Article

pubs.acs.org/IECR

Control of Heat Integrated Pressure-Swing-Distillation Process for

Separating Azeotropic Mixture of Tetrahydrofuran and Methanol
Yinglong Wang,* Zhen Zhang, Huan Zhang, and Qing Zhang
College of Chemical Engineering, Qingdao University of Science and Technology, Qingdao 266042, China
S Supporting Information
*

ABSTRACT: Dynamic control of a pressure-swing-distillation process for separation of azeotropic mixture of tetrahydrofuran
and methanol is explored. The pressure-swing-distillation processes involved with no, partial, and full heat integration are
simulated using Aspen Plus Dynamics. The inuences of the selection of the sensitive temperature stage in the low-pressure
column on the dynamic responses in the pressure-swing distillation with dierent heat integration were investigated. The results
indicate that a suitable temperature control stage in the low-pressure column is crucial to achieve ecient control of the process.
In addition, more time is needed to reach the quality specications under feed disturbances for both components when heat
integration is added in the distillation sequence.

1. INTRODUCTION
The separation of azeotropic mixtures is an interesting and
important topic for academic research and industrial
application.13 The binary mixture of tetrahydrofuran (THF)
and methanol exhibits a minimum boiling azeotrope because of
molecular repulsion between the two chemical components,
thus leading to diculty in obtaining high-purity products. In
our previous work, the separation of this mixture using
pressure-swing distillation (PSD) was reported using steadystate simulation.4 However, dynamic control of the PSD was
not involved in that study. It is of great importance to study the
dynamic control process, which can help maintain the purity of
the products in the face of disturbances in feed ow rate and
feed composition.
Dynamic control schemes for separating azeotropic mixtures
have been studied in recent years.518 Skogestad5 proposed a
systemic and model-based method for the control structure
design. Luyben13,15 studied the separation of azeotropic
mixtures in detail from steady-stage design to dynamic control.
Since the controller design that based on heuristic methods
draws the attention of most researchers, inuences of control
variables on dynamic performances of control schemes are
studied widely.1928 For example, Wang22 investigated the
eect of entrainer loss on plantwide control of an isopropanol
dehydration process. Jones27,28 focused on the selection of the
primary and secondary control variables in the plantwide
control system design. More importantly, all the studies on
control structure and the selection of control variables can
promote the applications of dynamic control in the chemical
processes.17,29,30 For instance, Luyben29 showed the control
schemes for the multiunit heterogeneous azeotropic distillation
for separating the mixture of ethanol and water with benzene as
the entrainer. Wei17 gave an ecient control structure of the
PSD for separating the mixture of dimethyl carbonate and
methanol with the large feed ow and feed composition
disturbances. In addition, the dynamic control of the divided
wall column attracts great attention due to its investment and
2015 American Chemical Society

energy savings compared with traditional distillation

schemes.3135
Design, modeling, and optimization of the PSD processes
have been studied recently.36,37 Some researchers also
concentrated on the dynamic controllability of the PSD
process.13,17,38,39 Luyben13 demonstrated the basic control
structure of the PSD process for separating the acetone/
methanol azeotropic mixture in detail. The possibilities of the
heat integration for the PSD process, thermally coupled
distillation sequences, and dividing wall distillation make
them more competitive than traditional distillation.4044
Among them, in addition to energy and capital saving, the
PSD process has a prominent merit that no third component is
introduced in the distillation system for separating azeotropic
mixtures. Thus, the dynamic control of the PSD with heat
integration for separating the azeotrope deserves further
study.3739,45 Li37 gave a composition/temperature cascade
control structure for the new PSD with partial heat integration,
which can handle the feed ow and feed composition
disturbances perfectly. Yu39 presented various control structures for the PSD with full heat integration when the mixture of
methylal and methanol was used. Also, the comparison between
the PSD and extractive distillation is reported according to the
steady-stage and/or dynamic control performance.36,46
The temperature controller detects the signal from the
control stage of a column and then transmits it to the
manipulated variable. The selection of the temperature control
stage based on the slope criterion was advised by Luyben.15
The slope criterion is feasible for distillation column with a
large change in temperature from stage to stage. As for the
distillation column that the temperature prole shows no
obvious temperature change from stage to stage, Luyben
suggests choosing the stage according to the sensitivity
Received:
Revised:
Accepted:
Published:
1646

November 6, 2014
January 9, 2015
January 19, 2015
January 19, 2015
DOI: 10.1021/ie505024q
Ind. Eng. Chem. Res. 2015, 54, 16461655

Article

Industrial & Engineering Chemistry Research

Figure 1. Temperature proles and temperature slope value plots of two columns.

