(239b) Impeller Power Dissipation and Power Number for a Retreat-Blade Impeller Rotating In a Torispherical-Bottomed Tank Under Different Baffling Configurations

Mechanically stirred glass-lined reactors and tanks are commonly used in the pharmaceutical and specialty chemicals industries, in order to avoid material compatibility issues between the vessel and its contents, thus minimizing possible product contamination problems.  Because of the fabrication difficulties associated with glass lining metal surfaces, including baffles and impellers, these reactors are typically equipped with a single retreat curve impeller placed close to the bottom of the tank and a single baffle mounted from the top.  Despite their common use in the industrial practice, these systems have not been extensively studied.

In this work, the power dissipation and the power number, Po, were experimentally obtained as a function of the impeller agitation speed in a 60-L laboratory plastic vessel having the same geometric shape, including a torispherical bottom, and the same geometric ratios of a typical industrial glass-lined reactor system.  A typical scale-down version of single three-blade retreat-blade curved impeller was used in all experiments.  Different baffling systems were studied, including a completely unbaffled tank, a partially baffled tank equipped with a single beaver-tail baffle commonly used in industry, and a fully baffled tank provided with four wall baffles.  Water and aqueous sucrose solutions of different concentrations were used to vary the liquid viscosity and hence the Reynolds number.  Torque was experimentally measured with a strain gage-based rotary torque transducer connected to an external multi-channel signal conditioner, display, and controller. 

At lower Reynolds numbers (Re@100), all systems produced similar values of the power number, irrespective of baffling.  However, as Re increased, the power dissipated by the baffled system increased appreciably above that of the partially baffled system and especially that for the unbaffled system.  In addition, the power number decreased with increasing Re in all cases, but more significantly so for the unbaffled system.  The value of Po approached an asymptotic value of about 0.7 for the baffled system.  However, for the unbaffled system Po decreased with Re over the entire range of Reynolds numbers explored here (100<Re<300,000), finally approaching values for Po in the range 0.2-0.3 at large Re’s.  Po for the partially baffled case was intermediate between the fully baffled and unbaffled cases, with Po approaching 0.5, but was in general more similar in trend to the fully baffled system. 

The Po-vs.-Re experimental data were fitted with a single type of regression curve containing three adjustable parameters, which were found to vary monotonically with increasing baffling effects.  A good fit was obtained between the fitted curve and the experimental data.  This equation can be used to predict the value of Po and hence the power dissipated in industrial systems.