Alternative modeling of an perforated plate (not as a porous medium)
 

Is it possible to modelling a perforated plate as a “pressure drop” function of the velocity? I know I can realise that by a porous medium, but therefore I must generate a thin porous domain. Sometimes I got mesh-problems with that.

In CFX Pre is an Option for a “Domain Interface” called “pressure change”. Can I use that instead of a porous medium? Therefore I only need a baffle, which is easier to generate.

I have tried to model it by define the pressure drop as an expression:

-((velocity v)^2) * PressuredropCoefficient[kg m^-3]

But the simulation crashed after some iteration. (Velocity at monitor points rises up to supersonics)

If I use an averaged velocity, the simulation runs ok, but it is not physically correct….

-((areaAve(velocity v)@baffle_side1)^2) * PressuredropCoefficient[kg m^-3]

So anyone have an idea what I can do?

Looks like some numerical instability which is fixed by averaging across an area. I assume the averaging approach is not correct for your application as the flow varies across the plate. In that case I would consider breaking the plate up into several small bits and averaging inside each bit. Then you should get some resolution of the flow variation but still get some numerical stability.