Figure 2. Basic temperature control structure for the PSD process without heat integration.

criterion.15 It is known that temperatures on all stages can be

quickly aected by manipulating heat input. Luyben46
investigated the eect of the two alternative control stages in
extractive column on the tight control of an extractive
distillation system for separating CO2 and ethane in the
enhanced oil recovery processes, suggesting that composition
control is required.

In this work, the PSD processes containing a low-pressure

column (LPC) and a high-pressure column (HPC) are used for
separation of azeotropic mixture of tetrahydrofuran and
methanol. Two alternative control stages are selected in the
LPC, respectively, while one stage is chosen as the control stage
in the HPC. We develop eective control structures of the PSD
without and with heat integration for investigating how the two
1647

DOI: 10.1021/ie505024q
Ind. Eng. Chem. Res. 2015, 54, 16461655

Article

Industrial & Engineering Chemistry Research

Table 1. Controller Tuning Parameters for Dierent PSD Processes
no integration
controlled variable
manipulated variable
transmitter range
controller output range
ultimate gain
ultimate period
gain Kc
integral time I
controlled variable
manipulated variable
transmitter range
controller output range
ultimate gain
ultimate period
gain Kc
integral time I

TC1
T1,33
QR1
376.2476.2 K
010.951 GJ/h
2.27
5.40 min
0.71
11.88 min
TC2
T2,18
QR2
286.2386.2 K
04.489 GJ/h
32.01
3.00 min
10.01
6.60 min

partial integration

full integration

T1,33
QR1
376.2476.2 K
010.951 GJ/h
2.76
4.80 min
0.86
10.56 min

T1,33
QR1
376.2476.2 K
010.951 GJ/h
3.72
4.80 min
1.16
10.56 min

T2,18
RR2
286.2386.2 K
05.980
58.36
21.00 min
18.24
46.20 min

T2,31
RR2
290.2390.2 K
06.817
21.84
10.20 min
6.82
22.44 min

(1) Feed is ow-controlled (reverse acting).

(2) The pressure in each column is controlled by
manipulating the heat removal of condenser (reverse
acting).
(3) Reux drum levels in both columns are held by
manipulating distillate ow (direct acting).
(4) Sump levels in both columns are held by manipulating
bottom ow (direct acting).
(5) The temperature on stage 33 in the HPC is controlled by
manipulating the reboiler heat input in the HPC (reverse
acting).
(6) The temperature on stage 18 in the LPC is controlled by
manipulating the reboiler heat input in the LPC (reverse
acting).
In the control scheme, reux ratios are 1.79 for the HPC and
2.99 for the LPC, which are held constantly in each column.
Relay-feedback tests are run on the two temperature controllers
to determine ultimate gains and periods, and the Tyreus
Luyben tuning rule is used in two temperature controllers.
Table 1 lists the temperature transmitter ranges, controller
output ranges, and tuning parameters. To further model the
other dynamics in the control system, a deadtime of 1 min is set
for each temperature controller. All level controllers are
proportional with gain (Kc) of 2.
Figure 3 shows the dynamic responses for the basic control
structure of the PSD without heat integration. It is observed
that when new stable regulatory controls are achieved for
positive 20% (solid lines) and negative 20% (dashed lines) ow
rate disturbances at 1 h in feed stream, the purities of THF in
bottom stream of HPC (B1) return to 99.87 and 99.92 mol %
(Figure 3a), respectively. However, the purity of methanol in
the bottom stream of the LPC (B2) has a large deviation from
the desired value of 99.90 mol % when the control structure
encounters positive 20% feed ow rate disturbance. The
responses as feed composition disturbances from 25 to 30 mol
% methanol (solid lines) and from 25 to 20 mol % methanol
(dashed lines) are also simulated and similarly the obtained
results are for the purity of THF and methanol as 20% step
changes in feed composition disturbance at 1 h (Figure 3b).
Composition/Temperature Cascade Control Structure.
Temperature control has the advantage of being fast, but it

control stages in the LPC aect the dynamic performance of

the PSD process for separation of THF and methanol.

