NACA 0021 Airfoil in Deep Stall - Diverging during start-up phase

General
1

Hi PyFR community,

I am trying to recreate the computations from Park (2017) and Witherden (2020) on a NACA 0021 airfoil in deep stall.
I run into an issue almost immediately during the low-order (P1) start up phase without anti-aliasing, where I get a divergence error (“RuntimeError: Minimum sized time step rejected”)

I have also tried running with a lower Re=27K and only after afterwards changing to the nominal Re=270K like explained in Park, but it still diverges once the Reynolds number is increased.
In addition, I have tried both Riemann-solver types (roem and rusanov) that were used in both papers, but no difference was seen.
Lastly, I am seeing this phenomenon with both the P3 and P4 meshes provided in Park.

Can you share any other tips to solve this? Am I missing something else here?
I am attaching the config file used.

Much appreciated,
Michael Laufer

naca0021.ini (1.68 KB)

2

I am also trying to solve an airfoil at deep stall (90 degrees), and are having some problems doing so. If you figure out the problem, please share!

fredag 6. mars 2020 14.00.37 UTC+1 skrev Michael Laufer følgende:

3

Hi Michael,

My start-up procedure was to run at p = 1 and Re = 27k with no
anti-aliasing until t = 100. Then, restart with the ‘main’ .ini file
(final p, correct Re, anti-aliasing/filtering as required).

Regards, Freddie.

Edit by WillT: Fixed missing text from mailing list import.

4

Freddie,
Thank you,

I see my error now. After running p = 1 and Re = 27k, I restarted with p = 1 and Re = 270k, instead of p = 4.
I will give it another shot with the ‘main’ .ini

Michael Laufer