[bean]

Hi,

I am trying to simulate an aerofoil at high angles of attack (18+), but when I do, the residuals do not converge and neither do the lift or drag coefficients.

So, I attempted to simulate it as an unsteady simulation. Is this the correct way to do this? If so, what kind of time step etc would be needed? The simulation itself runs fairly quickly, with convergence in about 2000 iterations.

Any help would be really appreciated

Thanks

[ryancoe]

The flow at high angles of attack is likely to be inherently unsteady, thus requiring an unsteady formulation.

The proper timestep will depend on a number of things (e.g., CFL number and to what degree you hope to understand the unsteady fluctuations).

In this case, as in most, it is important to judge convergence by both the progress of residuals, but also by the progress of flow measurements, such as airfoil lift and pitching moment.

[bean]

Well, i set-up my case with 0.2s time step and over a 1.2 seconds. However, I kept an eye on both the residuals and the lift and drag convergence. On all time steps it converges to a set value, except for the last time step when both the residuals and lift and drag do not converge at all, instead they are all over the place.

I tried it at 1sec time step, and then 1.2, and on both it did this on the final time step. Do you know why this may happen? Would editing the CFL number help this?

Thanks for your help.

[ryancoe]

I don't know the rest of the details of your simulation (Reynolds number, cell size), but my guess is that both of those time steps are too large. See what happens if you set the time step so that the CFL is on the order of 1 (I assume you are using an implicit formulation).

CFL number: http://en.wikipedia.org/wiki/Courant...Lewy_condition