2. PROCESS STUDY AND SELECTION OF

TEMPERATURE CONTROL STAGE
In our previous work, the optimized owsheet processes for
separation of azeotropic mixture containing 75 mol % THF and
25 mol % methanol were available via the steady-state PSD with
and without heat integration.4 The pressure dierence between
two adjacent stages is assumed to 0.0068 atm in the LPC and
HPC. Furthermore, an auxiliary condenser in the PSD process
is also integrated into the process for the purpose of partial heat
integration.
It is of great importance to select the temperature control
stage in both columns. Figure 1 shows temperature proles and
corresponding slope values as functions of column stages for
the HPC and LPC. It can be seen that the HPC could be
eectively controlled because its temperature prole exhibits a
fairly sharp temperature break; however, the shape of the
temperature prole has no breaks for the LPC and the slope
values of the temperature prole stay almost the same in each
stage. According to the slope criterion suggested by Luyben,15
stage 33 in the HPC is selected as a control stage to keep its
temperature constant, while stage 18 or 31 in the LPC can be
selected as a control stage in the temperature control loop
structure.
3. CONTROL STRUCTURES FOR THE PSD PROCESSES
The volumes of reux drums and sumps are specied to
provide holdup for 5 min when the vessels are lled with 50%
using the heuristic method.15 The ratios of height to diameter
of reux drums and sumps are both set as 2. The tray-sizing
option in Aspen Plus is selected to calculate the diameters of
two columns. Pumps and valves are sized to give adequate
pressure drops to handle changes in ow rates.
3.1. Control of the PSD without Heat Integration.
Basic Temperature Control Structure. Figure 2 shows the
basic temperature control structure for separating binary
mixture of THF and methanol without heat integration. It is
noticed that nine controllers in this scheme are included as
follows.
1648

DOI: 10.1021/ie505024q
Ind. Eng. Chem. Res. 2015, 54, 16461655

Article

Industrial & Engineering Chemistry Research

First, stage 18 in the LPC is selected as a temperature control

stage as the same as that in the basic control structure (Figure
4). Table 2 lists the tuning parameters of the composition
Table 2. Tuning Parameters of the Composition Controller
with Dierent Deadtimes for the PSD Process without Heat
Integration
deadtime
ultimate gain
ultimate period
gain Kc
integral time I

2 min

3 min

4 min

889.80
6.00 min
278.06
13.2 min

345.98
11.40 min
108.12
25.08 min

70.18
22.20 min
21.93
48.84 min

controller with dierent deadtimes for the PSD process. It can

be seen that the integral time is positively correlated with the
deadtime of the composition controller. The shorter the
deadtime of the controller, the less time the control scheme
needs to reach a new stead state from the disturbed state. In
reality, however, the detectors have specic measurement
deadtimes, and the measurement of the composition controller
has larger deadtime than the temperature controller. The
deadtime of the composition controller is set at 3 min, which is
consistent with most published papers.10,46,47 Notice that the
controller TC2 is on cascade with its set point from the
temperature of stage 18 in the LPC, and the composition
controller CC is tuned with TC2 on cascade. The output
signal from the controller TC1 is the heat duty of reboiler in
the bottom of the LPC, and the output signal from the
controller TC2 is the ratio of heat input of the reboiler in the
distillation of the LPC to the feed ow rate QR2/F. Figure 5
gives the dynamic response results for the composition/
temperature cascade control structure for the PSD without heat
integration. The solid lines stand for the increases in feed ow
rate, and the methanol content in the feed is increased from 25

