code_saturne User's Forum

Hello !

I use Code-Saturne 5.0.4 from CAELinux 2018.

My problem is about inlet/outlet and conservation of mass flow.

Case description :

You can find all the usefull files here to compile the case :

https://1drv.ms/u/s!Ao3sRlH5bsVXtlkT1T3 ... U?e=Bi5X0u

Direction of flow is given by arrows. The rest is "wall" with thermal flux or not. Velocity at the first inlet is set to 0.6 m/s for 0.18m2 with angle. An other inlet is set to -0.039 m/s (set to exterior) for 0.36m2. The third hole is directly next to the first inlet and is an outlet (0.18m2).


You can find the sum of all mass flow at each boundary for every iteration :
Blue = negative Z velocity
Yellow = positive Z velocity
As we can see, the air goes into bedroom from the outlet despite the thoerical imposed mass flux balance.
The air is theoretically drawn up the outlet when you calculate mass flow balance but it is not the case in the simulation with C-S ! It must be noted that simulation have already run on Ansys Fluent and the results seem correct with Fluent...
I can't understand what is wrong...

Hello,

When using free outlets with non-orthogonal cells in that area (for example tetrahedra), we often have issues, which you will see in the "listing" as the code tries to prevent rentry.

For these zones, it is better:
- either to extrude 1 or 2 cell layers to have a more regular outlet
- or to switch to imposed pressure outlet (but v 5.0.4 might still have a bug in the GUI when more than one such section is present, which does not seem to be the case in this particular mesh).

Regards,

Yvan

Hello !

What do you mean when you say :"Either to extrude 1 or 2 cell layers to have a more regular outlet." ?
Is it an extrusion at the outlet with a regular meshing (parallelepiped) ?

Like this ?

Hello !

an other question !

Why is velocity high at walls like on these captions ?

Hello,

Yes, the extrusion is good, though 1 or 2 layers are enough.

Regards,

Yvan