Boundary conditions for gas dispersion case
 

Hi all,

I am trying to model the dispersion of a type of gas at a chimney exhaust extending 10m from the ground to the atmosphere. My domain (box domain) extends approximately 400m from each direction of the exhaust and I am using the species transport model.

My question is this : How do I model the farfield of my domain for this sort of simulation?

Currently what I do is have velocity-inlet at the inlet, sides and the top, pressure-outlet at the outlet and wall at the floor of my domain. The velocity at the inlet is 12m/s with the direction specified to be the freestream direction. With this configuration I face a few problems :

1)There is a velocity gradient in the freestream direction. So if I set my inlet at 12m/s, it becomes 14-15m/s when it reaches the exhaust chimney. This is probably due to the pressure at the outlet being set to 0 (gauge). But how can I define a BC that would have constant freestream velocities from the inlet to the chimney exhaust?

2)The flow velocity speeds up quite rapidly (70-80 m/s) before exiting the lower portion of the outlet boundary (closer to the ground) and I am also getting reverse flows at the upper portion of the outlet boundary. I dont understand how this can be? Does anyone have any idea why? On top of this I am getting viscosity ratio limits as well at the vicinity of the outlet (perhaps due to the high velocity gradients of the reverse flow across the outlet).

I have tried numerous things to remedy the problem such as mesh refinement, increasing the domain size, using a domain that expands in the freestream direction etc but all solutions have the same problem, which leads me to think that I have defined the boundary conditions wrongly.

Hope you guys can shed some light on this.

Cheers
DC