Comments to “CFD simulation of stirred tanks: Comparison of turbulence models. Part I: Radial flow impellers” and “Part II: Axial flow impellers, multiple impellers and multiphase dispersions”

Abstract

Iread with great interest, articles of Joshi et al. (2011a,b) “CFD Simulation of Stirred Tanks: Comparison of Turbulence Models. Part I: Radial Flow Impellers” and “Part II: Axial Flow Impellers, Multiple Impellers and Multiphase Dispersions.” The articles cover a valuable review and analysis of CFD simulation of turbulent flow of mechanically agitated liquid and two-(or more-) phase mixtures with two frequently used high-speed impellers: standard Rushton impeller (RT) and pitched blade down flow turbine (PBT). Analysis of individual CFD procedures is made on the basis of results of laser Doppler velocimetry (LDV) in agitated systems with both investigated impellers. The commented articles bring very useful pieces of knowledge for application of CFD simulation when projecting and designing equipment with high-speed impellers and give many useful advices both to researches and people in industrial praxis for successful research and development of devices with investigated agitators. For deeper understanding and following usage of published items it seems that some of them could be clarified and completed: Experimental LDV data of velocity profiles (axial, radial and tangential) are presented in the time – and ensemble – averaged form, i.e. the average-time velocity is obtained over 360◦ of impeller revolution (Schafer et al., 1998), in contradistinction to the angle-resolved form, where angular (tangential) coordinate between two adjacent impeller blades is taken into consideration. It is not quite clear how the results of CFD simulation were evaluated by the same way when compared with LDV ensembleaveraged data. Moreover each velocity profile in observed agitated system (cylindrical baffled vessel with axially located impeller) depends on tangential position of the radial ray where the experiments were carried out. It seems from Figure 4 (Joshi et al., 2011a) and Figure 1 (Joshi et al., 2011b) that location of experimental radial rays was between opposite baffles, but presented experimentally determined radial profiles of all three components of the mean ensemble-average velocities better correspond to the radial ray in the midplane between adjacent baffles. The authors of commented articles described design of both investigated impellers. For PBT 30, 45 and 60 they do not mention a number of their blades, nor relative width of the blade. For RT they mentioned “standard Rushton turbine” and it is hoped that this impeller has six blades and exhibits the standardized simplexes of geometric similarity. Nevertheless, the power input of RT mentioned in the article of Joshi et al. (2011a) depends significantly on the thickness of the RT disk t (Bujalski et al., 1987; Rutheford et al., 1996) according to the formula for the impeller power number: