Most stable way to initialize this domain
 

Hi,
I'm looking for the most stable way to initialize this compressor setup:

https://www.cfd-online.com/Forums/at...1&d=1497332394

The sections marked with 1 should get suction side conditions (20[C], 1[bar]); the region labeled 2 discharge conditions (80[C], 3 [bar]). I tried using a CEL with inside() [which works well] but this gets unstable as soon as I try to implement a mechanical solver for 2 way FSI (the initial load has to be to much). Is there a more stable way or should I calculate 2 separate initialization file for both mechanical as well as CFX and try to start from there?

For the sake of completeness: 4 is the inlet, 3 is the piston that moves downwards as soon as this is stable ...

Thanks and regards

You are specifying an initial condition with 2 bar jumps in pressure. If this simulation is compressible this will cause shockwaves to fly in all directions (which is highly unstable). If this simulation is incompressible then the flow will go bezerk as it tries to find a rational solution to this initial condition.

So the problem is your unrealistic initial condition, not CFX.

If this is a reciprocating machine, often the best way to initialise them is to have atmospheric pressure everywhere and then run it for a few cycles to establish a realistic initial condition.

Hi,
well running several cycles is challenging on computation time. And as I said, its working as long as there is no coupling in between, but I take your criticism and think over the initialization process again.