Figure 3. Dynamic responses for the basic control structure for the
PSD without heat integration: feed ow rate and feed composition
disturbances.

may not keep the product purity constant. Composition control

is slow, but it will drive product purity to the desired value. To
integrate the advantageous features of both, a cascade
combination of composition and temperature control structure
is developed according to the performance of the basic control
structure. As illustrated in section 2, both stage 18 and 31 in the
LPC are possible temperature control stages, where the
temperature can be controlled by manipulating the heat duty
of reboiler in the LPC. In this section, we will discuss the
inuence of the temperature control stage in the LPC on
cascade control dynamic performance.

Figure 4. Composition/temperature cascade control structure for the PSD process without heat integration.
1649

DOI: 10.1021/ie505024q
Ind. Eng. Chem. Res. 2015, 54, 16461655

Article

Industrial & Engineering Chemistry Research

to 30 mol % methanol. The dashed lines represent the decrease

in feed ow rate and the methanol content in the feed
decreased from 25 to 20 mol %. It is noticed that decreasing the
methanol in the feed produces less distillates (D1) and more
bottoms (B1) in the LPC, and temperature decreases and
reboiler duty increases in the LPC. Both of the two columns are
well-controlled.
Then, Figure S1 shows the cascade control structure for the
PSD process with stage 31 in the LPC as the temperature
control stage. The temperature controller TC1 and the
composition controller CC are retuned. Figure S2 gives the
responses of the control structure encountering disturbances in
feed ow rate and feed composition as the above. Both product
purities are held close to their specications. It is noticed that
the temperature of stage 18 in the LPC is more closely
controlled than that on stage 31. The reason for this is that the
transmitted temperature signal from stage 18 is quicker than
that from 31 because the stream is fed at stage 20 in the LPC.
Thus, stage 18 in the LPC is selected as the control stage in the
cascade control scheme of the PSD without heat integration.
3.2. Control of the PSD with Partial Heat Integration.
Basic Temperature Control Structure. As for the PSD process
with partial heat integration, the heat input of the LPC is
completely provided by the HPC with combined reboiler/
condenser. The excessive heat of the HPC is removed by an
auxiliary condenser added in the process. The area of 16.67 m2
for the combined reboiler/condenser is calculated according to
the steady-state simulation. The heat duties of HPC condenser
and the LPC reboiler are 979.37 and 586.09 kW, respectively.
The heat removal rate of the auxiliary condenser is equal to
393.28 kW, and the negative sign of this value is specied as the
initial output signal of the PC1 controller. The overall heat
transfer coeciency is specied as 0.002 044 8 GJ/(h m2 K).
Figure S3 gives the basic temperature control structure and the
controller faceplates. In this control structure, the ow
controller has the normal setting of 0.50 for Kc and 0.30 min

Figure 5. Dynamic responses for the composition/temperature

cascade control structure for the PSD without heat integration: feed
ow rate and feed composition disturbances.

Figure 6. Improved composition/temperature cascade control structure for the PSD process with partial heat integration.
1650

