Process Assessment of Heterogeneous Azeotropic Dividing-Wall

Jul 25, 2016 - ... and Engineering, Chemical Engineering Research Center, and School of Chemical Engineering and Technology, Tianjin University, Tianj...
2 downloads 12 Views 3MB Size
Subscriber access provided by CORNELL UNIVERSITY LIBRARY

Article

Process Assessment of Heterogeneous Azeotropic Dividing-Wall Column for the Ethanol Dehydration with Cyclohexane as an Entrainer: Design and Control Ye Li, Ming Xia, Weisong Li, Junwen Luo, Lei Zhong, Shanyuan Huang, Jian Ma, and Chunjian Xu Ind. Eng. Chem. Res., Just Accepted Manuscript • DOI: 10.1021/acs.iecr.6b01244 • Publication Date (Web): 25 Jul 2016 Downloaded from http://pubs.acs.org on July 29, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Industrial & Engineering Chemistry Research is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

Process Assessment of Heterogeneous Azeotropic Dividing-Wall Column for the Ethanol Dehydration with Cyclohexane as an Entrainer: Design and Control Ye Li,1 Ming Xia,2 Weisong Li,1 Junwen Luo,1 Lei Zhong,1 Shanyuan Huang,1 Jian Ma,1 and Chunjian Xu1,* 1

State Key Laboratory of Chemical Engineering, Collaborative Innovation Center of Chemical Science and Engineering, Chemical Engineering Research Center, and School of Chemical Engineering and Technology, Tianjin University, Tianjin 300072, China 2

Institute of Coal Chemistry, Chinese Academy of Sciences, Taiyuan 030001, China Tel: +86 022-27404440. Fax: +86 022-27404440. E-mail: [email protected].

ABSTRACT: Heterogeneous azeotropic dividing-wall column (ADWC) features significant reduction of capital investment and energy consumption. Conventional azeotropic distillation sequences can be thermally coupled into two types of configurations of ADWC, namely, the original ADWC with simple azeotropic/recovery section and the other with recovery section also severing as preconcentrator. The ADWC with combined recovery/preconcentrator section has been investigated on both steady design and dynamic controllability. However, the original ADWC has only been proved to be more energy saving than conventional azeotropic/recovery column sequence, not compared with conventional azeotropic/stripper column sequence. Furthermore, the dynamic performances of this ADWC are still unclear. Note that all the benefits of ADWC are reasonable only under good dynamic controllability. Thus, in this work, we develop a comparative study of optimal

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

design and dynamic performance of ADWC with two conventional sequences by demonstrating ethanol dehydration using cyclohexane as an entrainer. The results show that ADWC has significant capital investment reduction and energy saving than two conventional sequences. As two reboiler duties are still preserved in ADWC, it has comparable dynamic controllability with conventional azeotropic/recovery column sequence and superior dynamic controllability with conventional azeotropic/stripper column sequence. Key words: heterogeneous azeotropic, dividing-wall column, ethanol dehydration, economic optimization, control strategy 1. INTRODUCTION Dividing-wall column (DWC), as a promising technology for process intensification, has more advantages than conventional distillation column since it usually provides significant reductions in capital costs and energy consumptions1-3. Recently, much more attention has been paid on the extension of DWC for extractive4-9, reactive10-13 and azeotropic14-19 distillation. Heterogeneous azeotropic dividing-wall column (ADWC) featuring the integration of azeotropic and entrainer recovery columns into one shell, has attracted several researchers as it has promising energy saving potential. In conventional heterogeneous azeotropic distillation, different ternary systems with various residue curve maps favor different distillation sequences. These sequences can be thermally coupled and mainly converted into two types of ADWC configurations as shown in figure 1b and e, with their corresponding equivalent scheme shown in figure 1c and f. Generally speaking, the feed with concentrated component close to the azeotrope composition enters into the azeotropic column of conventional sequence as shown in Figure 1d and into the left part of ADWC as shown in Figure

ACS Paragon Plus Environment

Page 2 of 37

Page 3 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

1e20-22. In case of the diluted feed, the recovery column can be usually combined with the preconcentrator, and thereby the diluted feed is fed into the recovery column of conventional sequence as Figure 1a and into the right part of ADWC as depicted in Figure 1b20,23,24. For the type 1 depicted in Figure 1a, b and c, Wu investigated the design and control of this configuration through pyridine dehydration17 and 1, 4-dioxane dehydration18. They found that significant energy saving and total annual cost (TAC) reduction can be obtained, and the control performance was comparable to original two-column sequence. Yu19 applied this configuration to t-butanol dehydration with similar conclusions as Wu. For the type 2 depicted in Figure 1d, e and f, Sun14 and Kiss15 applied this configuration to ethanol dehydration. Le16 applied this configuration to the separation of water, acetic acid and an organic component. All of their researches revealed that ADWC provides energy saving compared with conventional azeotropic/recovery column sequence. For energy saving of the conventional azeotropic distillation column sequence, recovery column can be replaced by stripper, and it has been proved that this modified sequence gains further energy saving20. However, no paper in the open literature compared the energy consumption of ADWC with conventional azeotropic/stripper column sequence. On the other hand, as the integration of two columns into one shell will change the operating mode and controllability of such thermally coupled DWC, all the benefits of reduction of capital investment and energy consumption are reasonable only if a good control strategy is available. However, there is no paper in the open literature has studied the control performance of this configuration ADWC. Thus, this study focuses on the energy saving potential and dynamic performance of this ADWC, investigating by comparing with the conventional two sequences.

