Control of BenzeneâCyclohexane Separation System via Extractive

Article pubs.acs.org/IECR

Control of Benzene−Cyclohexane Separation System via Extractive Distillation Using Sulfolane as Entrainer Jiwei Qin,† Qing Ye,*,† Xiaojuan Xiong,† and Ning Li‡ †

Institute of Petrochemical Technology, Changzhou University, Changzhou 213100, People’s Republic of China Institute of Petrochemical Technology, China University of Petroleum, Beijing 102200, People’s Republic of China

‡

ABSTRACT: Benzene and cyclohexane (CYH) can be separated via extractive distillation using sulfolane (SULF) as entrainer. In this Article, the steady state of an extractive distillation system has been simulated using the RadFrac model in Aspen plus, and the controllability of this system with feed flow rate and feed composition disturbances has been studied in Aspen Dynamics. Sensitivity analysis is used to obtain optimal operation conditions. The common control scheme for extractive distillation is infeasible. Two control structures are demonstrated in this Article. Each control structure has three main temperature controllers, two in the extractive distillation column and one in entrainer recovery column. Although both control structures can solve all disturbance issues and maintain the product purities very close to the set points, the dynamic responses of initial control structure have large transient deviation when +20% feed flow rate disturbance is introduced, while the improved control structure can overcome this obstacle. ethyl acetoacetate.6−8 Most of the current research on the benzene−cyclohexane separation system focuses on the steadystate design and entrainers, while only a small amount of research is on the dynamics. Xu et al.9 experimentalized the batch extractive distillation using the solvents of N,N-dimethylformamide and dimethyl sulfoxid. Vega et al.10 studied the solvent selection using non-steady-state gas chromatography. Also, Kao et al.,11 Yildirim et al.,12 and Sarkhel et al.13 studied the pervaporation of the benzene and cyclohexane system by membranes. The design and control of extractive distillation are reported in many papers. Most of them are for a special separation system,14−19 to optimize the steady state and put forward control structures. Vazquez et al.18 used a stochastic global optimization algorithm to obtain optimal process designs, Wang et al.14 adopted total annual cost to achieve optimal design, and Errico et al.19 analyzed the different possible process sequences to conclude the most energy-saving configuration. Some other papers proposed complex configures of extractive distillation.20−25 MurrietaDueñas et al.,20 Gutiérrez-Guerra et al.,21 and Ibarra-Sánchez et al.22 studied the optimization and control of thermally coupled extractive distillation. Bravo−Bravo et al.,23 Xia et al.24 and Errico et al.25 did research on the extractive dividing wall column. In addition, most of the papers discussed the control property of the processes. Distillation control, which is another important aspect for extractive distillation, can be mainly categorized into two types: composition control and temperature control. The composition control is the more accurate, but it brings in long drag time and costs more in both equipment and maintenance. Thus, the temperature control is more widely used in extractive distillation for different azeotropic systems and entrainers. Wang et al.14

1. INTRODUCTION Aromatic hydrocarbons are important raw materials in the petrochemical industry. Benzene, toluene, and xylene are widely used not only in the production of synthetic rubber, synthetic resin, and synthetic fiber, but also as raw materials of synthetic detergent, plasticizer, explosive, dyes, etc. There are many methods to produce benzene. The production of benzene from catalytic reforming and pyrolysis gasoline accounts for 38% of the total production worldwide.1 Because the aromatic hydrocarbons easily form azeotropes with nonaromatics, such as benzene and cyclohexane, the aromatic first must be separated from the mixture of hydrocarbons. Although distillation is popular for the separation of the homogeneous liquid mixture, the azeotropic system cannot be separated by distillation unless using some special methods, such as azeotropic distillation, extractive distillation, pressure swing distillation, etc. The pressure swing distillation is not feasible when the azeotropic composition is insensitive to pressure changes. Of the other two alternatives, the heterogeneous azeotropic distillation is preferable to the homogeneous azeotropic distillation because of the easier separation of the entrainer through liquid−liquid splitting in a decanter. However, the extractive distillation can be more energy-efficient than the heterogeneous azeotropic distillation, especially when thermally integrated sequences are used.2−5 The most common processes industrially used to recover aromatics are the liquid−liquid extraction and the extractive distillation; the former is more complicated and requires higher capital investment and operation cost than the latter. As for the entrainers of the extractive distillation processes, the MORPHYLANE process, the DISTAPEX process, and the SED process use N-formylmorpholine, N-methylpyrrolidone, and sulfolane, respectively, while some other applications use mixtures, such as a mixture of benzoic acid, maleic anhydride, and/or phthalic anhydride, a mixture of dimethyl sulfoxide and optional cosolvent like water, or a mixture of methyl acetoacetate and © XXXX American Chemical Society

