Micromixing time predicted

Comparison between measured and theoretically predicted results for micromixing time... 

For non-isothermal or non-linear chemical reactions, the RTD no longer suffices to predict the reactor outlet concentrations. From a Lagrangian perspective, local interactions between fluid elements become important, and thus fluid elements cannot be treated as individual batch reactors. However, an accurate description of fluid-element interactions is strongly dependent on the underlying fluid flow field. For certain types of reactors, one approach for overcoming the lack of a detailed model for the flow field is to input empirical flow correlations into so-called zone models. In these models, the reactor volume is decomposed into a finite collection of well mixed (i.e., CSTR) zones connected at their boundaries by molar fluxes.4 (An example of a zone model for a stirred-tank reactor is shown in Fig. 1.5.) Within each zone, all fluid elements are assumed to be identical (i.e., have the same species concentrations). Physically, this assumption corresponds to assuming that the chemical reactions are slower than the local micromixing time.5... 

This poses the problem of prediction and scale-up of mixing parameters (especially micromixing times). Macromixing is closely related to circulation times, which vary as N (5-1). 

Micromixing Models. Hydrodynamic models have intrinsic levels of micromixing. Examples include laminar flow with or without diffusion and the axial dispersion model. Predictions from such models are used directly without explicit concern for micromixing. The residence time distribution corresponding to the models could be associated with a range of micromixing, but this would be inconsistent with the physical model. 

For the first two cases Da << 1 (slow reactions) and Da >> 1 (very fast reactions) adequate closure models are available in many commercial CFD codes. For the third case, where the time scale for chemical conversion approximately equals the time scale for turbulent micromixing, moment methods are inappropriate and other methods should be used. In this situation the reactor performance may be significantly affected by mixing efficiency. Here the engineer is faced with the difficult problem of predicting the overall conversion and/or selectivity of the chemical process. In the last three decades this problem has received considerable attention in three scientific areas, namely, chemical reaction engineering, fluid mechanics and combustion, and various approaches have been followed. 

It is interesting to note that calculations of turbulent flows during fast chemical reactions, predicted that the chemical reaction rate constant influences the effective diffusion coefficient and accelerates micromixing, due to an increase of the local reactant concentration gradients [13]. The dependence of the lower boundaries of the reaction front macrostructure formation, in particular, the plane and the torch front, which characterise different scales of liquid flow mixing, on the values of the chemical reaction constants is experimental evidence of the correlation between the kinetic and diffusive parameters of the process. At the same time, one can suppose that the formation of the characteristic reaction front macrostructures is defined by the mixing at the macro- and microlevels. 

The fluid flow and tnrbulence in many plant reactors is more complex than in vessels modeled by academics. CFD is useful for understanding the flow patterns in such reactors however, at the time of writing CFD still does a poor job of predicting the local turbulence quantities vital to micromixing and mesomixing analysis. This is another area where greater understanding is needed. 

---