DOI: 10.1021/ie505024q
Ind. Eng. Chem. Res. 2015, 54, 16461655

Article

Industrial & Engineering Chemistry Research

for integral time (I) and the level controllers have a Kc value of
2. After running relay-feedback tests on the two temperature
controllers, the ultimate gains and the periods on two
temperature controllers are determined and listed in Table 1.
The owsheet equations function is used to achieve the
partial heat integration (Figure S4a). The rst equation is used
to calculate the reboiler duty of LPC, while the heat removal
rate of HPC is calculated by the second equation.
Figure S5 gives the dynamic responses of the basic control
structure for the PSD with partial heat integration. The control
structure is tested by making fresh feed ow rate disturbances
(Figure S5a) and feed composition disturbances at 1 h (Figure
S5b). It can be seen that the THF in B1 stream can keep almost
the desired purity when the control loop arrives at a new steady
state after encountering 20% changes in feed ow rate and
feed methanol. The purity of methanol in B2 stream has a large
deviation from the initial value when the feed ow rate varies
from 100 to 120 kmol/h, and the methanol content in the feed
is changed from 25 to 30 mol %.
Improved Composition/Temperature Cascade Control
Structure. The basic temperature control structure could not
achieve eective disturbance rejection for feed ow rate and
feed composition disturbances. The reason for this is that the
almost constant heat input of the LPC could not provide
enough energy to produce sucient vapor up to the LPC when
the disturbances occur. Thus, the low purity of methanol in B1
was discharged. To solve this problem, the multiplier QR2/F
and the composition controller are added to provide the heat
for better dynamic performance based on the basic temperature
control structure (Figure S5). The inuence of temperature
control stage in the LPC on the control structure of the PSD
with partial heat integration will be investigated in this section.
First, Figure 6 shows the improved cascade control structure
with the detective temperature signal of the LPC from stage 18.
Notice that only one equation is used in the owsheet
equation (Figure S4b) because the ratio of QR2/F is controlled
by TC2. Figure 7 gives the performance of the cascade control
structure when positive (solid lines) and negative (dashed
lines) 20% disturbances in the feed ow rate and feed methanol
are encountered at 1 h. It can be seen that the purity of THF in
B1 stream is brought back to 99.92 and 99.87 mol % when the
methanol content in the feed stream varies from 25 to 30 mol
% and from 25 to 20 mol %, respectively, while the methanol in
B2 stream can be kept constant after 12 h using the cascade
control structure for the feed ow and feed composition
disturbances.
Second, Figure S6 shows the cascade control structure for
the PSD process with partial heat integration when stage 31 is
selected as the temperature control stage in the LPC. Figure S7
gives results for feed ow and feed composition disturbances. It
is noticed that the purity of methanol in B2 stream can return to
the desired value of 99.90 mol % after transient deviations when
the feed ow and feed composition disturbances occurred at 1
h. However, the THF in B1 has a fairly large deviation after 4 h
when the feed methanol varies from 25 to 20 mol %. It can be
concluded that the composition/temperature cascade control
structure with stage 18 or 31 in the LPC as the temperature
control stage can handle the 20% disturbances in the feed
ow rate and feed composition very well. The dierence
between the two cascade control structures is the way of
variation of the methanol in the B2 stream. So stages 18 and 31
of the LPC can be selected as control stages in the control loop
of the PSD with partial heat integration.

Figure 7. Dynamic responses for the improved composition/

temperature cascade control structure for the PSD with partial heat
integration: feed ow rate and feed composition disturbances.

3.3. Control of the PSD with Full Heat Integration.

Basic Control Structure. The large temperature dierence
between the condenser of the HPC and the reboiler of the LPC
makes the full heat integration possible for the PSD process. As
for the fully heat-integrated process, the only heat input that is
required is from the heat duty QR1 of the HPC reboiler because
the heat duty of the LPC reboiler is provided by the combined
condenser/reboiler. The result of overall heat transfer
coeciency (0.002 044 8 GJ/(h m2 K)) multiplied by the
heat transfer area and the temperature dierence between the
condenser of the HPC and the reboiler of the LPC equal to the
heat duty of combined condenser/reboiler. The owsheet
equations are employed to achieve the full heat integration in
Aspen Plus Dynamics. In Figure S4c, the rst equation was
used to calculate the heat duty input of the LPC, and the
second equation was used to make the equivalence of the heat
input rate of the LPC reboiler and heat removal rate of the
HPC condenser. The basic control structure is given in Figure
S8. After running the relay-feedback tests on two temperature
controllers, the calculated gains and periods by TyreusLuyben
tuning are 1.06 and 10.56 min, respectively. Figure S9 gives the
eects of feed ow rate and feed composition disturbances on
the control structure. Both of the temperatures on stage 33 in
the HPC and on stage 18 in the LPC are controlled very well
by manipulating the reboiler duty of HPC and reex ratio
(RR2) of LPC. The product purities in two bottom streams are
1651

DOI: 10.1021/ie505024q
Ind. Eng. Chem. Res. 2015, 54, 16461655

Article

Industrial & Engineering Chemistry Research

Figure 8. Improved pressurecompensated temperature control structure for the PSD with full heat integration when stage 31 in the LPC is selected
as the control stage.

