[Laurence Wallian]

hello,

I suppose other people have already been concerned with a similar case : modelling a warm flux in a rather steady cold chamber.

Usualy, I used to make cold air coming from one side of the chamber at low speed (Velocity-inlet=0.1m/s) and leaving the domain on the opposite side defined as an outflow.

Here is another way to do : an equal number of sources and wells are defined on the limits of the domain, which faces are chosen as velocity_inlet.

as examples you can have a look at files joined:
- Geo3D.txt ( meshing -> gambit -inp Geo3D.txt) : a box having an outlet on its up face, and a velocity-inlet (warm air) on its down face is placed in the cold domain
- UDF.c : UDF for fluent 12.1
- Calcul.txt : Calcul journal file for fluent 12.1 -> fluent 3d -g < Calcul.txt
- image.jpg : sources and wells positions and view in the plane Z=0.5
- debit.txt : mass flow rate on box and domain boundaries

note 1 : as sources and wells are defined by velocity-inlet having the same speed and the same temperature, the global flow rate on these boundaries (Cf. debit.txt) can be bigger by zero du to the mixing warm-cold air : in the example, the difference is equal to 1% of the warm box mass rate . That why, I think this solution can only be used for small temperature difference between warm and cold air

note 2 : to limit the global flow rate on domain boundaries, it is sometimes useful to change the position of sources and wells : in the example here, sources and wells alternate along horizontal lines, for others case, it might be better to have horizontal (or vertical) lines of sources and wells (which is like alternate wells and sources along diagonal..) ?!