Received: April 7, 2013 Revised: July 8, 2013 Accepted: July 11, 2013

A

dx.doi.org/10.1021/ie401101c | Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX

Industrial & Engineering Chemistry Research

Article

Figure 1. The process flow sheet of extractive distillation.

separated methylal and methanol using N,N-dimethylformamide as entrainer, and they used simple temperature control structure to control one temperature at each bottom of the two columns and obtained good dynamic responses with the ratio of reflux flow rate/feed flow rate (R/F) schemes. Arifin et al.15 separated isopropanol and water using dimethyl sulfoxide as entrainer, and they also used the two temperature control structure with R/F ratios. Luyben16 studied the acetone and methanol separation system by extractive distillation using water as solvent, and he used the two temperature control structure as well with the fixed reflux ratios. Gil et al.17 separated the ethanol and water using glycerol as entrainer, and the control structure was similar to that of Luyben16 except for the bottom level control of entrainer recovery column. In this Article, the extractive distillation process is based on an actual industrial SED process. When this process was put into industrial practice, the feed condition fluctuated and the process became difficult to control. In this work, the benzene and cyclohexane separation system via extractive distillation using sulfolane as entrainer has been studied. The feed condition and other operation conditions are all very close to the industrial practice. The nonrandom two liquid (NRTL) model, which can predict the system very well, is used to calculate the physical properties of system.26 The sensitive analysis for multiple variables has been done to obtain optimal operation conditions. In industrial practice and some papers,14−17 the extractive distillation column (EDC) and the entrainer recovery column (ERC) are, respectively, controlled by one-stage temperature, which manipulates the reboiler duty. The same control scheme has been tried in this study, and the results revealed that this control scheme could not solve the feed flow rate and feed composition disturbances. After many attempts, it finds that using two temperatures to control the extractive distillation column is available in this case. Two control structures are proposed, and the dynamic responses for feed flow rate and composition disturbances are evaluated in the two control structures.

from the lower part and the entrainer feed (EF) of pure sulfolane from the upper stage are fed into the extractive distillation column. The presence of sulfolane alters the relative volatility between benzene and cyclohexane to make cyclohexane move toward the top part and benzene move to the bottom part of the column. Consequently, the cyclohexane is produced from the top of the EDC, while the benzene and the entrainer are collected from the bottom of the EDC. The bottom production of EDC is fed into the entrainer recovery column to produce almost pure benzene in the distillate and almost pure sulfolane in the column bottom. Sulfolane as a heavy entrainer will be recycled back to the EDC. A small fresh sulfolane stream is fed into EDC to make up the minor losses. 2.2. Sensitivity Analysis. To establish the operating conditions for the extractive distillation process, a sensitivity analysis is carried out. The analyzed parameters are as follows: the pressure of reflux drums of ERC, the reflux ratios (RR), the molar ratio of entrainer to feed (EF/F), the mixture feed stage (FS), the entrainer feed stage (EFS), and the entrainer temperature (ET). The original case of the process is as below. Because the process is based on an actual industrial process, the operation conditions (feed flow rate, feed composition, total stage number, and so on) should get very close to the real process. The feed flow rate is assumed to be 100 kmol/h, and the composition is equal mole fraction of benzene and cyclohexane. The two top product purities are both 0.999 (kmol/kmol). The number of total stages for EDC is 29, while the ERC is 10 (the aspen notation is that the condenser is stage 1 and the reboiler is the last stage). The initial entrainer feed stage is 3 of the EDC, and the fresh feed stage is 19. The ERC feed (F2) stage is 5. The reflux drum of EDC is operated at 1 atm, and the reflux drum of ERC is 0.08 atm. Effect of the Pressure of ERC. The boiling points are quite different for benzene (80.1 °C) and sulfolane (285 °C); therefore, the reboiler duty will be significantly increased if the ERC is operated at atmospheric pressure instead of 0.08 atm. Table 1 gives the comparison between the two operating conditions. The reboiler duty of ERC (QR2) at atmospheric condition is 62% higher than that at 0.08 atm. What is more important is that the base temperature of the ERC is only 205 °C on the vacuum operation so that the thermal decomposition of SULF can be

