In summary, the tradeoff comes from the competition between the reaction in the column base and the side reactor. If we increase the reaction in the column base, it will reduce the loading of the side reactor and consequently the side reactor will require a smaller side stream flowrate and a smaller percentage of equilibrium conversion, as shown in Figure 16.35.
For the case of one side reactor, the TAC only increases by a factor of 5% compared to that of the reactive distillation design, and we conclude that an additional external reactor will not be necessary for the side reactor configuration. Specifically, the energy cost of the side reactor configuration increases by a factor of 3% while using 1.67 times the amount of the catalyst, as compared to the reactive distillation process. Figure 16.36 shows the optimized flowsheet where the side stream flowrate is 3.69 times the organic reflux flowrate and the catalyst loading in the column base and side reactor are Wcat bot = Wcat,RD and Wcat,SR = 0.67 x Wcat,Ro, respectively. This corresponds to 95% of the equilibrium conversion for the side reactor. Moreover, 89% of the total conversion occurs in the column base while the remaining 11% happens in the side reactor. Composition profiles for the side reactor configuration process (Fig. 16.37) show that the composition profile of the side reactor configuration is almost the same as that of the reactive distillation (see Fig. 16.24). The parameter values of the steady-state designs of all three configurations (reactive distillation, single reactive tray, and side reactor configuration) are summarized in Table 16.6.
16.5 CONTROL OF COLUMN/SIDE REACTOR PROCESS FOR ETHYL ACETATE SYSTEM
The number of studies on the control of reactive distillation columns has grown steadily in the past decade. Luyben and colleagues propose eight control structures for the neat reactive distillation of the ideal system (A + B , C + D): CS1-CS6,8 CS6-CS7,9 and CS7-CS8.1,3 Chapter 13 explores control structure design for acetic acid esterification with different alcohols (ranging from C1 to C5), which are categorized into three types of flowsheets. As for the control of various coupled reactor/column systems, several articles have been published, such as Yi and Luyben7 and Chiang et al.10 However, the control of the side reactor configuration has not been discussed in the literature. In this section, a systematic approach to design the control structure is suggested. The design steps are the following:
1. Determine manipulated variables.
2. Use NRG (nonsquare RGA) to determine temperature control trays.
8M. A. Al-Arfaj and W. L. Luyben, Comparison of alternative control structures for an ideal two-product reactive distillation column, Ind. Eng. Chem. Res. 39, 3298-3307 (2000).
9M. A. Al-Arfaj and W. L. Luyben, Comparative control study of ideal and methyl acetate reactive distillation, Chem. Eng. Sci. 57, 5039-5050 (2002).
10S. F. Chiang, C. L. Kuo, C. C. Yu, and D. S. H. Wong, Design alternatives for amyl acetate process: Coupled reactor/column and reactive distillation, Ind. Eng. Chem. Res. 41, 3233-3246 (2002).
3. Use a decentralized PI controller.
4. Use RGA for variable pairing.
5. Use relay feedback to find Ku and Pu.
6. Use Tyreus -Luyben tuning to find controller settings.
Compared to reactive distillation, the side reactor configuration has 1 extra manipulated variable, side stream flowrate FSS, in addition to two feedflows (FEtOH and Fhac), two reboiler duties (QR,RD and QR,STR), two condenser duties, one subcooling duty, aqueous and organic distillate flowrates, organic reflux flowrate Rorg, and stripper bottoms flowrate. similar to the control of the reactive distillation system, the alcohol feedflow is selected as the throughput manipulator. We are left with 11 manipulated variables.
In the EtAc side reactor configuration process, there are six inventory control loops required, which include the control of four liquid levels (the reactive distillation column base level, the organic phase level in the decanter, the aqueous phase holdup in the decanter, and the stripper base level) and the two column pressures of the reactive distillation and stripping columns. Basic inventory and related loops are arranged in the following ways. The top pressures of two columns are controlled by manipulating the condenser duties. The decanter temperature is controlled by changing the chilled water flow. As for liquid inventories, except for the reactive distillation column base, the remaining liquid levels are controlled by their liquid outlet flow as shown in Figure 16.38.
For the reactive distillation column base-level control, we have three candidates: side draw flow FSS, organic reflux flow Rorg, and reboiler duty QR,RD. Because the side draw flow has 0 gain effect on the base level and the organic reflux flow gives slow dynamics,11 we choose the heat input of the reactive distillation column QR,RD to control the base level. In doing this, we have four manipulated variables left: FSS, FHAc, R, and QR,STR.
After selecting the inventory controls, the remaining manipulated variables (HAc feed, organic reflux flow, and stripper heat duty) are used for maintaining the stoichiometric balance and for product quality control (purities of the ethyl acetate and water products). In the stripper, the acetate product purity is controlled by the stripper heat input QR,STR, and a temperature is used to infer the product composition. From the RCM (Fig. 16.22), the water purity is determined by the tie line of the LL equilibrium, so it is left uncontrolled. In order to maintain the stoichiometric balance, we use a temperature control to adjust the feed ratio FHAc/FEtOH (FR). Thus, we have a 2 x 2 multivariable system. From a unitwise perspective, the pairing for the manipulated variables and the controlled variables are FR-TRD and QR STR-Tstr.
We have two manipulated variables left, side draw flow FSS and organic reflux flowrate Rorg. For the side stream flowrate, one can choose to ratio the side stream flowrate to the feed or simply fix FSS to its setpoint. The latter is utilized here because it is a relatively large flowrate (FSS = 1500 kmol/h) compared to the reflux flowrate (R = 407 kmol/h) and frequent perturbations may introduce unnecessary disturbances to the lower section of the column. Two possible approaches can be taken for the organic reflux. One is to fix the reflux ratio RR. The other is to ratio the reflux to the feed (R/FEtOH). Therefore, we
11I. L. Chien, Y. P. Teng, H. P. Huang, and Y. T. Tang, Design and control of an ethyl acetate process: Coupled reactor/column configuration, J. Process Control 15, 435-449 (2005).
Figure 16.38 (Continued).
have two possible control structures: CS1 for keeping RR constant and CS2 for fixing the reflux to feed ratio as shown in Figure 16.38.
Sensitivity analyses are performed for +0.01% variations in the corresponding manipulated variables (FR and QR,STR). Results are shown in Figure 16.39 for CS1 and Figure 16.40 for CS2. In both control structures, an increase in QR,STR leads to a decrease in the tray temperatures of the reactive distillation column because of a larger recycle flowrate for the reactive distillation column. As for the feed ratio (FR) changes, an increase in the heavy reactant results in the main column (reactive distillation column) tray temperatures increasing while the stripper temperatures show little variation. At the completion of the sensitivity analyses, NRG12 is used to find the temperature control trays in each column. The largest row sum of the NRG in each column is selected as the temperature control trays (Fig. 16.41). Note that the temperature control trays selected by the NRG are consistent with the most sensitive trays in Figures 16.39 and 16.40. The analyses select temperature control trays 7rD,15 and TStR4 for CS1 and rRD,15 and TStR3
Was this article helpful?