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

V1

Page 4 of 37

D2

Rrec

Makeup

Organic Phase

Rref Organic Phase

Organic Phase

L Makeup Makeup

C1

C2

D1

LR

LR Aqueous Phase

F

Aqueous Phase

F Aqueous Phase Dividing wall F

A

B

A

B

(a)

B

A

(b)

V1

(c)

D2

Rrec

Makeup

Organic Phase

Organic Phase

Rref L

Organic Phase

Makeup

Makeup

C1

F

C2

D1

F

F

LR Aqueous Phase

LR

Aqueous Phase Aqueous Phase Dividing wall

A

B

A

B

(e)

(d)

A

B

(f)

Figure 1. (a) Type1: conventional azeotropic distillation sequence with combined recovery/preconcentrator column (b) Type1: thermally coupled scheme for ADWC (c) Type1: equivalent scheme of ADWC (d) Type2: conventional azeotropic distillation sequence with simple azeotropic and recovery column (e) Type2: thermally coupled scheme for ADWC (f) Type2: equivalent scheme of ADWC

In this study, ethanol dehydration using cyclohexane as entrainer is selected to compare the relative merits of ADWC with two conventional sequences, involving the issues of both steady-state economics and dynamic controllability. Conventional temperature control structures are established to explore the dynamic performance of ADWC, compared with two conventional sequences. 2. STEADY STATE DESIGN 2.1. Thermodynamic Model and Residue Curve Maps. For ethanol dehydration through heterogeneous azeotropic distillation, cyclohexane is commonly selected as the entrainer to form a ternary heterogeneous azeotrope with ethanol and water.

ACS Paragon Plus Environment

Page 5 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

In this study, the NRTL-RK model is selected and the built-in binary interaction parameters in Aspen Plus are used to predict the vapor-liquid equilibrium and liquid-liquid equilibrium of the system. Residue curve maps(RCMs) is an important and useful method in the conception design of distillation. Figure 2 shows the residue curve maps of ethanol/water/cyclohexane system at 101.33 kPa. Three binary azeotropes and one ternary minimum-boiling azeotrope can be found in the system. The ethanol/water mixture has an azeotrope with composition of 89.21 mol% ethanol and an azeotropic temperature of 351.31K at 101.33 kPa. The ternary minimum-boiling azeotrope with temperature of 335.21K at 101.33 kPa is inside the LLE envelope. The residue curve maps (RCMs) of the ternary system is divided into three distillation regions by the distillation boundaries.

Figure 2. Ternary phase diagram and residue curve map at 101.33 kPa (molefr. basis).

2.2. Optimization The feed composition and flowrate are selected as the same as Sun14, with composition of 90 wt% ethanol and flowrate of 1000 kg/h. The product purities of ethanol and water are both specified at 99.9 mol%. The conventional azeotropic/recovery column sequence is called A1, while the sequence

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

with stripper is called A2. In this work, the total annual cost (TAC) suggested by Douglas25 including annual capital costs and operating costs, is used as the objective function to be minimized to screen the optimal design among feasible alternatives. It can be expressed by the equations TAC (103 $/year) = OC + Cf + (ir + im )CI And CI = installed column shell cost + installed column tray cost + installed heat exchanger cost, where OC is the operating cost, mostly utility consumption (stream, cooling water, and electricity); Cf is the annual fixed cost, such as maintenance expense and wages; CI is the fixed capital investment; ir is the fixed capital recovery rate applied to CI; and im is the minimum acceptable rate of return on CI. Cf is assumed to be 10% of CI, and ir + im is assumed to be 20% of CI in this optimization. The details of both equations for determining OC and CI are mentioned in Douglas’ book25. The costs of the entrainer and entrainer makeup are not included in OC because they are much lower than the costs of the heat duties. Other details can be referenced in supporting information. 2.3.1. Optimization of Sequence A1. The operation pressure of the two columns is selected to permit the temperature of overhead vapor sufficiently higher than the available cooling water temperature so the use of cooling water in the condenser is available. The distillate composition of entrainer recovery column has significant influence on the reboiler duties and the diameters of the columns, which influences the energy consumption and capital cost 20,24

. Figure 3 shows the TAC of different distillate concentration of entrainer recovery column. There

exists a minimum TAC with the optimal ethanol concentration 74.0 mol%.

ACS Paragon Plus Environment

Page 6 of 37

Page 7 of 37

840 830 820 810 TAC($1000/annum)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

800 790 780 770 760 750 740 0.71

0.72

0.73

0.74

0.75

0.76

0.77

xD2(molefrac of ethanol)

Figure 3. Summary of TAC plots at various mole fraction xD2, ethanol.

As the distillate ethanol concentration of entrainer recovery column is fixed at 74.0 mol%, a number of other parameters for sequence A1 should be determined: total stage numbers (NT1, NT2), reflux ratios (RR2) in the entrainer recovery column, organic reflux rate(OR), feed locations (NF, NFA) of the fresh feed and aqueous stream. They all should be varied to find the most economical design under the specified product yields and purities. The iteration procedure of the optimization is shown in Figure 4.

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 4. Global optimization sequence.

The optimized design flowsheet by minimizing TAC can be seen in Figure 5. The tray numbers of heterogeneous azeotropic distillation column and entrainer recovery column are 45 and 18, respectively.

ACS Paragon Plus Environment

Page 8 of 37

Page 9 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

V1

D2

3188.08 kg/h, 53.7602 kmo/h Cyclohexane=48.51 mol % Ethanol=32.79 mol % Water=18.70 mol %

900.15 kg/h, 21.3332 kmo/h Cyclohexane=5.170 mol % Ethanol=74.00 mol % Water=20.83 mol %

Cond. Duty -593.09 kW

Cond. Duty -436.81 kW

Cyclohexane Makeup 1.0000E-7 kmol/h 1 Reflux Ratio=0.8407 kg/kg 14

F 1000 kg/h, 25.0867 kmol/h Cyclohexane=0 Ethanol=90.00 wt % Water=10.00 wt %

C1

Organic Reflux 2187.98 kg/h, 26.8899 kmol/h Cyclohexane=92.89 mol % Ethanol=6.820 mol % Water=0.2976 mol %

C2 11

Aqueous Flow

N1=45 D=0.56 m

1000.04 kg/h, 26.8695 kmol Cyclohexane=4.110 mol % Ethanol=58.77 mol % Water=37.12 mol %

N2=18 D=0.46 m

Reb. Duty 634.59 kW

Reb. Duty 460.92 kW

B1

B2

900.00 kg/h, 19.5478 kmol/h Cyclohexane=2.3484E-11 mol % Ethanol=99.90 mol % Water=0.1000 mol %

99.89 kg/h, 5.5363 kmol/h Cyclohexane=3.5130E-21 mol % Ethanol=0.1000 mol % Water=99.90 mol %

Figure 5. Optimal design of sequence A1.

