I know this is a general question so please let me know what would be helpful. Sorry in advance for being crude with much of this. I am self tought and I have no one to ask.

Hi Adam,

Which Fluid model did you select in the domain?

Air Ideal Gas

Also SST Turbulence model

From my experience, I think the SST turbulence model doesnt capture the flow separation. Cause, I also ran different simulations for different angles of attack using SST model but didnt get any separation.

I used Gamma Theta Transition model with SST turbulence model and got good results with the prediction of flow separation. So, I suggest you do the same thing.

Hi,

The SST turbulence model can model separations, however it has certain accuracy problems in predicting the exact separation point (both in terms of chordal location and AOA for stall). You will find all RANS turbulence models have problems accurately predicting this so it is not unique to SST.

What Reynolds number are you using, and what airfoil?

Glenn Horrocks

Hi Glenn,

I have tried with Mach No.=0.3 and Re=0.54 Million for AOA=5,8,10 and 12 deg.The flow was fully turbulent and the model was SST.But I didnt get any separation. On the other hand, Using GAmma Theta Transition model with SST Turbulence model under the same condition, I got separation for AOA=8,10 and 12 deg.

Thanks

Farid

NACA 0012.

Hi,

To get accurate results you need a refined mesh, sufficiently small timesteps (if transient) and sufficient convergence. Have you check all these things are OK?

Regards, Glenn Horrocks

Hi Glenn,

Thanks for your help.

Actually, I ran steady state simulation but the convergence was not very good.

For mesh quality: I made very refined hexa mesh around the airfoil(NACA0012)with the inflation layer upto 4 times the height of the airfoil and with proper Y+ value. But after that I did domain interfacing with prism mesh where for the sake of time and memory I made coarse mesh(the interfacing nodes are almost 4 times larger than those of inner hexa domain).