Olllom
commented
2 years ago Hi @robert-DL,
sorry, just saw this. Did you figure something out?
It should not matter whether you use density or pressure variation as a stopping criterion, as there is a linear relation between the two.
Pressure is probably a bit more convenient. For a fixed tolerance it gives you the same quality initial solution independent of Mach number. So checking for pressure facilitates setting a default tolerance.
Cheers, Andreas
PhiSpel
Hi,
I want to initialize proper Dirichlet Boundary Conditions, e.g. of the Form
\uD = \u{avg} \frac{6y(H-y)} {H^2} w(t),
with w(t) given as w(t) = 0.5 - 0.5cos(2pi*t) if t < 0.5 and 1 if t > 0.5.
In the Krüger et. al. book, there is a short chapter on proper initialization. They mention the method by Mei et. al. However, this approach is only for one single time step (most likely t=0). How to handle the above case. Use the iterative approach of Mei, repeat it for subsequent time steps and set the initial condition for every time step separately? Seems a little bit odd to me.
Furthermore, there is a method in the Simulation class, that initializes f such that it matches the given moments. It looks pretty similar to the approach of Mei et. al., with the only difference that the termination criterion is pressure-dependent, Mei et. al. propose a density-dependent criterion.
Any suggestions on that?
Best, Robert