In the optimization of distillation sequence, there exists tradeoff between the capital cost and the operating cost, which will complicate the benefit of the new design versus the original design. As the complete optimization is time costing, for simplicity purposes, we assume that the optimized total stage numbers of sequence A1 are preserved for the optimization of sequence A2 and ADWC. Namely, both total stage numbers of the heterogeneous azeotropic column of sequence A2 and ADWC are kept at 45, and that of entrainer recovery column of sequence A2 and stripper section of ADWC are kept at 18. This setting is also beneficial to the comparison of dynamic performance between these sequences. 2.3.2. Optimization of Sequence A2. For Sequence A2, stripper column is used so that energy carried by the overhead vapor is conserved instead of being removed in a condenser, which is considered to be more energy-saving than the conventional recovery column. However, the expensive compressor is also introduced at the same time, which will increase the capital investment and operating cost. Meanwhile, for the stripper

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 37

column, the design degree-of-freedom at the top of the conventional recovery column is lost. Therefore, the ethanol concentration of distillate vapor of stripper column can’t be manipulated by reflux ratio, which is actually an output variable. As the total stage numbers of heterogeneous azeotropic column and stripper column are fixed, feed locations (NF, NFA) of the fresh feed and aqueous stream, organic reflux rate(OR), are varied to optimize the design basing on TAC. Figure 6 shows the result of optimization of sequence A2. The ethanol concentration of distillate vapor in stripper column is reduced to 67.9 mol% compared with 74.0 mol% of sequence A1, which indicates larger distillate rate of stripper column. Comp. power 29.84 kW

V1

D2

5070.55 kg/h, 85.4583 kmo/h

1500.10 kg/h, 37.2621 kmo/h Cyclohexane=4.850 mol %

Cyclohexane=48.46 mol % Ethanol=33.02 mol % Water=18.52 mol %

Cond. Duty -943.05 kW 313.0 K

Ethanol=67.85 mol % Water=27.30 mol %

Cyclohexane Makeup 1.0000E-7 kmol/h 1

14

F 1000 kg/h, 25.0867 kmol/h Cyclohexane=0 Ethanol=90.00 wt % Water=10.00 wt %

C1

1

Organic Reflux 3470.76 kg/h, 42.6626 kmol/h Cyclohexane=92.85 mol % Ethanol=6.850 mol % Water=0.2973 mol %

C2

Aqueous Flow

N1=45 D=0.51 m

1599.99 kg/h, 42.7980 kmol Cyclohexane=4.220 mol % Ethanol=59.09 mol % Water=36.69 mol % Reb. Duty 525.87 kW

N2=18 D=0.46 m Reb. Duty 452.81 kW

B1

B2

900.00 kg/h, 19.5478 kmol/h

99.87 kg/h, 5.5359 kmol/h Cyclohexane=3.6454E-19 mol % Ethanol=0.1000 mol %

Cyclohexane=1.3247E-5 mol % Ethanol=99.90 mol % Water=9.9987E-2 mol %

Water=99.90 mol %

Figure 6. Optimal design of sequence A2.

2.3.3. Optimization of ADWC. In the ADWC, liquid sidestream from certain location in the heterogeneous azeotropic column enters into the top of stripper column. The overhead vapor of stripper column is compressed by a virtual compressor and enters into the heterogeneous azeotropic column at the same tray of liquid

ACS Paragon Plus Environment

Page 11 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

sidestream. Therefore, the heterogeneous azeotropic column gets a new degree of freedom as liquid sidestream rate. And stripper column loses a degree of freedom, which results in that ethanol concentration of vapor distillate can’t be specified. The left reboiler duty is manipulated to maintain ethanol product purity at 99.9 mol % and the right reboiler duty is manipulated to maintain water product purity at 99.9 mol %. The cyclohexane makeup flow is introduced to balance out the tiny loss of entrainer through the two product streams. When the total stage numbers of heterogeneous azeotropic column and stripper column are fixed, feed locations (NF, NFA) of the fresh feed and aqueous stream, liquid sidestream location(NLR) and flow rate(LR), organic reflux rate(OR), are varied to optimize the design basing on TAC. Figure 7 shows the result of optimization of ADWC. Cond. -475.38 kW

V1 2465.01 kg/h, 42.9580 kmo/h Cyclohexane=45.91 mol % Ethanol=32.08 mol % Water=22.01 mol %

D2 1492.48 kg/h, 31.0106 kmo/h Cyclohexane=18.63 mol % Ethanol=63.42 mol % Water=17.95 mol %

Organic Reflux 1663.77 kg/h, 20.4112 kmol/h Cyclohexane=93.27 mol % Ethanol=6.440 mol % Water=0.2977 mol %

1

1

8

F 1000 kg/h, 25.0867 kmol/h Cyclohexane=0 Ethanol=90.00 wt % Water=10.00 wt %

N1=45 D=0.50 m

Cyclohexane Makeup 1.0000E-7 kmol/h 16

C1

LR 791.13 kg/h, 14.0000 kmo/h Cyclohexane=36.37 mol % Aqueous Flow Ethanol=51.47 mol % 801.23 kg/h, 22.5467 kmol Water=12.16 mol % Cyclohexane=3.040 mol % Ethanol=55.29 mol % Water=41.67 mol %

C2 14

Reb. Duty 520.71 kW

N2=18 D=0.40 m Reb. Duty 340.85 kW

B1

B2

900.00 kg/h, 19.5477 kmol/h Cyclohexane=2.3414E-4 mol % Ethanol=99.90 mol % Water=9.9766E-2 mol %

99.89 kg/h, 5.5361 kmol/h Cyclohexane=0 mol % Ethanol=0.1000 mol % Water=99.90 mol %

Figure 7. Optimal design of ADWC.

It is noticed that the equivalent diameter (De) of the lower part of ADWC is calculated, and for practical consideration, it is used as the diameter of the ADWC in the optimization.17 The equivalent diameter of the ADWC is illustrated in Figure 8.

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

πDଶୣ A୐ + Aୖ = Aୣ = 4

AL

AR πDଶୖ Aୖ = 4

πDଶ୐ A୐ = 4

AL

AR

Figure 8. Illustration of how to calculate the equivalent diameter (De) on the lower part of dividing-wall column.