2. STEADY-STATE DESIGN 2.1. Flow Sheet. The process for the benzene and cyclohexane separation using sulfolane as the entrainer is shown in Figure 1. The fresh feed (F) of benzene−cyclohexane mixture B



Article

It is noticed that the energy requirement increases when the reflux ratio is larger for a fixed EF/F ratio in Figure 2. The same pattern occurs when the EF/F ratio increases for a fixed RR1 value, and RR1 affects more the energy requirement. Although the entrainer has selectivity, the entrainer flow of the EDC is subcooled, and it condenses some part of the vapor flowing up the EDC column. The condensed cyclohexane must be vaporized again. The higher is the EF/F ratio, the more vapor flowing up is needed. Finally, the high energy requirement corresponds to the high EF/F ratio for fixed reflux ratio. The reflux ratio has a similar effect on energy requirement for fixed EF/F ratio. The energy requirement cannot decrease infinitely by lowering the RR1 and the EF/F ratio because the product purity is fixed. Figure 3 gives the effect of the EF/F ratio and the RR1 on the product purity of EDC to find the minimum EF/F ratio. For a given EF/F ratio, the mole fraction of CYH will reach a maximum value with a certain reflux ratio. Given a fixed reflux ratio value, the mole fraction of CYH will increase with the rising EF/F ratio. The minimum EF/F ratio is between 0.9 and 1, and the range of 1.0−1.5 is selected to ensure the controllability of the system. Effects of the Molar Ratio of Entrainer to Feed and the Entrainer Temperature. On the basis of the assumption in some papers17,27,28 that the temperature of the entrainer is suggested to be 5−15 °C less than the top temperature of the EDC, the entrainer temperature is studied between 65 and 75 °C. Figure 4

Table 1. Results of Two Operating Conditions of ERC pressure (atm) feed stage condenser duty (GJ/h) reboiler duty (GJ/h) base temp (°C)

condition 1

condition 2

1 5 −2.3112 6.3972 289

0.08 5 −2.5729 3.9276 205

avoided. Unfortunately, the top temperature of ERC is only 15.7 °C, which is too low to use the conventional cooling water as the condensing intermedium. Effects of the Molar Ratio of Entrainer to Feed and Reflux Ratio of EDC. Figures 2 and 3 show the influences of the molar

Figure 2. Influence of EF/F ratio and RR1 on reboiler energy requirement of EDC.

Figure 4. EDC reboiler duty with different EF/F ratios and different EF temperature.

shows the influence of the molar ratio of entrainer to feed and the entrainer temperature on the reboiler duty of EDC while maintaining the production purities. If the temperature of the entrainer is fixed, there is a low energy requirement region when the EF/F ratio is between 1.1 and 1.2. The reboiler duty increases much less as the temperature rises for a fixed EF/F ratio when compared to the former effect. With these results, the optimal EF/F ratio is 1.1, and the entrainer temperature is 65 °C. Effect of the Entrainer Feed Stage and the Mixture Feed Stage. The effect of the entrainer feed stage and the mixture feed stage on reboiler duty of EDC is shown in Figure 5 while maintaining the production purities. It is noticed that the EDC reboiler duty is almost a constant when the F stage is in the range of 18−22; however, it will increase sharply when the feed stage is more than 22. For a fixed mixture F stage, as the number of EF

Figure 3. Influence of EF/F ratio and RR1 on product purity of EDC.

ratio of EF/F and reflux ratio of EDC (RR1) on the reboiler duty of EDC (QR1) and the product purity of EDC (XD1), respectively. For a given separation system and the entrainer, the EF/F ratio must exceed a limited value so that it can achieve the separation objective. However, the higher is the EF/F ratio, the more energy requirement it will be. The minimum EF/F ratio to reach the product purity requirement must be determined, and then the EF/F ratio is selected by contrasting the energy requirements and controllability. C



Article

stage is 4 and the fresh feed stage of the EDC is 20. The ERC feed stage is 5. It is noted that the EF includes the entrainer recycle and the fresh sulfolane to make up for the small losses of the entrainer. The reflux ratios of two columns are 1.218 (RR1) and 0.5 (RR2), respectively. The ratio of the EF/F is 1.1, and the temperature of EF is 65 °C.

