Stirred Vessels Liquid Mixing Time

There is some conflict in the literature as to the effect of gassing on liquid mixing time in the homogeneous regime. However, the effects can all be related to the 

When multiple impellers are used, care must be taken in their selection for gas-liquid systems. For example, a vessel of H = 3T with three radial flow Rush-ton turbines gives rise to compartmentalization of the flow with poor overall top-to-bottom mixing mixing times can be very much longer than with similar specific power input in a tank with H = T mixed by a single impeller (Cooke et al., 1988). Mixing times for a combination of one to three radial flow impellers for Re 4400 were well correlated by 

CFD (as described above) has been used to predict mixing times in liquid-phase systems. In the authors experience, once the tracer input condition has been carefully modeled to match an experiment, reasonable agreement with experimental values has been obtained for axial flow impellers. For radial flow impellers, the predicted mixing times were longer than the measured values this appears to be caused by inadequate description of the vertical transfer between the blade vortices in the impeller discharge stream. The effects of gassing have, however, not been explored. 

This section deals only with turbulent flow conditions. In this regime power dissipation is the controlling factor for mixing and phase dispersion. For in-line mixers the power is derived from the flow energy of the fluid, and for stirred vessels it is obtained from the impeller and, where density differences occur, from buoyancy forces. 

Noting that power = volumetric flow rate x pressure drop, the overall power per unit mass of liquid is straightforward to calculate for single-phase systems given the friction factors and voidage fraction in the mixer as supplied by mixer manufacturers or measured in the laboratory. For gas-liquid systems the volume of fluid in the mixer must be multiplied by (1 — c )) to obtain the liquid volume, so the gas fraction 4) must be known (see Section 11-5). It has been found that the Lockhart-Martinelli (1944) correction for the effect of the gas phase on pressure drop in pipe flow can be applied to static mixers with reasonable accuracy ( 20%). 

---

Continuously operated stirred tanks can only attain homogenization if q > q (where q is the liquid throughput through the vessel). The mixing time is then 0 oc V/q (where V is the liquid volume of the vessel). Knowledge of q, in the case of propeller stirrers, enables the flow velocity along the heat exchanging surfaces of one or more concentric cylindrical coils round the stirrer (and consequently the heat transfer rates) to be calculated. Such an installation may be indispensable in case of extremely exothermic reactions. 

One of the main tasks of the stirred tank is homogenization of the inflowing starting material and the reaction mixture with the aid of a suitable stirrer. To maintain a homogeneous reaction mixture in the vessel, the mixing time should be at most 10 % of the time constant of the reaction (initial concentration divided by the reaction rate). Apart from the use of stirrers, it is also possible to mix the reactor contents by using jet mixers (injection of the circulating liquid) and loop reactors. 

Liquid residence-time distributions in mechanically stirred gas-liquid-solid operations have apparently not been studied as such. It seems a safe assumption that these systems under normal operating conditions may be considered as perfectly mixed vessels. Van de Vusse (V3) have discussed some aspects of liquid flow in stirred slurry reactors. 

The calculation of k using Eqs. 9.2.11 and 9.2.12 requires a priori estimation of the exposure time or the surface renewal rate s. In some cases this is possible. For bubbles rising in a liquid the exposure time is the time the bubble takes to rise its own diameter. In other words, the jacket of the bubble is renewed every time it moves a diameter. If we consider the flow of a liquid over a packing, when the liquid film is mixed at the junction between the packing elements, then is the time for the liquid to flow over a packing element. For flow of liquid in laminar jets and in thin films, the exposure time is known but in these cases it may be important to take into account the distribution of velocities along the interface. In the penetration model, this velocity profile is assumed to be flat (i.e., plug flow). For gas-liquid mass transfer in stirred vessels, the renewal frequency in the Danckwerts model s may be related to the speed of rotation (see Sherwood et al. 1975). 

In a stirred reactor, it takes time for liquid added at the surface or at any point in the tank to become blended with the bulk of the liquid. The mixing process can be followed by observing the color change after a basic solution with an indicator is neutralized by suddenly adding a slight excess of acid. If this test is done in a 2-liter vessel, most of the solution will appear free of base in less than a second, but wisps of color may persist for 2-3 seconds until the mixing is complete. If the same test is carried out in a 5000-liter... 

Consider a single-phase liquid in a stirred tank to which a volume of tracer material is added. The mixing time is the time measured from the instant of addition until the vessel contents have reached a specified degree of uniformity when the system is said to be mixed . 

---