2.4. Comparison. Table 1 shows the comparison of these three sequences. The total reboiler duty of ADWC decreases by 21.36% and 11.97% than sequence A1 and A2, respectively. The operating cost of ADWC decreases by 21.07% and 14.93% than sequence A1 and A2, respectively. Note that sequence A2 with stripper is indeed more energy saving than sequence A1, whereas not better than ADWC. The fixed capital investment of ADWC decreases by 14.09% and 23.35% than sequence A1 and A2, respectively. The utilization of expensive compressor of sequence A2 makes it inferior to sequence A1 on fixed capital investment, which eventually leads to the almost equivalent TAC with sequence A1. However, ADWC shows a significant reduction of TAC about 18.5% than both sequence A1 and A2. Therefore, from the standpoint of steady-state design economics, ADWC is the best. Table 1. Economic comparison among the optimal designs of sequences A1, A2, and ADWC

ACS Paragon Plus Environment

Page 12 of 37

Page 13 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

items\sequences

A1

A2

ADWC

saving gaina/%

QR, total (kW)

1095.51

978.68

861.56

-21.36/-11.97

OC (103$/year)

429.894

398.846

339.313

-21.07/-14.93

CI (103$/year)

846.591

948.872

727.303

-14.09/-23.35

TAC (103$/year)

683.871

683.508

557.504

-18.48/-18.43

a

: saving gain = (itemADWC, total - itemAi, total)/itemAi, total×100 %, where i = 1, 2; item = QR, total, OC, CI, TAC.

The main reason for the significant reduction of TAC and energy requirement can be explained by using the residue curve maps (RCMs) and material balance lines of sequence A1 and ADWC as an example. Figure 9 (a) shows the conceptual diagram of the separation via sequence A1. Figure 9 (b) shows the conceptual diagram of the separation via ADWC. Compared with sequence A1, in sequence ADWC, main column has the liquid side stream LR, which results in the vapor distillate D2 of stripper column increasing and moving towards bottom right and closing to distillation boundary, with composition reducing to 63.4 mol%. Thus the boilup ratio of stripper column decreases leading to the decreasing of reboiler duty. The liquid side stream LR makes the split point of B1 and LR moving from B1 to S. So |SM2|/|SV1| decreases which results in the distillate vapor V1 of ADWC decreases roughly. Thus the energy consumption of the main column of heterogeneous azeotropic dividing column decreases. This indicates a smaller organic reflux flow rate back into the main column and also smaller aqueous outlet flow rate into the stripper column, which in turn requires less reboiler duty for the two columns to generate the vapor rate going up the column and smaller diameters for the two columns. Though ADWC has the minimum TAC, the dynamic performance of ADWC is still need to be investigated.

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(a)

(b)

Figure 9. (a) Ternary phase diagram and residue curve map for the optimum design of sequence A1 (b) Ternary phase diagram and residue curve map for the optimum design of sequence ADWC (molefr. Basis).

3. DYNAMIC CONTROL In this section, dynamic control of the proposed design flow sheet will be explored. A conventional temperature control structure and an improved temperature control structure with two feedforward ratio controllers QR1/F and QR2/F are established. The sizes of equipment are necessary to convert a steady-state simulation to dynamic one. The “Tray Sizing” section in Aspen Plus is used to define the sizes of equipment such as decanter, reflux drum and column base. The column bases and reflux drums are sized to provide 5 min of holdup when at the 50% liquid level. Pumps and valves are inserted to provide adequate pressure drops so that the flow sheet is fully pressure-driven and the valves have good range ability. Most valves are specified to have pressure drops of 300 kPa with the valve half open at the design flow rate. For ADWC, the steady-state RadFrac model in Aspen Plus consists of two column sections: a main column (with both reboiler and condenser) and a stripper column (with only reboiler). When exploring the dynamic performance of ADWC, a three-column model is implemented using Aspen

ACS Paragon Plus Environment

Page 14 of 37

Page 15 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

Dynamics. This three-column model is the equivalent flowsheet of ADWC, of which the main column is divided into the upper column section C1 and the lower column section C12, with stripper column unchanged. Pumps and valves are vital for a realistic dynamic simulation. The valves must have sufficient pressure drop for effective dynamic controllability. Control valves and pumps are sized to give the rangeability required to handle the disturbance in the dynamic simulations. In addition, several fictitious items are inserted into the flowsheet of ADWC. Two fictitious compressors are added to compensate the pressure drop. The first one is added to the vapor sidestream outlet. The other is added to the overhead vapor stream of lower column section. A fictitious Fsplit is inserted between the upper column section and the lower column section/stripper column. Two fictitious valves are inserted. One lies on the liquid sidestream from the Fsplit. The second lies on the other stream from Fsplit. A fictitious pump is added on the bottom stream of upper column section to compensate the pressure drop over the downstream fictitious valves. 3.1. Basic Temperature Control Structure. In this study, sensitive tray is selected by “slope criterion” suggested by Luyben.26 Figure 10 shows the temperature profile and composition profile of main column and stripper column of ADWC at the design conditions. The location of the tray with the largest slope is the 38th stage of the main column (the 22th stage of the lower column section) and the 17th stage of stripper column. Both sensitive trays of main column and stripper column are close to the reboiler. Therefore, to ensure quick and effective response of the loop, changing reboiler duty to control the temperature of 38th stage of main column and changing reboiler duty to control the temperature of 17th stage of stripper column. The control loop is listed as below:

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(1) Fresh feed rate is flow controlled. (2) The pressure of main column is controlled by the valve in the overhead vapor line of the main column. (3) The base level of upper column section is controlled by manipulating the bottom flowrate. (4) The liquid sidestream from the upper column section to the stripper column is flow controlled, which is ratioed to the bottom stream of upper column section. And the setpoint signal of the flow controller is the output signal of the LR/L controller. So this flow controller is “on cascade”. (5) The temperature of the decanter is controlled by manipulating the heat removal of the condenser on the upstream line of the decanter. (6) Aqueous flow rate is flow controlled. Aqueous level of decanter is level controlled by manipulating the aqueous flow rate. The setpoint of the flow controller is the output signal of the level controller, so this flow controller is “on cascade”. (7) As the losses of cyclohexane are too small to measure, the control of makeup is not considered. (8) Organic reflux rate is flow controlled, which is ratioed to the sum of fresh feed flowrate and vapor distillate flowrate of stripper column using the controller R/SUM1. The setpoint of the flow controller is the output signal of R/SUM1 controller, so this flow controller is “on cascade”. (9) The pressure of lower column section is controlled by manipulating the rotate speed of the fictitious compressor. (10) The base level of lower column section is controlled by manipulating the bottom flowrate. (11) The pressure of stripping column is controlled by manipulating the rotate speed of the fictitious compressor.