3. CONTROL SYSTEM DESIGN 3.1. Selecting Temperature Control Trays. According to some papers and books, the main criteria for selecting temperature control trays include five aspects: (1) slope criterion, choose a tray that has steep change in temperature from tray to tray; (2) sensitivity criterion, select the tray that has the largest change in temperature for a change in the manipulated variable; (3) SVD criterion, use singular value decomposition to analysis; (4) invariant temperature criterion, select the tray that the temperature does not change as feed composition changes while producing the desired distillate and bottoms purities; and (5) minimum product variability criterion, choose the tray that produces the smallest changes in product purities when it is held constantly in the face of composition disturbances. In this Article, only the slope criterion and SVD criterion are used.4,16,29 Figure 8 shows the temperature profiles of the two columns. The upper plot of Figure 8 shows the temperature profile in EDC, stage 14 has the steepest temperature slope, and stage 23 also has a steep temperature slope. The second plot of Figure 8 shows that the steep temperature slopes appear at stages 5, 7, and 8 in ERC. The SVD criterion is used to select the two control temperatures of EDC. First, it must make a sensitivity analysis of the manipulated variables (RR1 and QR1) in EDC. Adjust one of the variables to a very small change and maintain the other variable invariant. Next, rerun the simulation in Aspen to obtain a new temperature profile. The difference between the temperatures in the same tray is divided by the change of the manipulated variable. So it gets a matrix with 29 rows and 2 columns (the matrix is assumed as X). Finally, the matrix is decomposed by the SVD program in Matlab into three matrices: K = UΣVT. Here, the program in Matlab is [U, S, V] = svd(X, 0), which can generate three matrixes. The vectors of X and U are plotted against the stage number, respectively, and the results are shown in Figure 9. There are peaks at stages 11, 12, 14, and 23 in the upper plot of Figure 9. The lower plot of Figure 9 shows the two U vectors, which reveal peaks at 3, 11, 12, 14, and 23. From these results and the manipulated variables, the control system selects control stage 14 temperature with reflux ratio of EDC and stage 23 temperature with reboiler heat input of EDC. Because the feed of ERC has only slight CYH, it can be treated as a simple binary system. There are three temperature candidates in ERC from Figure 8. The control system selects the stage 7 temperature with reboiler heat input of ERC. 3.2. Two Control Structures. In this section, the Aspen plus file will be imported into Aspen Dynamics. Before the conversion, the sizes must be specified for the major equipments, such as the reflux drums, column bases, etc. The Tray-Sizing tool of Aspen plus can calculate the diameters of columns conveniently. Usually, the reflux drums and column base sizes provide 10 min of the liquid holdup. This is a pressure-driven dynamic simulation. All of the pumps require the necessary pressure increase so that the control valves can drop enough pressure and have adequate dimensions to pass streamflow when they are half open. Each control valve has 3−10 atm pressure

Figure 5. EDC reboiler duty at different EF and F locations.

stage changes, there is little effect on the reboiler duty of the EDC. Eventually, the optimal EF stage range is 3−6 and the optimal F stage range is 19−21 for the optimal low energy requirement area. It is worth noting that the distance between the EF stage and the F stage must be a suitable value so that the operation can accomplish the separation target. In addition, the rectifying section of EDC must maintain enough space to avoid entrainer loss at the top distillate; the stripping section of EDC also needs enough space to avoid the CYH loss at the bottom. Effects of the Reflux Ratio and the Mixture Feed Stage of ERC. Figure 6 shows the results of the ERC reboiler duty affected

Figure 6. The reboiler duty of ERC with different RR2 and feed stage.

by the reflux ratio (RR2) and the feed stage (F2) location while maintaining the production purities. It shows that the reboiler duty increases as the RR2 is added when the feed stage location is fixed. With a fixed RR2 value, the location of the feed stage almost does not affect the reboiler duty except at the two ends of the column. What is more is that the reflux ratio has an effect on the controllability of ERC. The ERC is more controllable, when the reflux ratio of ERC increases suitably. There is an offset between the controllability and energy requirement. Therefore, the optimal feed stage location is at 3−7 and RR2 is 0.5, which is in the low reboiler duty area and advisable controllability. 2.3. The Optimal Flow Sheet. The final process flow sheet of steady-state design is shown as Figure 7. The entrainer feed D



Article

Figure 7. The steady-state process flow sheet of extractive distillation.

