Numerical Study of Mixing of Different Newtonian and Non-Newtonian Fluids in Stirred Tank

Abstract

Mixing has the most common occurrence in process industries like chemical, food and polymer and plays a significant part in overall success of the processes. Stirred tanks are commonly used for mixing various types of Newtonian and non-Newtonian liquids. Impeller is the movable part and is used as the rotating device in stirred tank systems for achieving mixing. An impeller while it rotates imparts shear force in the vicinity along the peripheral zone. Literature is rich with information on various experimental and theoretical findings on the hydrodynamics and mixing behaviour of Newtonian fluids in stirred tank systems. However, with non-Newtonian fluids, limited published literature is available on the hydrodynamic behaviour of the mixing process in stirred vessels. A few available experimental works in literatures successfully explained the mixing process in a non-Newtonian system using Rushton turbine (impeller commonly used in industry). But unavailability of the theoretical prediction of the same is basically explains the motivation behind the study on the mixing of non-Newtonian fluids in stirred tank with Rushton turbine.
For mixing highly viscous liquids, helical ribbon impellers are most suited. In this thesis work, it was aimed to study the computational aspects of the hydrodynamic performance of helical ribbon impeller in a highly viscous non-Newtonian system and comparing the results with helical screw ribbon impeller through computational fluid dynamics (CFD) simulation. Entropy generation minimization study is an integral part of this thesis work. Mostly, the earlier works involve use of analytical expressions from basics of mass, energy and entropy balance which has got certain limitations because of many assumptions. Here, we aimed for a detailed numerical study on the same. Also, the understanding of residence time distribution (RTD) study in a stirred tank system gives an idea on the distribution of flow structure. Although, this particular aspect has been studied by various research groups, however, some of the experimental data are not compared with numerical findings for validation. In this work it was aimed to predict RTD numerically especially by using swept volume of the impeller into consideration.
A computational fluid dynamics study using Ansys Fluent was carried out to determine the mixing performance of a tank stirred with Rushton turbine. The predicted profiles of the velocity components were validated with literature data. The non-parametric Spearman’s rank order test was used to find the interaction of velocity profiles with the impeller Reynolds number and flow behavior index. The characteristic performance parameters such as power number and flow number of the impeller were predicted. The variations of entropy generation due to only viscous dissipation with Reynolds number, tank geometry, etc. were calculated for the isothermal tank. The entropy generation minimization (EGM) approach was used to optimize the performance of the non-isothermal continuous stirred tank with respect to the system parameters like inlet Reynolds number, impeller speed, and impeller clearance and impeller blade width.
The numerical study of the stirred tank with helical ribbon (HR) and helical ribbon with screw (HRS) impellers was carried out successfully. The CFD models were successfully validated with the experimental power number given in literature. The power constant for Newtonian fluid (Kp) and non-Newtonian fluid (Kp(n)) were calculated and compared successfully with the literature data. The Metzner Otto or geometry constant, Ks were computed following four different methods and the best one was identified by predicting successfully the generalized power curve. The flow numbers of HRS impeller were predicted for wide range of impeller Reynolds number. The non-dimensional mixing times were varied in scattered way with impeller Reynolds number, and the dispersive flow away from the impeller shaft was observed. The entropy generations were increased with the impeller Reynolds number, and an empirical model of entropy generation with impeller Reynolds number was developed. The non-isothermal stirred tank with HR and HRS impellers were optimized employing the entropy generation minimization technique.
The hydrodynamic and the residence time distribution (RTD) behavior of the viscous Newtonian fluid was studied using a tracer age distribution function, I(θ). The experimental tracer age distribution functions were predicted by CFD tools using tracer injection and swept volume methods. The predicted results were found in good agreement with the literature data. The mixing behaviour was changed from dispersion to ideal mixing state with increasing the tank Reynolds number and impeller rotations. The mixing performance parameters like holdback, segregation, number of ideal continuous stirred tank in series equivalent to single actual continuous stirred tank were also calculated to identify the necessary flow parameters and their magnitude to obtain the ideal flow distribution in the tank.