ACS Paragon Plus Environment

Page 16 of 37

Page 17 of 37

(12) The temperature of overhead vapor entering into the upper column section is controlled by manipulating the heat removal of the fictitious condenser. (13) The base level of stripping column is controlled by manipulating the bottom flowrate. 375

380

370

Temperature(K)

Temperature(K)

370

365

360

350

360

350

355

0

10

20

30

38 40

340

50

0

5

Main Column Stage

15

17

20

1.0 Cyclohexane Ethanol Water

Cyclohexane Ethanol Water

0.8 Liqiud Mole Fraction

0.8

0.6

0.4

0.6

0.4

0.2

0.2

0.0

10 Stripper Column Stage

1.0

Liqiud Mole Fraction

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

0

10

20

30

40

50

0.0

0

Main Column Stage

5

10

15

20

Stripper Column Stage

Figure 10. Temperature profiles and liquid composition profiles in ADWC.

Most of the loops discussed above are standard distillation control strategies. All of the controllers are PI(proportional and integral) controllers. Because of the existence of measurement and actuator lags in any real physical system, a 1-min dead time is inserted in the two temperature control loops. Common PI settings are utilized for the control loops. Relay-feedback tests are run on the two temperature controllers to achieve the ultimate gains and periods. Then, Tyreus−Luyben turning is applied to obtain the gain KC and integral time constant τI. Details of tuning and setting are covered in Luyben’s book26. Figure 11 shows the basic temperature controller structure TCS1 and

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 37

controller faceplate. The tuning parameters of the two temperature controllers are listed in Table 2.

Figure 11. Overall control strategy of ADWC using basic control structure TCS1. Table 2. Controller tuning parameters for TCS1 controlled variable

manipulated variable

controller gain KC (%)

TC22

T22

QR1

1.60

ACS Paragon Plus Environment

controller integral time τI (min) 11.88

Page 19 of 37

TC17

T17

QR2

1.93

9.24

The effectiveness of the temperature control structure TCS1 for the ADWC is tested by making 20% disturbances in the feed flow rate and 2 wt% ethanol concentration disturbances in the feed composition. Both of feed flowrate and feed composition disturbances are introduced at the time equal to 0.2 h. Figure 12 gives the dynamic responses for feed flowrate disturbance. The results showed that both temperature control points are able to return back to their set points quickly. Both product purities are maintained fairly close to their desired values. Figure 13 gives the dynamic responses for feed composition disturbance. Under +2 wt% feed composition disturbance, both temperature control points and both product purities are able to return back to their set points quickly. However, for -2 wt% feed composition disturbance, a small negative offset (about −0.5 wt %) of ethanol product purity maintains under the new stable steady state, while water product purity and both temperature control points return back to their set points quickly. As the total stage numbers of ADWC are not optimized but selected as the same as sequence A1, when the feed is diluted by -2 wt% ethanol, it is most probably that total stage numbers of azeotropic and recovery section are not sufficient for the separation of the diluted feed. 1.0000

1.0000

0.9995

0.9995

xB2(molefrac of water)

xB1(molefrac of ethanol)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

0.9990

0.9985

0.9990

+20% feed flowrate -20% feed flowrate

0.9985

+20% feed flowrate -20% feed flowrate

0.9980

0

2

4

6

8

10

0.9980

0

2

Time(h)

4

6 Time(h)

ACS Paragon Plus Environment

8

10

374.0

372

373.5

T17(K)

374

370

+20% feed flowrate

368

366

2

4

6

8

+20% feed flowrate -20% feed flowrate

372.5

-20% feed flowrate

0

Page 20 of 37

373.0

372.0

10

0

2

4

Time(h)

6

8

10

Time(h)

1.000

1.0000

0.998

0.9995

xB2(molefrac of water)

xB1(molefrac of ethanol)

Figure 12. Dynamic responses to feed flowrate disturbance using control structure TCS1.

0.996 92wt% ethanol 88wt% ethanol 0.994

0.992

0.9990

92wt% ethanol 88wt% ethanol

0.9985

0

2

4

6

8

0.9980

10

0

2

4

Time(h)

6

8

10

Time(h)

371.00

374.00

370.75

373.75

T17(K)

T22(K)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

T22(K)

Industrial & Engineering Chemistry Research

370.50

92wt% ethanol

370.25

373.50

92wt% ethanol 88wt% ethanol

373.25

88wt% ethanol

370.00

0

2

4

6

8

10

373.00

0

2

Time(h)

4

6

8

10

Time(h)

Figure 13. Dynamic responses to feed composition disturbance using control structure TCS1.

3.2. Improved Temperature Control Structure with Two Feedforward Ratio Controllers QR1/F and QR2/F. In many researches, various feedforward ratio controllers are found very effective to reject the disturbances. For ADWC, as two reboiler duties are still preserved, two feedforward ratio controllers

ACS Paragon Plus Environment

Page 21 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

with fixed QR1/F and QR2/F ratio are added in the control loops, which are expected to improve the dynamic performance. The improved temperature control structure is called TS2. Each of the ratio loops is set up by adding a multiplier block into control structure. Each of the reboiler heat input of heterogeneous azeotropic distillation column and stripper column is ratioed to the corresponding feed flowrate. The tuning parameters of the two temperature controllers are listed in Table 3. Table 3. Controller tuning parameters for TCS2 controlled variable

manipulated variable

controller gain KC

controller integral time

(%)

τI (min)

TC22

T22

QR1/F

1.97

10.56

TC17

T17

QR2/F2

1.74

9.24

Figure 14 shows the improved temperature controller structure TCS2 and controller faceplate.

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 14. Overall control strategy of ADWC using Improved temperature control structure TCS2.

Figure 15 gives the dynamic responses for feed flowrate disturbance. Robust control is achieved with both product purities and both temperatures returned close to their specifications. As the reboilers of two columns are ratioed to feed flowrate, the directly and quickly response of the reboilers under the feed flowrate disturbances determines the narrower but larger transient deviations of the temperature control points. Figure 16 gives the dynamic responses for feed composition disturbance. A small negative offset (about −0.5 wt %) of ethanol product purity still maintains under the new stable steady state, while the controlled temperatures return back to their set points quickly. Narrower and larger transient deviations of both the temperature control points still appear. In general, the control loop with two feedforward ratio controller features quicker response against the feed flowrate and composition disturbances than the original control loop. Thus, TS2 is applied to the two conventional sequences to compare the dynamic performance with ADWC.

