In the present work, the relationship between the flow pattern and blending
has been investigated. The flow patterns generated by around 40 axial flow
impellers have been examined. The impellers differed in blade angle, blade
twist, blade width, impeller diameter, impeller location, and pumping dire
ction. The mean-flow and turbulence characteristics generated by all of the
impellers have been measured using laser doppler velocimetry (LDV). On the
basis of available LDV data, the flow pattern throughout the vessel was es
tablished by employing computational fluid dynamics (CFD) and subsequently
used for the simulation of the blending process. The predicted mixing times
were found to be in excellent agreement with the experimental measurements
. It has been shown that the dimensionless mixing time (theta) varies inver
sely with the secondary flow number of the impeller. Comparison of the impe
llers on the basis of equal power consumption per unit mass has shown that
theta(mix) proportional to (NpTNQS)-T-1/3-N-2/3/. The present CFD model has
shown the possibility of reducing the eddy diffusivity to about 20% of the
actual value and still achieving the same mixing time. This reduction in e
ddy viscosity represents substantial savings in operating costs.