Hi,

I am simulating air being sucked out of my domain (negative velocity inlet) but only the segregated solver will work, the coupled solver does not. I have strong natural convection in the domain so I would like to use the coupled solver. Is this possible?

Use a pressure outlet to model flow leaving the domain; it has a target mass flow option where you can specify a mass flow rate.

What courant number are you using?

It doesn't have anything to do with the courant number,i'm sure of that. I ramp the courant number from a very small number after initialisation. Inf fact, the solution never actually takes off in the first place, i.e. air is not entrained from the pressure outlet. I would like to know what causes this but If you were successful with the model setup in this way, then I must be doing something else wrong.

The coupled solver is more sensitive to initialization than the segregated so one way round is to run the segregated solver for 50 iterations or so until the flowfield starts to set up then turn to the coupled solver to resolve the field to an acceptable level of accuracy.

You can also try to use a pressure ramp...creat a field function or use the standard ramp tool at your boundary (a pressure boundary as suggested before) and ramp the pressure slowly from atm to desired value...that should work! (already tried)
Bye