Motivation and objects

Many industrial processes which involve blending, suspension, dispersion, heat transfer and chemical reactions are performed using continuously stirred tank reactors (CSTRs). In CSTRs, impeller is used to stir the fluid into a vessel. Heat, mass and momentum transfer depend strongly on the flow field developing in the vessel. Achieving a good mixing in the vessel is costly and time-consuming and CSTRs usually represent the economical bottleneck of a whole industrial process. Therefore, research work [Ranade and Joshi, 1991, Armenante and Chou, 1996, Kresta and Wood, 1993a, Kresta and Wood, 1993b, Sahu and Joshi, 1995, Sahu et al.,1999, Yianneskis et al., 1987, and Zhou and Kresta, 1996] is now aimed at optimizing the equipment, the geometry and the operative conditions in CSTRs. Extensive experimental investigation and computational modeling are used to this end. In most cases, due to costs and objective difficulties, experiments are limited to pilot scale vessel, which allow to examine power consumption and flow field in different operating conditions. Detailed information on the flow field developing in the vessel can be obtained by experiments only using complex and costly methods like Particle Image Velocimetry (PIV) and Laser Doppler Anemometry (LDA) [Jaworski et al., 1991, Ranade and Joshi, 1989]. A flexible alternative to sophisticated experimental methods is offered by computational methods. Once the computational techniques have been validated by suitable experiments, they can be used to optimize the number of experiments and to test new concepts, geometries and operative conditions.

In this work, experiments and computational modeling are used to characterize the fluid dynamics of industrial size CSTRs. A laboratory scale CSTR is examined in the first place, using both experiments and numerical modeling. Experimental data are used to validate the computational methods used in this work. Second, numerical simulation for the laboratory scale vessel are compared with simulations made for the industrial size CSTR to verify scale-up possibilities. Third, different configurations of the industrial size CSTR are examined: in particular, the pumping capability and the pumping efficiency of two different impellers are evaluated. This piece of research is a prerequisite which is necessary in order to obtain design guidelines leading to optimization of industrial CSTRs of the type investigated. The main object of this work is to obtain relationships between agitation and circulation and the power consumption which compare with previous experiments and existing empirical correlations [Nagata, 1975]. These evaluations are expressed through proper dimensionless groups $-$ Reynolds number, Power number, Discharge flow number, Secondary circulation flow number, Pumping efficiency $-$ and can be used as a design tool. This object is obtained performing computational simulations of the fully three dimensional and time dependent flow field in three different CSTRs: i) one laboratory scale CSTR; ii) reactor CE12500, equipped with retreated curved blade impeller; iii)i reactor BE12500, equipped with the turbofoil turbine. Further goals are listed as follows: comparison between experimental data collected for the laboratory scale reactor and results of numerical simulations allow to assess the modeling capability of the computer code used in this work. Comparison between results obtained for the laboratory scale and the industrial size vessel CE12500 allows identification of possible scale-up problems. Comparison between CE12500 and BE12500 allows to investigate on the fluid mechanics, the efficiency and the power consumption of the two different impellers.