[wouter]

Hello All,

In my model I have two scales:
scale 1:
the main part is a reactor with a diameter of 2 m and a hight of 10 m. In this reactor filled with liquid I would like to examine the mixing due to gas bubbles. The superficial liquid speed is approx 1-5 m/h

scale 2:
The liquid and the air come into the reactor by means of a distribution system of small holes (1 cm for the gas and 10 cm for the liquid) in a pipe.
The speed in the holes is approx. 1-5 m/s

I made the mesh with Salome (netgen - moderate)
I tried to solve it with interFoam (1.7) but the problem is the timescale. Because the wholes in the distribution pipe need small cells and there also is a high velocity so the courant number gets big easy and so the time step needs to get very small (3e-6 - 3e-11) it would take forever to solve.

Is there a way to simulate the distribution system with one timestep and then have an interface between the distribution system and the reactor. This interface is the input to the reactor and gets calculated with another much larger time step. A big problem with this approach will be the liquid flowing back into the distribution system from the reactor.
So dividing the mesh in two domains with each a different time steps.

Is this a viable approach or is this not possible?
If it is possible, has anybody done this before?
any ideas to the steps to take?

thanks in advance
Wouter