ACS Paragon Plus Environment

Page 22 of 37

Page 23 of 37

1.0000

1.0000

xB2(molefrac of water)

xB1(molefrac of ethanol)

0.9995

0.9990

0.9985

0.9980

+20% feed flowrate -20% feed flowrate

0.9995

0.9990

0.9985

+20% feed flowrate -20% feed flowrate

0.9975

0.9970

0

2

4

6

8

0.9980

10

0

2

4

Time(h)

6

8

10

Time(h)

375

375

374

T17(K)

T22(K)

370

373 +20% feed flowrate

365 +20% feed flowrate -20% feed flowrate

360

0

2

4

6

8

-20% feed flowrate

372

371

10

0

2

4

A (Time)

6

8

10

Time(h)

Figure 15. Dynamic responses to feed flow disturbance using control structure TCS2.

1.000

1.0000

0.998

0.9995

xB2(molefrac of water)

xB1(molefrac of ethanol)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

0.996 92wt% Ethanol 88wt% Ethanol 0.994

0.992

0

2

4

6

8

10

0.9990

92wt% ethanol 88wt% ethanol

0.9985

0.9980

0

2

Time(h)

4

6 Time(h)

ACS Paragon Plus Environment

8

10

Industrial & Engineering Chemistry Research

372

374.0

373.5

T17(K)

371 T22(K)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 37

92wt% ethanol

92wt% ethanol

370

88wt% ethanol

369

373.0

0

2

4

6

8

88wt% ethanol

372.5

10

372.0

0

2

Time(h)

4

6

8

10

Time(h)

Figure 16. Dynamic responses to feed composition disturbance using control structure TCS2.

3.3. Comparison of Dynamic Performance of ADWC and Sequence A1 Using Temperature Control Structure TCS2. The dynamic performance of sequence A1 is also implemented using temperature controller structure TCS2 to compare with sequence ADWC. Figure 17 shows the improved temperature controller structure TCS2 and controller faceplate of sequence A1.

ACS Paragon Plus Environment

Page 25 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

Figure 17. Overall control strategy of sequence A1 using control structure TCS2.

Figure 18 gives the dynamic responses of sequence A1 and ADWC using temperature control structure TCS2 under feed flowrate disturbances. It can be seen that sequence A1 and ADWC have almost similar dynamic performances. Both product purities and both temperature control points can return to the specifications quickly. Larger transient deviations of the variables of ADWC reveal the tighter and more interactive controllability of ADWC than sequence A1. Figure 19 gives the dynamic responses for composition disturbance. Sequence A1 and ADWC have almost similar dynamic performance. A small offset (0.2 wt%) of ethanol product purity of sequence A1 is observed

under -2 wt% ethanol disturbance. This small offset indicates the

existence of the balance between the steady state design and dynamic controllability of sequence A1. As the offset is such small, more total stages of azeotropic column and recovery column of sequence A1 may easily solve this little shortcoming. As ADWC is thermally coupled with interactive control loops, it is reasonable that the offset of ethanol product purity of sequence A1 is expanded for

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research

ADWC under the same total stage numbers of azeotropic and recovery column. 1.0000

1.0000 +20% feed flowrate

+20% feed flowrate xB2(molefrac of water)

xB1(molefrac of ethanol)

0.9995

0.9990

0.9985 sequence A3 sequence A1

0.9980

0.9995

0.9990

sequence A3

0.9985

sequence A1

0.9975

0.9970

0

2

4

6

8

0.9980

10

0

2

4

Time(h)

6

8

10

Time(h)

1.0000

1.0000 -20% feed flowrate

0.9995

xB2(molefrac of water)

xB1(molefrac of ethanol)

-20% feed flowrate

0.9990

sequence A3 0.9985

0.9980

sequence A1

0

2

4

6

8

10

0.9995

0.9990

sequence A3 sequence A1

0.9985

0.9980

0

2

4

Time(h)

6

8

10

Time(h)

375.0

375 +20% feedflow rate

+20% feed flowrate 374

370.0

T17(K)

372.5 T22/T26(K)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 37

sequence A3(T22) sequence A1(T26)

373 sequence A3(T17) sequence A1(T17)

367.5

365.0

372

0

2

4

6

8

10

371

0

2

Time(h)

4

6 Time(h)

ACS Paragon Plus Environment

8

10

Page 27 of 37

372

375 -20% feed flowrate 374 -20% feed flowrate

366

T17(K)

T22/T26(K)

369

363

373 sequence A3(T17)

sequence A3(T22) sequence A1(T26)

sequence A1(T17)

372

360

0

2

4

6

8

371

10

0

2

4

Time(h)

6

8

10

Time(h)

Figure 18. Dynamic responses to feed flow rate disturbances for sequences ADWC and A1 using structure TCS2.

1.0000

1.0000

0.9995

0.9995

xB2(molefrac of water)

xB1(molefrac of ethanol)

92wt% ethanol

92wt% ethanol

0.9990

0.9985

sequence A3

0.9990

sequence A3

0.9985

sequence A1

sequence A1

0.9980

0

2

4

6

8

0.9980

10

0

2

4

Time(h)

6

8

10

Time(h)

1.000

1.0000 88wt% ethanol 88wt% ethanol

0.998

xB2(molefrac of water)

xB1(molefrac of ethanol)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

sequence A3 sequence A1 0.996

0.994

0.992

0

2

4

6

8

10

0.9995

0.9990 sequence A3 sequence A1

0.9985

0.9980

0

2

Time(h)

4

6 Time(h)

ACS Paragon Plus Environment

8

10

Industrial & Engineering Chemistry Research

372

374.0

370

373.5

92wt% ethanol

T17(K)

T22/T26(K)

92wt% ethanol

368

373.0 sequence A3(T17)

sequence A3(T22) sequence A1(T26)

366

364

0

2

4

6

8

sequence A1(T17)

372.5

372.0

10

0

2

4

Time(h)

6

8

10

Time(h)

372

374.0 88wt% ethanol 88wt% ethanol

373.5 370 T17(K)

sequence A3(T22)

T22/T26(K)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 37

sequence A1(T62)

373.0 sequence A3(T17)

368

sequence A1(T17)

372.5

366

0

2

4

6

8

372.0

10

0

2

Time(h)

4

6

8

10

Time(h)

Figure 19. Dynamic responses to feed composition disturbances for sequences ADWC and A1 using structure TCS2.

