The flow pattern map coordinates (Us,l, Us,g) presented in Fig. 6,
are established from eight flow-rates pairs (Ql, Qg) and for three liquid
solutions. In the present study, the plotted frontier corresponds
to the transition between the two flow regimes observed
within the experimental range, which are the ‘‘dispersed’’ regime
that allows a complete incorporation of the gas, and the so called
‘‘slug flow’’ regime that occurs at high gas flow-rates. Other types
of existing flow patterns (plug, annular, stratified, etc.) are not observed
over the present range of conditions. Theoretical foam void
fractions are mentioned on the maps by contours of iso-values;
they are straightforwardly deduced of the flow-rates, but of course
the validity is limited to the bubble regime area. Some representative
experimental values taken at the frontiers are gathered in
Table 5. Operating with a set of velocities located at the left of the
limit will ensure that there is no gas loss in the process, and hence
a better process yield. Increasing the gas velocity beyond the frontier
results in a gas waste, then the foam void fraction cannot be
controlled. The limits obtained for SMX10 and SMX+6 seem to be
consistent with the previous hydrodynamic study: for a given liquid
matrix the dispersed regime is reached for a higher gas
flow-rate in the case of a smaller hydraulic diameter in the
SMX+6. Thus, reducing the SMX diameters does not induce radical
changes in the flow map frame, so that no particular issues are
brought to light in the scale-down. It can also be observed that
the limit shifts toward lower gas velocity with increasing liquid
phase viscosity. Most experiments have been made in triplicate
in the case of SMX10 and are at least doubled in the case of
SMX+6 (ranges of deviation are shown for each flow condition in
Fig. 6). Repeat experiments show a 5% standard deviation for the