[JayDeeUU]

I'm computing buoyancy-driven circulation within a vessel that possesses symmetry about the vertical axis (see attached images and those at http://www.cfd-online.com/Forums/ope...tml#post791089). To create the needed .stl files I manipulated a sphere (radius 6 m) and a cylinder (radius 0.75 m) in Salome, making various cuts so as to represent the vessel walls by means of four separate shells. The lowest shell ("shell1") is a partial sphere, with a flat base and an open top. "Shell2" has an open base matching the top of shell1, and an open top of radius 0.75 m. Above shell2 is an open cylinder ("cyl1", radius 0.75 m) and finally "cyl2" is a matching cylinder with open base and closed upper face.

The walls of each layer were assigned different temperatures, the fluid itself a uniform initial temperature, and I ran buoyantSimpleFoam. Ideally, the solution should evolve to an azimuthally-symmetric steady state, however I see no sign of that state emerging despite having executed a large number of steps (to a "time" of 15 hours with the "timestep" initially 1 sec and at largest 10 sec - although "time" is metaphorical in the context of this steady-state solver, which drops the partial derivative in time). The residuals had ceased to decrease well before the 15 hours "time". I tried to push the solution on with more demanding convergence settings in controlDict, to no improvement.

As a newcomer to OpenFoam, I'm unsure how to interpret this solution, or more specifically, what could be the cause of the asymmetry that makes it (the solution field) unacceptable in principle. Viewed in paraFoam the mesh produced by snappyHexMesh looks okay (see attached images), although checkMesh did report "2 highly skew faces detected which may impair the quality of the results... 2 faces with concave angles between consecutive edges". It seems to me improbable that a minor irregularity of the mesh could explain the wildly asymmetric solution that is emerging (an indication of which is given in the last of the attached images).

[JayDeeUU]

Unfortunately it had never occurred to me that the gravity vector "g" might be defined to align with the y (or x2) axis rather than with z (or x3)! My axis of symmetry is z, but I had carried over a "g" file that has gravity aligned with y! As a meteorologist I'm habituated to using z as vertical coordinate... but was well aware that's not the case in other fields, so ought to have checked.