Figure 8. Temperature profiles of the two columns.

Figure 9. SVD analysis for EDC.

drop when half open. Pressure-check is done for the whole steady-state flow sheet, and then the Aspen profile is exported to Aspen Dynamics. Some similar papers reported the control scheme of extractive distillation.14−17 The extractive distillation column and the entrainer recovery column are, respectively, controlled by onestage temperature, which manipulates the reboiler duty. On the basis of these papers, this work first used the same control

scheme, but it was ineffective. After many attempts, it found that two control structures as follows were available. 3.2.1. Initial Control Structure for Extractive Distillation System. The initial control structure and the control panel are shown in Figure 10. The basic control schemes are as below. (1) The levels of the two reflux drums are both maintained by manipulating the flow of top productions (direct acting). E



Article

Figure 10. The initial control structure.

Table 2. Temperature Control Tuning Parameters for Two Control Structures Initial Control Structure parameters controlled variable manipulated variable controller input range (°C) controller output range (GJ/h) ultimate gain ultimate period (min) gain, KC integral time, τI (min) parameters controlled variable manipulated variable controller input range (°C) controller output range (GJ/h)/(kmol/h) ultimate gain ultimate period (min) gain, KC integral time, τI (min)

TC11 T1,23 QR1 90−100 0−2.191 9.046 5.94 2.827 13.068

TC12 T1,14 RR1 70−90 0−2.436 (kg/h)/(kg/h) 0.390 5.1 0.122 11.22 Improved Control Structure TC11 T1,23 QR1/F 90−100 0−0.022 8.163 6.60 2.551 14.52

TC12 T1,14 R1/F 75−90 0−1.263 (kg/h)/(kg/h) 3.004 6.0 0.939 13.2

TC21 T2,7 QR2 40−200 0−7.856 3.238 5.7 1.012 12.54

TC3 TRecycle QHX 40−90 0−6.311 0.510 2.1 0.159 4.62

TC21 T2,7 QR2/F 50−190 0−0.079 0.399 5.4 0.125 11.88

TC3 TRecycle QHX/F 60−70 0−0.063 0.506 1.8 0.158 3.96

(4) Feed flow rate is controlled to achieve constant flow (indirect acting). (5) EF flow rate is maintained by manipulating the flow of entrainer recycle (indirect acting). To solve the feed disturbances, the EF flow rate is cascaded with F flow rate by ratio control.

(2) The column base level of the EDC is maintained by manipulating the flow of bottom product (direct acting). However, the column base level of the ERC is maintained by manipulating the flow of entrainer makeup (indirect acting). (3) Both of the pressures in two columns are maintained by manipulating the top condenser duties (indirect acting). F



Article

Figure 11. Dynamic responses for initial control structure with 20% feed flow rate disturbances.

The temperature of entrainer recycle is maintained by manipulating the cooler duty (indirect acting). (9) Each temperature control loop adds a 1 min dead time to fit the practical operation.

(6) The mass reflux ratio of ERC is constant. (7) The temperature of 23 stage of EDC is maintained by manipulating the reboiler duty of EDC (indirect acting). The temperature of 14 stage of EDC is maintained by manipulating the reflux ratio in EDC (direct acting). (8) The temperature of 7 stage of ERC is maintained by manipulating the reboiler duty of ERC (indirect acting).

Common PI controllers are used for all controllers. The four level controllers are all with KC = 2. The proportional and integral (PI) settings of the top pressure control loops for the two G



Article

Figure 12. Dynamic responses for initial control structure with feed composition disturbances.

columns are set at KC = 20 and τI = 12 min. The proportional and integral (PI) settings of two feed flow controllers are set at KC = 0.5 and τI = 0.3 min. Relay-feedback tests are run to obtain the ultimate gains and periods, and Tyreus−Luben tuning is used for each temperature controller. Table 2 gives the tuning parameters for the initial control structure temperature controllers. Now the dynamic performance of the initial control structure is evaluated by the feed flow rate and composition disturbances.

