[Jannis997]

Greetings!
For my masters i have to do a series of 2d transient transitional SST simulations on an airfoil in the range of Reynolds number: 80k - 500k and angle of attack:0-9°. I want to simulate 10s. These simulations have to be stable but dont need to be extremely accurate but instead need to have low computational cost, since my computational power is limited.

What i have so far:
Fist i did a steady state simulation but the lift and drag coefficient started to oscillate after a while, which is an indication that a transient simulation is needed.
(1) Is that correct or are there other solutions?

My mesh is a C-type mesh with around 110,000 elements. I know it is coarse but condition of y+<1 is satisfied. The time step size is 0.0005 to keep the courant number under 1. If i refine the mesh, the time step size has to be reduced even further which is not possible for me because of limited computational power.
Edit: I read that the courant number is only important for explicit simulation but not for implicit, so i will refine the mesh to about 350,000 elements.

(2) Can i expect acceptable results with these parameters? Again i dont need extremely accurate results, i just need to achieve convergence and the simulation should represent the overall air flow incl. Laminar separation bubble.

(3) Can i reduce computaional cost even further by decreasing the iterations per time step (currently 20), reducing residual convergence check (1e-6) or reducing the time step size after a certain time, though i only found examples for that in temperature flow problems. Which one can i saftely do?

I already did one simulation, the lift coefficient started to oscillate after a while but calmed down in the end to the expected value.
(4) Are these oscillations normal?
Edit: I read about the van karman effect but in that case they shouldnt dissappear, am i right?

(5) For the first simulation, the drag coefficient is positive, for a second one where i changed nothing but the AoA from 0° to 6° it is negative. This makes no sense to me. Any ideas?
Edit: Solved! I forgot to adjust the monitor.

Sorry for so many questions, i did quite a bit of research but i think i need the opinion of more experienced people since most puplished studies and other example on the internet work with other models or with more computational power that i dont have. I will add a few pictures of the mesh etc. and if you need anymore info, i will provide it asap.