3.4. Comparison of Dynamic Performance of ADWC and Sequence A2 Using Temperature Control Structure TCS2. The dynamic performance of sequence A2 is also implemented using temperature controller structure TCS2 to compare with ADWC. Figure 20 shows the improved temperature controller structure TCS2 and controller faceplate of sequence A2.

ACS Paragon Plus Environment

Page 29 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

Figure 20. Overall control strategy of sequence A2 using control structure TCS2.

Figure 21 gives the dynamic responses of sequence A2 and ADWC using temperature control structure TCS2 for feed flowrate disturbance. Both sequence A2 and ADWC can handle the disturbances well. A noteworthy difference between ADWC and sequence A2 is that the ethanol product purity of ADWC under +20% feed flowrate disturbance undergoes a larger transient

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research

deviation than sequence A2. However, this is opposite under -20% feed flowrate disturbance. Figure 22 gives the dynamic responses for feed composition disturbances. It is obvious that sequence A2 isn’t able to hold ethanol product purity with a large negative offset (3.2 wt%) under -2 wt% ethanol disturbance, although the dynamic performances of all other variables are similar. This means that snowballing effect happens for sequence A2 under -2 wt% ethanol disturbance. The most probably reason is that the lack of reflux of stripper column makes the ethanol concentration of recycle flow unable to maintain at desired value. 1.0000

1.0000 +20% feed flowrate

+20% feed flowrate xB2(molefrac of water)

xB1(molefrac of ethanol)

0.9995

0.9990

0.9985 sequence A3 0.9980

sequence A2

0.9995

0.9990

sequence A3 0.9985

sequence A2

0.9975

0.9970

0

2

4

6

8

0.9980

10

0

2

4

Time(h)

6

8

10

Time(h)

1.000

1.0000 -20% feed flowrate

xB2(molefrac of water)

0.999 xB1(molefrac of ethanol)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 37

0.998

0.997

sequence A3 sequence A2

0.996

-20% feed flowrate

0.9995

0.9990 sequence A3 sequence A2

0.9985

0.995

0.994

0

2

4

6

8

10

0.9980

0

2

Time(h)

4

6 Time(h)

ACS Paragon Plus Environment

8

10

Page 31 of 37

376

374.0 +20% feed flowrate +20% feed flowrate

374 373.5

sequence A3(T22) 372 T17(K)

T22/T35(K)

sequence A2(T35)

370

373.0

sequence A3(T17)

372.5

sequence A2(T17)

368

366

0

2

4

6

8

372.0

10

0

2

4

Time(h)

-20% feed flowrate

10

-20% feed flowrate 374

T17(K)

370

373

sequence A3(T17)

sequence A3(T22)

365

372

sequence A2(T17)

sequence A2(T35)

360

8

375

375 T22/T35(K)

6 Time(h)

380

0

2

4

6

8

371

10

0

2

4

Time(h)

6

8

10

Time(h)

Figure 21. Dynamic responses to feed flow rate disturbances for sequences ADWC and A2 using structure TCS2.

1.0000

1.0000 92%wt ethanol 92wt% ethanol

0.9995

xB2(molefrac of water)

xB1(molefrac of ethanol)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

0.9990

sequence A3 0.9985

0.9980

sequence A2

0

2

4

6

8

10

0.9995

0.9990

sequence A3 0.9985

0.9980

sequence A2

0

2

Time(h)

4

6 Time(h)

ACS Paragon Plus Environment

8

10

Industrial & Engineering Chemistry Research

1.00

1.0000 88wt% ethanol xB2(molefrac of water)

xB1(molefrac of ethanol)

0.99

0.98 88wt% ethanol 0.97

0.96 sequence A3

0.94

0.9995

0.9990 sequence A3 sequence A2

0.9985

sequence A2

0.95

0

2

4

6

8

0.9980

10

0

2

4

Time(h)

6

8

10

Time(h)

372

374.0 92wt % ethanol 92wt% ethanol 373.5

T17(K)

T22/T35(K)

370

368

373.0 sequence A3(T17)

sequence A3(T22) 366

364

0

2

4

6

8

sequence A2(T17)

372.5

sequence A2(T35)

372.0

10

0

2

4

Time(h)

6

8

10

Time(h)

372

374.0 88wt% ethanol

88wt% ethanol

371 373.5 370 sequence A3(T22) 369

T17(K)

T22/T35(K)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 32 of 37

sequence A2(T35)

373.0 sequence A3(T17)

368

sequence A2(T17)

372.5 367

366

0

2

4

6

8

372.0

10

0

2

Time(h)

4

6

8

10

Time(h)

Figure 22. Dynamic responses to feed composition disturbances for sequences ADWC and A2 using structure TCS2.

4. CONCLUSION In this article, the heterogeneous azeotropic dividing-wall column (ADWC) is proposed by demonstrating ethanol dehydration. Overall assessment of ADWC is implemented by comparing the optimal design and dynamic controllability with two conventional sequences. Significant energy

ACS Paragon Plus Environment

Page 33 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

saving of 21.36% and 11.97%, total annual cost of 18.48% and 18.43% can be obtained compared with conventional azeotropic/recovery sequence and conventional azeotropic/stripper sequence, respectively. Conventional temperature control structure is applied to investigate the dynamic performances of ADWC compared with two conventional sequences. As two reboiler duties are still preserved

in

ADWC,

it

has

comparable

dynamic

controllability

with

conventional

azeotropic/recovery column sequence and superior dynamic controllability with conventional azeotropic/stripper column sequence. Therefore, the ADWC has the benefits to be able to apply to azeotropic distillation.

