Separation flowsheet synthesis examples

In industry, it is necessary to deal with very complicated mixtures for which structural matrices can serve as an instrument of separation flowsheets synthesis. In Wahnschafft (1997), the example of plant for separation of coal tar in South Africa (20 components, more than 200 azeotropes) consisting of 40 columns is given. 

Choice of Process Configuration. The sequence of processing steps and flowsheet configuration must be selected concurrently with the choice of extraction chemistry. Some of the established heuristics of separation process synthesis may be helpful here. For example, in multimetal fractionation it is probably better to tiy selective stripping from a single organic stream than to do multiple selective-extraction oper-... 

In the given example mixture 2,3,4, that is bottom product of the first column, is zeotropic therefore, its separation in the second and third columns presents no difficulty. Figure 8.23 shows nine preferable separation flowsheets for this example, obtained by means of synthesis. All these flowsheets separate four-component mixtures using a minimum number (three) of columns without additional entrain-ers, but with application of autoentrainer. Component 2 or 4 or mixtures 2,4 or 3,4 are used as an autoentrainer. Therefore, each of nine separation flowsheets contains one column with one or two feedings where recycle flow of autoentrainer is brought in. 

In the discussed examples, synthesis of separation flowsheets is carried out automatically on the basis of formal rules. In more comphcated cases, in the process of synthesis there can be no other variants other than variants of obtaining azeotropes as some products. These cases are discussed in the following section. 

As was described in the review of previous work, over the last ten years MINLP optimization models have been reported for the synthesis of process flowsheets, heat-exchanger networks, separation sequences, reactor networks, utility plants, and design of batch processes. Rather than describing in detail each of these works, we will briefly highlight several examples from our research group at Carnegie Mellon to illustrate the capabilities and the current limitations of the MINLP approach. 

Heuristic methods were developed by well-experienced engineers and researchers. The first attempt to develop a systematic heuristic approach for the synthesis of multicomponent separation sequences was made by Siirola and Rudd. Common example is hierarchical heuristic approach.Heuristic rules are applied at five design levels to generate and evaluate the alternatives using economic criteria. The hierarchical heuristic method emphasizes the strategy of decomposition and screening. It allows for a quick location of flowsheet structures that are often near ... 

Within each of the three general approaches toward process synthesis, key decisions are made about the flowsheet design that have a bearing on the operability characteristics of the plant. For example, in a hierarchical procedure (Ref. 6) we will make decisions about whether the plant is batch or continuous, what types of reactors are used, how material is recycled, what methods and sequences of separation are employed, how much energy integration is involved, etc. In a thermodynamic pinch analysis, we typically start with some flowsheet information, but we must then decide what streams or units to include in the analysis, what level of utilities are involved, what thermodynamic targets are used, etc. In an optimization approach, we must decide the scope of the superstructure to use, what physical data to include, what constraints to apply, what disturbances or uncertainties to consider, what objective function to employ, etc (Ref. 7). 

---