Figure 11 shows the results of initial control structure when going through the fresh feed flow rate disturbances. The results show the dynamic responses of this control structure corresponding to the 20% step changes in the feed flow rate at 0.5 h. The system can reach a new steady state in 2 h after introducing the disturbances, and the product purities come very close to the set points. All of the controlled temperatures arrive at the set points as well. However, the product purities have large transient deviations from the set points when the feed flow rate increases H



Article

Figure 13. The improved control structure with feed forward control.

20%, and the CYH mole fraction nearly closes to 0.98. The reason is that the entrainer flow rate increases very fast when the disturbance is introduced, and then the bottom product of ERC increases sharply. However, the cooler of the entrainer is unable to cool the recycled entrainer to the set point immediately. Thus, in the last plot of Figure 11, the transient entrainer temperature is 15 °C higher than the set point. The EF stage temperature may allow part of the vapor benzene to escape from the top of EDC. Figure 12 shows the results of initial control structure when going through the feed composition disturbances. In the real industrial process, the composition of benzene is within the range of 0.45−0.54. The composition of fresh feed will, respectively, change to 55 mol % benzene and 45 mol % benzene at 0.5 h. The results (or responses) are better than the feed disturbances, although the transient deviation is still large when 55 mol % benzene disturbance is introduced. The product flow rates reach the new steady state within 2 h, and the controlled temperatures restore to the set points with small transient deviations. The entrainer temperature largest transient deviation is within 0.5 °C. This is much better than the response of the feed flow rate change. 3.2.2. An Improved Control Structure for Extractive Distillation System. To obtain better dynamic responses, the initial control structure must be adjusted for some improvements. The improved control structure is demonstrated in

Figure 13. In Figure 11, the entrainer temperature has large transient deviations when undergoing fresh feed flow rate disturbances. This problem can be solved by adding feed forward control structure; the direct control duty of entrainer cooler is replaced by controlling the ratio of cooler duty to mole feed flow rate (QHX/F). The improved control structure adds two R/F ratio schemes to replace the fixed reflux ratios. In addition, two feed forward controls of the temperature are added. The controlled temperatures at the bottom parts of the columns are maintained by manipulating the ratios of reboiler duty to the mole feed flow rate (QR1/F and QR2/F). The tuning parameters of the temperature controllers for improved control structure are shown in Table 2. Next, the improved control structure is evaluated by feed flow rate and feed composition disturbances. The dynamic responses for 20% feed flow rate disturbances are demonstrated in Figure 14, while dynamic responses for composition disturbances with benzene 55 mol % and benzene 45 mol %, respectively, are shown in Figure 15. The disturbances are introduced at 0.5 h as well. In Figure 14, the improved control structure gives tighter control with smaller peak transient deviations and shorter settling time. The system reaches a new steady state within 3 h, and the product purities are kept fairly close to the set points. The product purities that have small transient deviations from the set points are higher than 0.9985 when the feed flow rate increases I



Article

Figure 14. Dynamic responses for improved control structure with 20% feed flow rate disturbances.

20%. It is worth noting that the fluctuations of controlled temperatures are much smaller than initial control structure performances with the same disturbances. The largest transient deviations of controlled temperatures are not higher than 3 °C in EDC, and the system can rapidly achieve the steady state. For the temperature of entrainer recycle, the transient deviations are smaller than 4 °C, which is much better than the responses of initial control structure with the same disturbances.

In the Figure 15, the system reaches the new steady state within 2 h, while the initial control structure performs at about 2.5 h. What is more important is that the fluctuation of each variable in the Figure 15 is smaller than the same performances in Figure 12, which means this control structure is more stable. The lowest transient product purity of CYH is higher than 0.9977, and the lowest transient product purity of benzene is about 0.9982 when the disturbance of benzene 55 mol % is introduced. These are better than the responses J



Article

Figure 15. Dynamic responses for improved control structure with feed composition disturbances.

distillation using sulfolane as entrainer are studied in this Article. From the sensitivity analysis, the main operating conditions of the system are established, and the effects of the design variables are investigated. The ratio of EF/F has great effects on the energy requirement and purity of cyclohexane; however, it should be operated at a suitable value, which is a little higher than the minimum value. Reflux ratio is useful in compensation changes in some operating

of initial control structure with the same disturbances. The largest transient deviation of the controlled temperature of ERC is about 20 °C with composition disturbances, while it is over 25 °C for the initial control structure with the same disturbances.

