91 Reactor Engineering Methodology

A general reactor engineering methodology is shown in Fig. 1.10. Based on available information concerning the chemistry and catalysis of the process under consideration, the first step in reactor engineering is to select a suitable reactor type. Krishna (1994) discussed a systems approach for reactor selection. He advocates setting up a 'wish list' for reactor selection. The subject of reactor selection is not discussed further here, and interested readers may refer to the original paper (Krishna, 1994). It must, however, be emphasized that setting up of such a 'wish list' is one of the most important steps not only for the selection of reactor type but also for any reactor engineering or mathematical modeling activity. The success of the application of mathematical (or otherwise) modeling to any reactor engineering project depends on setting up such 'wish lists' which act as maps or guides for the selection and application of relevant tools. The results obtained by these various tools and the 'wish lists' are used to evolve a suitable reactor engineering solution.

For the development of a new reactor technology, a typical 'wish list' could be (from Krishna, 1994):

• operability within technologically feasible region;

• intrinsically safe operations;

• environmentally acceptable;

• maximum possible conversion of feed stocks;

• maximum selectivity of reaction to the desired products;

• acceptable impurity profiles;

• lowest capital and operating costs.

To enhance the performance of an existing reactor technology/hardware, a typical wish list could be:

• more throughput per unit volume;

• improved selectivity and better quality product;

• safer operation;

• reduced energy consumption;

• more environment friendly operation.

The next step is to translate the wish list into a quantitative form and establish a relationship between items in the wish list and reactor hardware and operating protocols. The reactor engineer's task is to design and tailor the reactor hardware and operating protocols to realize the wish list. Several activities are involved in this process. It may often turn out that some of the items in the wish list require contradictory options of hardware and operation. In such cases, a careful analysis of different items in the wish list must be made to assign priorities. Operability, stability and environmental constraints often receive precedence over throughput and energy consumption when such conflicting requirements arise.

Some of the tasks of the reactor engineer when establishing the relationship between reactor configuration/operation and performance are shown in Fig. 1.10. Examination of these tasks emphasizes the need for developing a multilayer modeling strategy. Some of the tasks, such as examining the influence of reactant flow rate and operating temperature on the performance of the reactor (conversion, selectivity, stability and so on), can be answered by developing conventional reaction engineering models. In these models, some assumptions are made regarding the flow and mixing of various species in the reactor, instead of solving the fluid dynamics equations. Thus, although these models cannot directly relate the reactor hardware with performance, these models are computationally much less demanding than CFD-based models and can give a quick understanding of the overall behavior of the reactor. These models can be used to identify the important parameters/issues, which may require further study. Of course, the class of conventional chemical reaction engineering models itself contains a variety of models. It will be useful to distinguish between 'learning' models and 'design' models at this stage.

'Learning' models are developed to help to understand basic concepts and to obtain specific information about unknown processes. The results obtainable from such models may not lead directly to design information but are generally useful to take appropriate engineering decisions. 'Design' models, on the other hand, yield information or results, which can be used directly for reactor design and engineering. It is first necessary to develop design models to estimate reactor sizing and to evolve a preliminary reactor configuration. Several 'learning' models can then be developed to help understand various reactor engineering issues, such as:

• start-up and shut-down dynamics;

• multiplicity and stability of thermo-chemical processes occurring in the reactor;

• sensitivity of reactor performance with respect to mixing and residence time distributions;

• selectivity and by-product formations.

The understanding gained by development and application of these 'learning' models is helpful in identifying the needs for developing more sophisticated simulation models to establish the desired reactor design. These models are also useful in identifying the likely impact of reactor fluid dynamics on reactor performance. The results allow the reactor engineer to identify gaps between available knowledge and that required to fulfill the 'wish list'. The identified gaps can then be bridged by carrying out experiments in the laboratory and/or pilot plant(s), and by developing more comprehensive fluid dynamic models.

Computational flow modeling enters the reactor engineering activity at this point. Despite the advantages, conventional chemical reaction engineering models will not be directly useful for understanding the influence of reactor hardware on reactor performance. For example, how the design of the distributor for dispersed phase affects the radial distribution of dispersed phase and thereby the reactor performance, will be difficult to predict without developing a detailed fluid dynamic model (CFM) of the reactor or without carrying out experiments on a scale model. The CFM-based approach will make valuable contributions at this stage by providing the required insight, by helping to devise the right kind of experiments and by allowing the screening of alternative configurations and by providing tools for extrapolations and scale-up. Of course, the whole process of reactor engineering is not sequential! All steps interact with and influence each other. The results obtained in laboratory experiments on hydrodynamics and residence time distribution (RTD) or from the computational flow model may demand changes and revisions in the earlier analysis and the whole process is iterated until a satisfactory solution emerges. In this book, we are particularly concerned with the application of computational flow modeling to obtain the relevant information about reactor engineering. Translating reactor engineering requirements to formulate suitable flow models and the use of such flow models for reactor engineering is illustrated here with the help of a few examples.

Before we discuss the examples, some general comments on CFM for reactor engineering will be useful. Computational flow models can be built either as 'learning' models or 'design' models. For 'design' models, which are expected to yield directly applicable design results, relating reactor engineering objectives to computational flow modeling objectives is relatively simple and straightforward. Some special types of reactors, such as chemical vapor deposition reactors, are designed directly based on a comprehensive computational flow model. Such comprehensive CFD models enable the reactor engineer to directly relate reactor hardware (and operating protocols) to reactor performance. In several other cases, however, it may be necessary to use computational flow models to assist the process of reactor engineering decision-making. In such cases, correct formulation of the flow problem plays a crucial role. Developing computational models to obtain the required information about the behavior of industrial chemical reactors is a complex task and requires specialized knowledge and approach. Previous chapters have provided basic information about the elements of computational flow modeling. Part IV of this book contains separate chapters on three major reactor types, namely stirred reactors, bubble column reactors and flu-idized bed reactors. One chapter is included to cover miscellaneous reactors, along with fixed and trickle bed reactors. These chapters are designed to provide specialized knowledge pertinent to different reactor types, which will assist the reactor engineer wishing to develop reactor flow models. In this chapter, examples are discussed to illustrate the basic methodology and to relate results obtained from computational flow modeling to reactor engineering objectives.

0 0

Post a comment