The network-of-zones concept provides a simplified basis for construct
ing mixing models appropriate to tomography. For miscible single-phase
mixing in three-dimensions, of the order of 10(4) volume elements (vo
xels) are needed to interpret mixing results from 16 element electrica
l resistance tomographic (ERT) sensing rings stacked axially. Experime
ntal mixing and dispersion in passive scalar pulse tests can be acquir
ed at video frame rates by ERT and simulated by integration of sets of
first-order ordinary differential equations. Non-uniformities in gas
hold-up and solids suspension can also be detected and quantified in t
hree-dimensions by resistance tomography. Modelling the approximately
axisymmetric mixing of these two-phase examples can be achieved by a m
uch smaller number of algebraic equations. Tomographic modelling provi
des new insights into the fundamentals of mixing in a stirred vessel.
The reduced computational requirements of networks-of-zones make real-
time model-based control more feasible. (C) 1997 Elsevier Science Ltd.
All rights reserved.