16

FIGURE 10.32 Influence of the number of feed inlets on selectivity (from Ranade, 1993).

stirred reactors. For example, maps of volumetric mass transfer coefficients (kLa) for gas-liquid stirred reactors can be generated based on information about turbulence characteristics predicted using CFD models and empirical relationships between these characteristics and bubble diameters. Such maps are used to identify regions of high mass transfer rates and regions of high driving (for mass transfer) force (near feed pipe). It makes sense to provide the highest mass transfer rates in the region where concentration-driving force is also highest. Such a provision will be much more productive than the case where high mass transfer rates exist in a region of low concentration-driving force. Such models can therefore be used to select or to devise the best-suited reactor configuration to achieve the best mass transfer performance.

CFD models will also be used to carry out scale-up and scale-down analysis, especially for non-geometric scale-up. In many reactor-engineering applications, it is necessary to carry out laboratory-scale or bench-scale experiments to understand the behavior of large-scale reactors. It is essential to undertake systematic scale-down analysis to ensure that small-scale experiments mimic key features of the large-scale system. More often than not, it is necessary to use a geometrically dissimilar system in order to mimic key features of a large-scale system. Scale-down and scale-up analysis using CFD models may prove to be very valuable in such an endeavor and will help to derive maximum benefit from the small-scale experiments. CFD models also allow extrapolation of cold flow results to actual operating conditions (high temperature and pressure) and provide tools to interpret and extrapolate small-scale experimental data.

Detailed characterization of flow generated by various impellers also leads to extremely useful information about the sensitivity of the impeller fluid dynamics to fabrication details, providing evidence for the well-known saying that 'impellers, which look alike may not perform alike'. The insight gained via computational models will lead to better reactor and process engineering. Creative analysis of CFD simulations may lead to new impeller designs. Ultimately, it may be possible to create 'designer' flow fields to ensure better reactor performance. In many situations, CFD models may be used to generate information which will be used by other sets of models. Compartment models or zones-in-loops models mentioned in Chapter 1 fall into such a category. With a combination of two or more different modeling tools, it is possible to derive useful engineering information. For example, recently Vivaldo-lima et al. (1998) developed combined CFD and compartment models to simulate a suspension polymerization reactor. They showed that the combined model could capture the key features of complex interactions of coalescence and break-up processes with polymerization reactions. Thus, judicious combinations of CFD models with other modeling tools may lead to realistic simulations of complex multiphase reactors. Use of computational flow models to understand basic phenomena and to simulate complex industrial reactors (using a hierarchy of modeling tools) establishes a link between reactor hardware (and operating protocols) and reactor performance and eventually leads to better reactor engineering.

0 0

Post a comment