Pitched-blade stirrer

A typical stirred-tank reactor is shown in Fig. 5.4-3. It is a cylindrical vessel with elliptical or torospherical bottom and cover. It is equipped with an axially mounted stirrer rotating with a speed from 25 rpm (large scale) to 2000 rpm (laboratory). Fig. 5.4-4 shows the stirrers that are mostly used in fine chemicals manufacture, viz. the marine propeller, turbine, flat- or pitched-blade agitator, and anchor. Agitators move the fluid into axial and radial direction. Marine propellers and pitched-blade stirrers predominantly impose axial motion. 

The maximum dissipation rates for the dispersion of droplets and gas bubbles with different stirrer types such as turbine and pitched-blade stirrers, Lightnin A 310 and Cheminer HE 3 was determined using ID-LDA measurements, and the resulting turbulent fluctuating velocities were calculated using a model based on dimensional analysis. The essential parameters were identified using statistical analysis supported by sensitivity analysis. It was found that in the relationship e oc the stirrer speed dependence with n is correct, but that for the stirrer diameter with was set too low. The number of baffles was of secondary importance [608]. 

The proportionality constant fe depends upon the stirrer type. For propeller stirrers fe = 10 ]68], for turbine stirrers fe = 11.5 and pitched-blade stirrers fe = 13 [104]. For blade stirrers k = 2.5, for cross-beam stirrer fe = 4.1 and for helical ribbon stirrer fe = 6.0 [411]. Calderbank [66] found that when turbine stirrers were used with Bingham and pseudoplastic fluids fe = 10 and when used with dilatant liquids fe = 12.8 (d/D) 5. Lower fe values were found as the viscoelasticity of fluids increased [104]. In the case of close-clearance anchor stirrer fe depended upon the wall clearance [24]. 

H. Judat has comprehensively extended this work. In [246] Ne(Q) relationships are given for different stirrer types, namely for perforated disk stirrers with 12 holes (after Brauer [48]), for 3-wane propeller stirrers with a pitch a = 26°, for pitched-blade stirrer with a pitch a = 45° and for conventional turbine stirrers with different numbers of blades i = 6 12 18 24 (Fig. 2.6). The measuring range of Q was extended for turbine stirrers to Q 1.0. It was shown that in the Q range 0.2 to 1.0 expression (2.23) did not apply. With increasing Q, Ne decreases ffirther. 

Fig. 2.6 Power characteristic Ne(Q ) in the air/water system for four different stirrer types turbine stirrer with i = 6 and 18 paddles pitched blade stirrer propeller stirrer perforated disk stirrer from [246].

In the turbulent range, Jahoda and Machon [234] found no difference between the n6 Re) values of a turbine stirrer and that of a pitched blade stirrer (a = 45°). Since the latter, under otherwise the same conditions, only required a quarter of the power, it is more suitable for homogenization. 

This concept for the determination of the effective viscosity for mixing processes is also transferable to stirred vessels, because in the laminar flow range the above circumstances also apply for propeller loop reactors [351]. This was proved with tanks with pitched-blade stirrers [540]. The nO values could be better correlated with RSefr numbers, which were produced according to expression (3.25), than those which contained ft g according to expression (1.45) after Metzner-Otto. 

W. Muller (Technical University Dortmund) [384] has utilized differently concentrated aqueous solutions of glucose (Newtonian), CMC (pseudoplastic) and PAA (viscoelastic) in experiments in baffled tanks with pitched-blade stirrers and found that in the Re,.fr range of Repff = 10 -10 the nOg values were about a factor of 100 to 10 higher with PAA than with glucose, whereas these values for CMC and glucose only differed by a factor of 5. Only above Rdeff = IO did the non-Newtonian properties cease to have an effect on this relationship (Fig. 3.8). 

Fig, 3.8 Mixing time characteristics for a pitched-blade stirrer in baffled tanks with Newtonian and non-Newtonian fluids in the conventional representation. For legend see Fig. 3.9 from [384]. 

Measurements with the pitched-blade stirrer (type PTD) could not be well correlated in the framework of expression (4.64) instead of which it was found that [461] ... 

Buurman et al. [63] have established through measurements (l-s criterion) in two differently sized tanks (D = 4.3 and 0.48 m, H/D= 1), equipped with a 4-pitched blade stirrer (a = 45°, d/D = 0.4), that in the flow range investigated by them the scale-up rule P/V = const applies and that thin stirrer paddles are advantageous for saving power, because they produce eddies in the lower range of the... 

Judat [244] investigated the dispersion action of rapidly rotating stirrer types 3-vane propeller stirrer, 6-pitched-blade stirrer (a = 45°), 6-blade turbine stirrer and a rotor-stator arrangement (rotor 6-turbine stirrer, stator 24 vertical paddles whose angle could be varied between —60° and +60° with respect to the radial position). 

Due to the liquid circulation caused by the stirrer in the tank, helical coils should only be used with axially operating stirrers (propeller and pitched-blade stirrers), which are, on the other hand, not suitable for gas dispersion. Thus in bioreactors vertical heat exchange tube bundles are preferably used as baffles (see Fig. 1.3c). Occasionally, baffles are executed as hollow bodies, so that they can serve as heat exchangers. 

In striking contradiction with the above-discussed experimental data, are measurements [358] of heat transfer for suspensions of macromolecular materials (copolymers with dp = 0.1-2 mm and p = 1054 or 1025 kg/m = 34%) in aqueous polyvinyl alcohol or gelatin solution of comparable density, in which a pitched-blade stirrer with 2 blades (h = 45° D/d = 2 h/d = 0.1) and an anchor stirrer with D/d = 1.07, h/d = 0.1 were utilized. The tank of D = 300 mm with a hemispherical bottom was not baffled. The result... 

---