Supporting Information (1) Binary interaction parameters for the Aspen Plus NRTL-RK model; (2) Comparison between predicted and experimental azeotropic compositions and bubble points at 101.3 kPa; (3) The details of calculating TAC. REFERENCES

(1) Dejanović, I.; Matijašević, Lj.; Olujić, Ž. Dividing Wall Column−A Breakthrough towards Sustainable Distilling. Chem. Eng. Process. 2010, 49, 559. (2) Yildirim, Ö.; Kiss, A. A.; Kenig, Y. E. Dividing Wall Columns in Chemical Process: A Review on Current Activities. Sep. Purif. Technol. 2011, 80, 403. (3) Kiss, A. A.; Bildea, S. C. A Control Perspective on Process Intensification in Dividing-Wall Columns. Chem. Eng. Process. 2011, 50, 281. (4) Xia, M.; Yu, B. Y.; Wang, Q. Y.; Jiao, H. P.; Xu, C. J. Design and Control of Extractive Dividing-Wall Column for Separating Methylal−Methanol Mixture. Ind. Eng. Chem. Res. 2012, 51,

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

16016. (5) Xia, M.; Xin, Y. P.; Luo, J. W.; Li, W. S.; Shi, L.; Min, Y.; Xu, C. J. Temperature Control for Extractive Dividing-Wall Column with an Adjustable Vapor Split: Methylal/Methanol Azeotrope Separation. Ind. Eng. Chem. Res. 2013, 52, 17996. (6) Zhang, H.; Ye, Q.; Qin, J. W.; Xu, H.; Li, N. Design and Control of Extractive Dividing-Wall Column for Separating Ethyl Acetate/Isopropyl Alcohol Mixture. Ind. Eng. Chem. Res. 2014, 53, 1189. (7) Sun, L. Y.; Wang, Q. Y.; Li, L. M.; Zhai, J.; Liu, Y. L. Design and Control of Extractive Dividing Wall Column for Separating Benzene/Cyclohexane Mixtures. Ind. Eng. Chem. Res. 2014, 53, 8120. (8) Tututi-Avila, S.; Jiménez-Gutiérrez, A.; Hahn, J. Control analysis of an extractive dividing-wall column used for ethanol dehydration. Chem. Eng. Process. 2014, 82, 88. (9) Bravo-Bravo, C.; Segovia-Hernández, J. G.; Gutiérrez-Antonio, C.; Durán, A. L.; Bonilla-Petriciolet, A.; Briones-Ramírez, A. Extractive Dividing Wall Column: Design and Optimization. Ind. Eng. Chem. Res. 2010, 49, 3672. (10) An, D. H.; Cai, W. F.; Xia, M.; Zhang, X. B.; Wang, F. M. Design and control of reactive dividing-wall column for the production of methyl acetate. Chem. Eng. Process. 2015, 92, 45. (11) Ignata, R. M.; Kiss, A. A. Optimal design, dynamics and control of a reactive DWC for biodiesel production. Chem. Eng. Res. Des. 2013, 91, 1760. (12) Wang, S. J.; Cheng, S. H.; Chiu, P. H.; Huang, K. J. Design and Control of a Thermally Coupled Reactive Distillation Process Synthesizing Diethyl Carbonate. Ind. Eng. Chem. Res. 2014, 53, 5982. (13) Dai, X.; Ye, Q.; Yu, H.; Suo, X. M.; Li, R. Design and Control of Dividing-Wall Column for the Synthesis of n-Propyl Propionate by Reactive Distillation. Ind. Eng. Chem. Res. 2015, 54, 3919.

ACS Paragon Plus Environment

Page 34 of 37

Page 35 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

(14) Sun, L. Y.; Chang, X. W.; Qi, C. X.; Li, Q. S. Implementation of Ethanol Dehydration Using Dividing-Wall Heterogeneous Azeotropic Distillation Column. Sep. Sci. Technol. 2011, 46, 1365. (15) Kiss, A.A.; Suszwalak, D. J. Enhanced bioethanol dehydration by extractive and azeotropic distillation in dividing-wall columns. Sep. Purif. Technol. 2012, 86, 70. (16) Le, Q. K.; Halvorsen, I. J.; Pajalic, O.; Skogestad, S. Dividing wall columns for heterogeneous azeotropic distillation. Chem. Eng. Res. Des. 2015, 99, 111. (17) Wu, Y. C.; Lee, H. Y.; Huang, H. P.; Chien, I. L. Energy-Saving Dividing-Wall Column Design and Control for Heterogeneous Azeotropic Distillation Systems. Ind. Eng. Chem. Res. 2014, 53, 1537. (18) Wu, Y. C.; Huang, H. P.; Chien, I. L. Investigation of the Energy-Saving Design of an Industrial 1,4-Dioxane Dehydration Process with Light Feed Impurity. Ind. Eng. Chem. Res. 2014, 53, 15667. (19) Yu, H.; Ye, Q.; Xu, H.; Zhang, H.; Dai, X. Design and Control of Dividing-Wall Column for tert-Butanol Dehydration System via Heterogeneous Azeotropic Distillation. Ind. Eng. Chem. Res. 2015, 54, 3384. (20) Ryan. P. J.; Doherty, M. F. Design/Optimization of Ternary Heterogeneous Azeotropic Distillation Sequences. AIChE J. 1989, 35, 1592. (21) Luyben, L. W. Economic Optimum Design of the Heterogeneous Azeotropic Dehydration of Ethanol. Ind. Eng. Chem. Res. 2012, 51, 16427. (22) Chien, I. L.; Zeng, K. L.; Chao, H. Y. Design and Control of a Complete Heterogeneous Azeotropic Distillation Column System. Ind. Eng. Chem. Res. 2004, 43, 2160. (23) Arifin, S.; Chien, I. L. Combined Preconcentrator/Recovery Column Design for Isopropyl Alcohol Dehydration Process. Ind. Eng. Chem. Res. 2007, 46, 2535.

ACS Paragon Plus Environment

Industrial & Engineering Chemistry Research

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(24) Wu, Y. C.; Chien, I. L. Design and Control of Heterogeneous Azeotropic Column System for the Separation of Pyridine and Water. Ind. Eng. Chem. Res. 2009, 48, 10564. (25) Douglas, J. M. Conceptual Design of Chemical Processes; McGraw Hill: New York, 1998. (26) Luyben, W. L. Distillation Design and Control Using Aspen Simulation; John Wiley & Sons: New York, 2006.

ACS Paragon Plus Environment

Page 36 of 37

Page 37 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

For Table of Contents Only

ACS Paragon Plus Environment