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

[Michael Laufer]

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 

[Stian Hjorteland]

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!

[Freddie Witherden]

Hi Michael,

On 06/03/2020 07:00, Michael Laufer wrote:
> I am trying to recreate the computations from Park

> <https://arc.aiaa.org/doi/10.2514/1.J055304> (2017) and Witherden
> <https://arc.aiaa.org/doi/10.2514/1.J058434> (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.

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.

[Michael Laufer]

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