4. CONCLUSIONS On the basis of a real industrial process, design and control of the benzene and cyclohexane separation process via extractive K



Article

(5) Widagdo, S.; Seider, W. D. Azeotropic distillation. AIChE J. 1996, 42, 96−130. (6) Berg, L. Separation of benzene from nonaromatic hydrocarbons by extractive distillation. U.S. Patent 4,514,262, April 30, 1985. (7) Berg, L. Separation of benzene from close boiling hydrocarbons by extractive distillation. U.S. Patent 5,458,741, October 17, 1995. (8) Gentry, J. C.; Berg, L.; Mclntyre, J. C.; Wytcherley, R. W. Process to recover benzene from mixed hydrocarbons by extractive distillation. U.S. Patent 5,399,244, March 21, 1995. (9) Xu, J.; Zhang, W. J.; Gui, X. Separation of benzene and cyclohexane by batch extractive distillation. Trans. Tianjin Univ. 2007, 13, 433−436. (10) Vega, A.; Díez, F.; Esteban, R.; Coca, J. Solvent selection for cyclohexane−cyclohexene−benzene separation by extactive distillation using non-steady-state gas chromatography. Ind. Eng. Chem. Res. 1997, 36, 803−807. (11) Kao, S. T.; Wang, F. J.; Lue, S. J. Sorption diffusion and pervaporation of benzene/cyclohexane mixtures on silver−Nafion membranes. Desalination 2002, 149, 35−40. (12) Yildirim, A. E.; Hilmioglu, N. D.; Tulbentci, S. Separation of benzene/cyclohexane mixtures by pervaporation using PEBA membranes. Desalination 2008, 219, 14−25. (13) Sarkhel, D.; Roy, D.; Bandyopadhyay, M.; Bhattacharya, P. Studies on separation characteristics and pseudo-equilibrium relationship in pervaporation of benzene−cyclohexane mixtures through composite PVA membranes on PAN supports. Sep. Purif. Technol. 2003, 30, 89−96. (14) Wang, Q. Y.; Yu, B. R.; Xu, C. J. Design and control of distillation system for methylal/methanol separation. Part 1: extractive distillation using DMF as an entrainer. Ind. Eng. Chem. Res. 2012, 51, 1281−1292. (15) Arifin, S.; Chien, I. L. Design and control of an isopropyl alcohol dehydration process via extractive distillation using dimethyl sulfoxide as an entrainer. Ind. Eng. Chem. Res. 2008, 47, 790−803. (16) Luyben, W. L. Comparison of extractive distillation and pressure−swing distillation for acetone−mehanol separation. Ind. Eng. Chem. Res. 2008, 47, 2696−2707. (17) Gila, I. D.; Gómez, J. M.; Rodríguez, G. Control of an extractive distillation process to dehydrate ethanol using glycerol as entrainer. Comput. Chem. Eng. 2012, 39, 129−142. (18) Vázquez-Ojeda, M.; Segovia-Hernández, J. G.; Hernández, S.; et al. Design and optimization of an ethanol dehydration process using stochastic methods. Sep. Purif. Technol. 2013, 105, 90−97. (19) Errico, M.; Rong, B. G.; Tola, G.; et al. Optimal synthesis of distillation systems for bioethanol separation. Part 1: extractive distillation with simple columns. Ind. Eng. Chem. Res. 2013, 52, 1612− 1619. (20) Murrieta-Dueñas, R.; Gutiérrez-Guerra, R.; Segovia-Hernández, J. G.; et al. Analysis of control properties of intensified distillation sequences: Reactive and extractive cases. Chem. Eng. Res. Des. 2011, 89, 2215−2227. (21) Gutiérrez-Guerra, R.; Segovia-Hernández, J. G.; Hernández, S. Reducing energy consumption and CO2 emissions in extractive distillation. Chem. Eng. Res. Des. 2009, 87, 145−152. (22) Ibarra-Sánchez, J. J.; Segovia-Hernández, J. G. Reducing energy consumption and CO2 emissions in extractive distillation: Part II. Dynamic behavior. Chem. Eng. Res. Des. 2010, 88, 135−145. (23) Bravo-Bravo, C.; Segovia-Hernández, J. G.; Gutiérrez-Antonio, C.; et al. Extractive dividing wall column: Design and optimization. Ind. Eng. Chem. Res. 2010, 49, 3672−3688. (24) Xia, M.; Yu, B. R.; Wang, Q. Y.; et al. Design and control of extractive dividing−wall column for separating methylal−methanol mixture. Ind. Eng. Chem. Res. 2012, 51, 16016−16033. (25) Errico, M.; Rong, B. G.; Tola, G.; et al. Optimal synthesis of distillation systems for bioethanol separation. Part 2: extractive distillation with complex columns. Ind. Eng. Chem. Res. 2013, 52, 1620−1626. (26) Yin, W.; Ding, S. H.; Xia, S. Q.; Ma, P. S.; Huang, X. J.; Zhu, Z. S. Cosolvent selection for benzene−cyclohexane separation in extractive distillation. J. Chem. Eng. Data 2010, 55, 3274−3277.

