[ic7]

Hi all,

In modelling a vortex with buoyancy, the axial (vertical) velocity does not seem to be coupled with the buoyant effects despite residual convergence.


Under the conditions listed below, the axial temperature difference between the plume and the surroundings reduces to near zero at about 1/3 of the domain height; it is logical that we would observe a decrease in buoyant effects. However, the observed axial velocity remains at 100% of the specified inlet value for the entire span of the domain. A reduction in the temperature differential seems to have no effect on the behaviour of the axial velocity in this regard.

• Axisymmetric Swirl/Standard k-omega
• Air density - piecewise linear
• Rectangular Domain (~200000 mesh elements)
  • Height - 2.1 [km]
  • Width - 0.9 [km]
• Inlet (r=3 [m]):
  • Temperature differential - 20 [K]
  • Axial Velocity - 2.0 [m/s]
  • Swirl Velocity - 0.5 [m/s]
• Both the outer and upper limits are pressure outlets

After finding the following, I am led to believe that that the k-omega turbulence models do not account for the inclusion of buoyancy effects on the specific dissipation rate. Is this correct?

What model would you recommend in order to realise the effects of buoyancy in a swirling plume?

Best,
Ian

[ic7]

To follow up, k-omega turbulence models indeed do not account for buoyant effects on the the specific dissipation rate. For this particular case (axisymmetric buoyant jet with swirl), the RNG k-epsilon model with a swirl modification appears to be best suited.

On the standard, RNG and realizable k-epsilon models

Ian