| This thesis covers two major fields in mixing: turbulence energy dissipation and liquid-liquid dispersions in agitated tanks. The main objective of the thesis was to examine the effect of tank and impeller geometry on these variables.; The average turbulence energy dissipation in the impeller regions was investigated for three impellers (the Rushton turbine (RT), the pitched blade turbine (PBT) and the fluidfoil turbine (A310)) using the macroscopic mechanical energy equation. The majority of the input power is dissipated in the small volume of the impeller region for all three impellers. Analysis of the distribution of energy between convective and turbulent flow shows that the A310 is the most efficient at generating convective flow; the RT generates the most turbulence kinetic energy and the PBT derives a large portion of its energy from the return flow.; The equation ({dollar}varepsilon{dollar}=Av{dollar}sp3{dollar}/L), used to estimate the local turbulence energy dissipation rate, was verified by comparing local and integral methods. Here v is a fluctuating velocity; A is a constant and L is macroscale length. The effect of rank geometry (number of baffles (N{dollar}sb{lcub}rm f{rcub}{dollar}), impeller diameter (D), and off bottom clearance (C or C/D)) on {dollar}varepsilonsb{lcub}rm max{rcub}{dollar} was investigated using three factorial designs for four impellers (PBT, A310, HE3 and RT). The dominant variable was shown to be the impeller diameter. This effect is in addition to the expected scaling with D{dollar}sp2.{dollar} Clearance is also an important variable, which is best quantified by its dimensionless form, C/D. The number of baffles has no significant effect on {dollar}varepsilonsb{lcub}rm max{rcub}.{dollar}; Silicone oil/water dispersions were also studied with varying tank geometries. The shape of the drop size distribution changes with rotational speed (N). Four characteristic distributions were found; in order of increasing N: long tail, double peak, skew, and skew-normal distribution. Two normal distributions can be combined to represent the last three distributions. The Kolmogoroff length scale ({dollar}eta{dollar}) cannot be used as an estimate of the minimum drop size present in the dispersions investigated. The cumulative number probability density less than {dollar}eta{dollar} is negatively correlated with {dollar}varepsilonsb{lcub}rm max{rcub}.{dollar} The Sauter mean diameter (d{dollar}sb{lcub}32{rcub}){dollar} is more closely correlated to {dollar}varepsilonsb{lcub}rm max{rcub}{dollar} and the interaction of {dollar}varepsilonsb{lcub}rm max{rcub}{dollar} with the mean flow than to P/{dollar}rho{dollar}V{dollar}sb{lcub}rm T{rcub}.{dollar} A new correlation for d{dollar}sb{lcub}32{rcub}{dollar} is proposed. |
