[Myda]

Hi everybody,
I'm an engineering student and i have to compute the flow around a Naca0012 in transonic speed for a project in turbulent flows. I already had tutorials in CFD last year with OpenFoam but it's the very first time i use SU2. I've read a lot of informations and tutorials to understand how to handle Su2. I'd like to thanks the community in this forum, i founded here a lot of explanation and advices.

Here is my problem:
First of all i computed my simulations in no-shock condition and the experimental and numerical data fit well.
But i have a problem at M=0.930, where i have a sonic shock: the data given by Su2 aren't consistent with the experimental data in the end of the airfoil.
In fact, they are ok until the end of the supersonic area. When the Mach number passes under 1, my Cp has to fall abruptly which is not really the case here.

For the simulations parameter:
I used a SST turbulence model with a scalar upwind scheme and a second order central central scheme JST. I also use a slope limiter which if i understander well, permit to limit the oscillations around discontinuities by switching in a first order scheme which is more stable.
I worked with implicit euler time discretization and in order to help the convergence I used a multigrid with 3 levels and W cycle.
When reading other threads, i understand that if the chord of the airfoil is 1 in my mesh, the Reynolds length has to be 1, i hope i didn't misunderstand.
I put the configuration file as attachment. And also the grid i used and the comparaison of the results.
I used several mesh and i tried in particular to have my y+<1.

Where can be the problem? I'm not an expert in cfd and Su2, so a mistakes can quickly happens.. If somebody can help me, i will be very thankful.


(here is my grid: https://www.dropbox.com/s/qs08a82oqr...ormal.su2?dl=0)

[hlk]

I have not had a chance to look at your grid; however based on the drop in Mach number in the experimental results it looks like there is a shock in the experimental results which for some reason is not captured in the CFD results. This could be because the grid is not refined enough in the region of the shock (in the chordwise direction, since you say you have y+<1). You may also want to try turning off the limiter or increasing the limiter coefficient (which reduces the amount of limiting) just in case the limiter has smoothed out a discontinuity that is actually supposed to be there. It could also be due to the turbulence model - turbulence will certainly affect a shock location, and it is difficult to match the level of turbulence present in an experiment, so try a laminar simulation as well, or changing the turbulence intensity.

[Martin Hegedus]

One must be very careful with wind tunnel results at this Mach number. The slots or porous walls which are required so the flow does not choke can, and do, have a large effect on the experimental results. I would suggest you compare to transonic 0012 CFD results that are published in papers or on the web first.

You can also try my code Aero Troll, http://www.hegedusaero.com/software, to compare SU2 to. It's free and has a GUI. Though the structured grid generation process is a little murky.