In CASE you find the simulation using the k-omega model

1) Regarding the splitting of the hexa cells I reimported the .med in Salome and it looks fine. The hexa I created are still there. Is it possible that the splitting of the hexa is only in paraview?
In the checkmesh file I found:
  Nombre d'éléments                   tria3 :   132132
  Nombre d'éléments                   quad4 :   149900
  Nombre d'éléments                  penta6 :    66066
  Nombre d'éléments                   hexa8 :    73800
so to me it seams only a graphical thing.

2) As you cas see comparing the results of the calculation between CS 2.2.0 and 3-Beta the flow is unphysical in the latter one. There is no flow separation. The turbulent viscosity is particulary strange but I don't think it's the only problem. I tried to disable the turbulence model and the result is the same. I also tested the velocity-pressure coupling with PISO but there is no improvement. For comparison, please find attashed the settings for CS 2.2.0 that produced the good results.
In the simolation the Co Number = 6, maybe a bit high but it worked with CS 2.2.0 and I also tested smaller time steps up to dt = 0.1 s resulting in Co Max = 2

3) Also in this case if I activate the Neumann pressure gradient extrapolation at boundaries I have pressure waves at the Inlet. Just to repeat what I wrote the other da on the forum, this doesn't seems to happen if the mesh is unstructured at the boundary.

PS) I also found a minor bug: once the case has been defined with a turbulence model, if this is disabled the calculation doesn't start correctly. It is necessary to edit the .xml to eliminate the turbulence initialization and boundary conditions. This is at least what happens when foumulas to specify k and omega are used.

All the screenshots are taken at T=600 s

Thenk you so much for the support again!!!!
