[richard222]

Hi all,


I am using a transient, two-phase Eulerian model in ANSYS Fluent (ver.16) to model a degassing process in gas-stirred tank. The geometry is a cylindrical tank filled with molten steel (patched to one half of the cylinder height). Argon gas is injected through a circular hole at the centre of the base of the tank (see attached image, where geometry is 2d axisymmetric and domain is rotated so that gas moves from left to right and gravity acts from right to left. X-axis is Symmetry axis).



I want to simulate (a) the flow field produced by the rising gas bubbles and (b) the rate of transfer of a chemical specie (hydrogen) from the steel phase to the argon phase.



When I run the simulation without species transport, it works fine and the gas bubbles rise to the top of the tank with a stable liquid surface.



However with species transport enabled, the gas remains concentrated at the bottom of the tank, rises at an extremely slow rate, and the free surface of the liquid becomes unstable after about 1 second and large waves begin to form (even before the gas bubbles have risen to the surface). The species transport seems to work as a plot of hydrogen concentration in the gas plume shows an increase in hydrogen None of these effects are observed when the species transport model is turned off.


Dimensionality: 2d axisymmetric
Mesh size: 0.015m
Dimensions: 6m tank diameter, 0.06m diameter of injection hole in tank base, 4m tank height.
Primary phase: steel + hydrogen gas (Initial mass fxn of hydrogen=0.000007)
Secondary phase: argon + hydrogen gas (initial mass fxn of hydrogen=0)
Inlet boundary condition: velocity inlet (100% volume fracction argon at 0.5m/s)
Outlet boundary condition: degassing
Species mass transfer model: Equilibrium model, Raoult’s law, Ranz marshall mass transfer coefficient.
Interfacial forces: continuum surface force (surface tension). No drag or lift forces.

I have tried to fix the problem by doing the following things: using 2d planar or 3d geometry, reducing the timestep, reducing grid size, different outlet boundary conditions (pressure outlet, outflow), changing the mass transfer coefficient model, changing the initial hydrogen concentration (mass fraction), disabling the diffusion energy source, changing the turbulence model and enabling drag and lift forces. None of these changes prevented the problem with the instabilities at the liquid surface.


Any ideas on what is causing this problem?