Turbulence models for CFD in the 21st century

Brian Spalding, of CHAM Ltd

Invited lecture presented at ACFD 2000, Beijing

Example 6. "Puff-jet" simulations.

• Results will be shown which pertain to two puff-jet simulations which differ only in the value chosen for the major empirical constant of MFM, namely CONMIX, the dimensionless micro-mixing constant.
• The two jets are supposed to be confined within an outer cylindrical pipe, in which the axial velocity and effective viscosity are as indicated here and here.
• The flow is from left to right; and the radial dimension has been increased ten-fold.
• The concentration dimension is divided uniformly into 50 equal intervals; so a 50-fluid model is in question.

  Click below for some animated fluid-concentration contours, first for CONMIX=1.0 and then for CONMIX=10.0.

  • Fluid 50; the suddenly-injected reactant
  • Fluid 1; the other reactant
  • Fluid 32, with a high concentration of injected material
  • Fluid 16, with a moderate concentration of injected material
  • Fluid 8, with a lesser concentration of injected material
  • Fluid 4, with a still smaller concentration of injected material
  • Fluid 2, with a very low concentration of injected material

  Suppose that fluid 32 is the one which produces the most colour. Then compare its concentration contours at the 20th time step for the two CONMIX values. The maximum values in the first case is 1.E6 times that of the second!

• It is evident that the effect of a ten-fold change in CONMIX is profound; therefore it should easily be possible, by suitable analysis of the experimental observations, to determine its appropriate value or functional dependence.

• Of course, unless better techniques are invented, the optical measurements will concern the total amount of radiation from all the fluids in the mixture. However, the above conclusion remains valid.

• A question: Have experiments of the "puff-jet" kind already been reported in the scientific literature?