[todda7]

I've been attempting to model air flow in a ventilated room, but I think I've got some of the basics wrong since I'm not too fluent in fluid mechanics.
The room has a simple air inflow with constant velocity and a passive outflow.

The general outline has been this:
Create separate (3D) grids for flags (fluid, obstacle), velocity, density and pressure
Flags contains the room geometry, which is modeled as 'obstacle' all around the boundary except for inflows and outflows.
Velocity grid initialized to 0, density to 1, pressure to 1
Then, loop: * Set density to 1 and velocity to some constant value normal to the walls at the boundary inflows
Advect density then velocity both with 2nd order Lagrange (MacCormack)
Set density to 0 at boundary outflows
Set zero normal velocity on walls/room geometry (no-slip boundary condition)
Perform pressure projection of the velocity grid

I feel a bit lost as to whether my model and parameters are sensible or not. Should the pressure and density grids be initialized to 1? Should pressure be constant at 1 in the inflows and outflows?