annular, ring airfoil
 

Hi all!

I'm simulating 3D annular/ ring airfoil. AR of the ring is 1, M=0.13 and Pgauge=100800 Pa

My boundaries are: front: 4 L (chord of airfoil) rear: 6 L Upper and lower: 3L

I use 400.000 grid (the CPU can't handle more than this), coupled-implicit and spallart almaras.

I simulate from 0 to 90 deg. The results for 0 to 20 deg are good, but for the rest are not. How can I improve the result for high AOA.

Thank you

Re: annular, ring airfoil
 

Accurately modeling highly separated flow from airfoils often requires the use of combined RANS/LES methods e.g. DES.

Read this: http://www.fulton.asu.edu/~squires/p...ions/dles5.pdf

This requires considerable computing resources.

Re: annular, ring airfoil
 

thank you for your answer.

do you know other alternatives since this one requires extensive computing resources.

thanks

Re: annular, ring airfoil
 

If the poor agreement is indeed caused by the use of RANS you have no choice but to run it using a different turbulence modelling approach.

Re: annular, ring airfoil
 

You say you are only using a 400k mesh for a 3D problem? That may simply be too coarse even for RANS.

Do a grid independence study.

Re: annular, ring airfoil
 

Good advice. A grid independance study is always advised. It may be so coarse that transient behaviour (such as vortex shedding) is masked. What is the Y+ and wall treatment? Can you do initial runs for separation point on a simpler (and quick to run) 2D profile first as a representative test? Try to use a structured grid to avoid excessive diffusion also with boundary layers on airfoil.

Re: annular, ring airfoil
 

Hi,

i tried to use 1.2 million using segregated solver and S-A, k-epsilon ad RANS. the results for low AOA are bad for the Cd, while for high AOA are bad for Cl.

I tried to run unsteady for high AOA (i simulate until 90 deg) but the results are not getting better. maybe because of the time step size. can you suggest reasonable value for the time step?

the y plus is around 300 for Re=1.2 and S-A. i tried to adapt the boundary but the computer is out-of-memory.

thanks.... :)