Given the product specification, the design optimization variables are number of reactive trays Ngx, numbers of rectifying and stripping trays NR and NS, and feed tray locations for heavy and light reactants NFheavy and NFlight. Instead of blindly exhausting all possible combinations of design variables, a systematic design procedure is used. Note that process knowledge is needed to place the reactive zone(s). A simple rule is to place the reactive zone(s) at locations where the reactants are most abundant. For the type Ip process, the reactive zone is located at the midsection of the column.
1. Set the reactants feed ratio to 1 (i.e., FR = Flight/Fheavy = 1).
2. Place the reactive zone in the midsection of the column, and fix the number of reactive trays Nrx..
3. Place heavy reactant feed NFheavy on the top of the reactive zone, and introduce light reactant feed NFught on the lowest tray of the reactive zone.
5. Run the simulation using the dynamic model with feedback control to meet the product specification.
6. Return to step 4 and change NR and NS until the TAC is minimized.
7. Return to step 3 and find the feed locations (NFheavy and NFlight) until the TAC is minimized.
8. Return to step 2 and vary Nnx until the TAC is minimized. 17.3.1 Type I: One Group
Type Ip LK + HK = LLK + HHK. This is the most popular reactive distillation configuration, which has been studied by a number of authors.3,4 In real chemical systems, the production of diphenyl carbonate belongs to this class.6 The net reaction can be expressed as dimethyl carbonate + 2 phenol o 2 methanol + diphenyl carbonate LK HK LLK HHK
Note that many real chemical systems have azeotropes associated with quaternary systems, and it may make the neat reactive distillation design infeasible. However, for the real chemical systems illustrated here, the placement of the reactive zone is the same as that of the ideal systems, despite having azeotropes. The boiling point ranking leads to an easy separation between the reactants and products. The two products leave the reactive section from opposite sides of the reactive zone while the two intermediate boilers (the reactants) are kept in the reactive zone. This is the most favorable boiling point arrangement possible for a quaternary reactive distillation system.
Figure 17.4a shows the effects of separation trays and reactive trays on the TAC. Because of the symmetry in the relative volatilities between the two products and the two reactants in the separation sections of the column, the numbers of trays in the rectifying section and the stripping section are assumed to be the same, that is, NR = NS. Figure 17.4a indicates that the optimum number of reactive trays Nnx is 16 and the optimum number of separation trays NR = NS is 4. The tradeoff comes from the fact that in the vicinity of the optimum the operating cost goes down and the capital cost goes up as we increase the number of separation trays. Figure 17.4b shows that the TAC is quite sensitive to changes in the feed tray locations, which is consistent with the findings in Chapters 7 and 18. When the heavy reactant feed and the light reactant feed are five trays apart, the design gives the lowest TAC, and the optimum corresponds to NFheavy = 16 and NFlight = 11. This results in a TAC of $254,170. Detailed process parameters are provided in Table 17.3.
6S. Fukuoka, T. Watanabe, and T. Dozono, Methods for producing a crystallized aromatic carbonate and a crystallized aromatic polycarbonate obtained thereby, US Patent 4,948,871, 1990.
Nr(Ns) N Flight
Figure 17.4 Relationship between TAC and design variables for type Ip process: (a) number of separation trays (NR and NS) versus TAC for different NRX and (b) feed tray locations (NFheavy and NFlight) with Nrx = 22 and NR (NS) = 4.
Figure 17.5 demonstrates that, for type Ip reactive distillation with products as LLK and HHK, the reactive zone is located at the middle of the column and the light product (LLK) and the heavy product (HHK) are removed from the top and bottoms of the column, respectively. This is intuitively correct because we have higher reactant (intermediate boilers) concentrations in the midsection of the column. As will become clear later, one of the most important elements in the conceptual design of reactive distillation is placing the reactive zone at the right location. The composition profile of the final design is given in Figure 17.6 and, as expected, significant amounts of reactants are present in the reactive zone (between two dashed lines) and product compositions increase gradually toward the top and bottoms of the column. Figure 17.6 also shows the fraction of total conversion (Rj/R,) in each reactive tray. These results indicate that all reactive trays are well utilized in the reactive zone.
Before leaving this section, we explore the effects of reactive holdups on the designs (assumption 3 in design). The weir height is varied from 5 to 40 cm (likely the lower and upper limits for real columns). Figure 17.7 shows that the TAC changes —10% (5 cm) to +10% (40 cm) from the base case design. Instead of using two extreme values, in this work a modest weir height of 10 cm is used in all designs.
Another issue is that the effect of pressure on the design is not explored here because we are assuming constant relative volatility systems. The column pressure is fixed at 8 bar in this work. Pressure is very important in reactive distillation because of the effect of temperature on both vapor-liquid equilibrium and reaction kinetics. For exothermic reactions, the optimum column pressure is affected by the competing effects of temperature on the specific reaction rates and the chemical equilibrium constant.
Type IR: LLK 1 HHK = LK 1 HK. This corresponds to the following reaction:
Was this article helpful?