convergence problems when using head loss region

[plaza]

Hi, I need to model a head loss region about 0.1 m thick with a pressure drop of about 800 Pa for a water velocity through the region of about 0.00173 m/s. It is also near an outlet, but I don't think that is the main problem as when I move the region away from the outlet I still have the same problem of divergence. I can get a converged solution easily for small pressure drops of 45 Pa (alpha of 5000) but when I increase alpha to 10000 or above, inflow starts at the outlet, blocking of part of the outlet is carried out by code saturne and the solution diverges. I have reproduced the problem in a simple vertical pipe problem with a tetrahedral mesh, attached in my post last week. I also have setup a prismatic mesh model, attached here including .hdf, .med, and code saturne, but still have the same problem. I would greatly appreciate some help.
Regards,
Floren

[Yohann Eude]

Hi plaza,

The value of 45Pa is obtained from an estimation or from the post process of your result.
How you apply your head losses coefficients (axx = ayy = azz or axx=ayy=0 , azz=5000)?
Also can you send us your listing file of your case, I would like to check your options.

Yohann