Computation of mixing and chemical reaction in stirred vessels

Abstract

Chemical reactions carried out in stirred vessels in industrial practice are commonly complex in that they have several steps which, for example, may complete for a reactant and form different products from it. Usually, one of these products is desired whereas the other is a waste product giving rise to extra costs of recycling or disposal. In some cases, the rate and pattern of mixing in the vessel influences the reaction selectivity to the desired product. One example of this occurs when one of the competing reactions is fast (its rate being mixing-controlled), while the other is slower. In such cases, it is found that the only successful predictive design or scale-up method available is to model the mixing process in detail and superpose the reaction kinetics, using computational fluid dynamics. This paper presents a suitable computational procedure, which first calculates the three-dimensional flow field in a stirred vessel, then solves the equations describing the mixing and chemical reactions that will occur at each place and time in that calculated flow field. Special attention is given to formulating the concentration transport equations in a manner that is efficient and simple to apply to various reaction schemes. The approach adopted uses the concepts of mixture fraction and reaction extent, and enables the program to handle almost any reaction scheme without any additional programming. Macro- and micromixing regimes are included. In the latter case, consideration is given to the interaction of turbulence and the local reaction rate, using a simple model based on a local mixing time scale. Example calculations are presented that show the ability of the method to calculate mixing times and reaction yields. Experiments with a competitive reaction scheme are described and their results compared with computed predictions, in a regime in which the traditional rules of thumb for scale-up are unsuccessful.