The response of a continuous flow mixing system to a step change in an input variable is discussed from the standpoint of dimensional analysis. This idea is presented as an alternative to using zone models for predicting residence time distributions in systems where imperfect mixing occurs.
Mixing experiments were performed in a cylindrical flat bottomed tank geometrically similar to tanks commonly used for industrial processes. The response of the system to a step change in feed concentration was observed. Initially the tank contained a salt solution. At the start of an experiment, a stream of salt-free diluent was introduced at the top of the tank and a stream of the salt solution was drained from the bottom, keeping the liquid volume in the tank constant. The salt concentration in the output stream was measured continuously after the start of the experiment.
The experimental results are correlated in terms of dimensionless variables, and the variables affecting the mixing process most strongly are determined. The results show that variations in throughput rate or in impeller shape or rotational speed affect the mixing process much more than does impeller position. The data are also compared with models proposed by other authors, and the model constants are evaluated.