conditions. The entrainer temperature and entrainer feed stage and the mixture feed stage also affect the energy requirements. This Article improves the control scheme that was proposed in literature and real industry. Two control structures are used for this extractive distillation process. SVD and slope criteria are used to select the temperature control trays. Feed flow rate and composition disturbances are applied to test their effectiveness. It is found that the initial control structure with a fixed reflux ratio has large transient deviations when +20% feed flow rate disturbance is introduced. To get a more robust control, the improved control structure with feed forward control is adopted. The dynamics simulation results demonstrate that the improved control structure has shorter settling time, and its controlled temperatures fluctuate much less than those in the initial control structure. The results demonstrate that the improved control structure with feed forward is much better than the initial control structure.

■

AUTHOR INFORMATION

Corresponding Author

*Tel.: +86 519 86330355. Fax: +86 519 86330355. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We are thankful for support from the project fund of the China Petroleum & Chemical Corp. (411024) and assistance from the staff at the Institute of Petrochemical Technology (Changzhou University and China University of Petroleum Beijing).

■

NOMENCLATURE F = the fresh feed of the extractive distillation column R/F = the mass ratio of reflux flow rate/flow rate of F EF = the entrainer feed EDC = the extractive distillation column ERC = the entrainer recovery column F2 = the feed of ERC RR = the reflux ratio EF/F = the molar ratio of entrainer to feed FS = the stage of fresh feed EFS = the entrainer feed stage ET = the entrainer temperature RR1 = the reflux ratio of EDC XD1 = the purity of top product in EDC XD2 = the purity of top product in ERC RR2 = the reflux ratio of ERC QR1 = the reboiler duty of EDC (GJ/h) QR1/F = the reboiler duty of EDC/molar flow rate of F QR2/F = the reboiler duty of ERC/molar flow rate of F QHX/F = the cooler duty/molar flow rate of F

■

REFERENCES

(1) Han, F. S.; Lin, K. Z. Development trends of aromatic production and technology in the world. Pet. Petrochem. Today 2006, 14, 30−35. (2) Knapp, J. P.; Doherty, M. F. Thermal integration of homogeneous azeotropic distillation sequences. AIChE J. 1990, 36, 969−983. (3) Lei, Z. G.; Li, C. Y.; Chen, B. H. Extractive distillation: A review. Sep. Purif. Technol. Rev. 2003, 32, 121−213. (4) Luyben, W. L.; Chien, I. L. Design and Control of Distillation Systems for Separating Azeotropes; John Wiley & Sons, Inc.: New York, 2010. L



Article

(27) Doherty, M.; Malone, M. Conceptual Design of Distillation Systems; McGraw-Hill: New York, 2001. (28) Knight, J. R.; Doherty, M. F. Optimal design and synthesis of homogeneous azeotropic distillation sequences. Ind. Eng. Chem. Res. 1989, 28, 564−572. (29) Luyben, W. L. Distillation Design and Control Using Aspen Simulation; John Wiley & Sons, Inc.: New York, 2006.

M


Control of BenzeneâCyclohexane Separation System via Extractive

Recommend Documents

Control of BenzeneâCyclohexane Separation System via Extractive