D. Pfleger et al., Hydrodynamic simulations of laboratory scale bubble columns fundamental studies of the Eulerian-Eulerian modelling approach, CHEM ENG SC, 54(21), 1999, pp. 5091-5099
The Eulerian-Eulerian model is used for the hydrodynamic simulation of a tw
o-phase gas-liquid flow in a laboratory scale bubble column. The behaviour
of the air-water system characterises a test case for bubbly flow with low
gas void fractions. A dynamic lest case with a centred gas sparger is chose
n for validation of the simulation models. Long-time-averaged liquid veloci
ty profiles and time series at specific points are compared with experiment
al data. The main focus lays on the influence of turbulence modelling. Lami
nar and turbulent simulations are carried out. A standard k-epsilon model i
s used to describe turbulence occurring in the continuous fluid. Additional
ly turbulent dispersion of the gas bubbles can be taken into consideration.
The results show that a turbulent model has to be considered to gain corre
ct results. The laminar model shows a chaotic behaviour and not the harmoni
c oscillations observed in experiments. In contrast good agreement of the r
esults can be obtained for three-dimensional calculations including turbule
nce. Distinct modelling of turbulent dispersion seems not to be necessary f
or the chosen test case. Furthermore, it can be concluded that a three-dime
nsional simulation with a sufficient fine resolution is necessary for accur
ate results. The test column depth, which is the determining length scale,
must be resolved meticulously to receive accurate turbulence intensity in t
he bubble column. The conclusion concerning grid refinement and independenc
e still needs further evaluations but fails up to now due to the limited co
mputational power.
Finally, two-phase bubbly flow calculations are carried out successfully wi
th commercial CFD software in a transient way with the two-fluid model. The
computational calculations still need very high resources. Further develop
ments and model discussions are strongly recommended. (C) 1999 Elsevier Sci
ence Ltd. All rights reserved.