parameters are listed in Table 1. The control structure and

control panel of all controllers are shown in Figure S13. Feed
ow rate and feed composition disturbances are added to test
the performances of the control scheme (Figure S14). This
pressure-compensated temperature control structure with stage
31 in the LPC as the temperature control stage shows the same
variation tendencies for the purities of two products and
temperature when stage 18 is selected as the control stage in
the LPC.
Improved Pressure-Compensated Temperature Control.
The improved pressure-compensated temperature control
structure is established based on the above pressurecompensated temperature control structure (Figure S11). A
composition controller that detects the methanol in the bottom
stream of the LPC and a deadtime element with deadtime of
3 min are added in the control loop. Figures S15 and 8 give the
improved pressurecomposition control structure with the
detected temperatures of stages 18 and 31 in the LPC,
respectively. The performances of the control structure are
tested by introducing the feed ow rate and feed composition
disturbances at 1 h (Figures S16 and 9). It can be seen that the
temperatures on two control stages of two columns deviate
from the initial values. The purity of THF in the bottom stream
of the HPC returns to close proximity of its setting point.
Methanol in the bottom stream of the LPC bounces back to its
steady-state value. It can be found from Figure 9 that the
improved pressure-compensated temperature control structure
can deal with the feed ow rate and composition disturbances
very well when stage 31 in the LPC is selected as the control
stage. The reboiler duties of HPC changed directly after
disturbances occurred, while the temperature on stage 31 of the
LPC controlled by manipulating the reex ratio (RR2) arrives at
a new steady-state value gradually. It is also observed that the
THF in B1 stream in the control loop with stage 31 in the LPC
as the control stage returns to the constant values in 3 h,
which is faster than the control loop when stage 18 in the LPC
is selected as the control stage. The purity of methanol in the

not maintained at their specications (99.90 mol %) when the

positive disturbances of feed ow rate and feed composition of
methanol are encountered. An increase of the feed ow rate
results in a temperature drop in the stripping section of the
LPC, leading to more THF escaping from the bottom.
Pressure-Compensated Temperature Control. Pressurecompensated temperature control has been described by
Buckley.48 Before setting up pressure-compensated temperature
control for the PSD with full heat integration, we investigate
the bubble point temperatures of the liquids on stage 33 in the
HPC at the pressure of 911 atm. Figure S10 shows the bubble
point temperature of the mixture as a function of pressure, and
a corresponding line is tted with a slope value of 5.016.
The pressure-compensated temperature is calculated by the
following equation, TPC = T1,33 (P 10.1325), where the
temperature is measured from stage 33 and the pressure is
obtained from the top of the HPC. The TPC from deadtime
element dead1 is sent to the temperature controller TC1. The
owsheet equation is applied to carry out the pressurecompensated temperature calculation (Figure S4d).39 In the
following text, we will investigate the eect of the control stage
in the LPC on the pressure-compensated temperature control
structure.
First, stage 18 in the LPC is selected as the control stage
(Figure S11) in the pressure-compensated temperature
structure, and corresponding dynamic responses of this control
structure on the 20% changes in feed ow rate and feed
composition are shown in Figure S12. It can be seen that the
pressure-compensated temperature control structure could not
handle the feed ow rate and feed composition disturbances
because the methanol in B2 stream changes from 99.90 to 93.50
mol % when the feed ow rate varies from 100 to 120 mol % at
1 h.
Second, the temperature of stage 31 in LPC is measured and
controlled. The controller is tuned by running relay-feedback
tests and using TyreusLuyben tuning rules. The temperature
transmitter ranges, controller output ranges, and tuning
1652

DOI: 10.1021/ie505024q
Ind. Eng. Chem. Res. 2015, 54, 16461655

Article

Industrial & Engineering Chemistry Research

fairly sharp temperature break; however, the shape of the

temperature prole has no breaks for the LPC and the slope
values of the temperature prole stay almost the same in each
stage. Therefore, the selection of the temperature control stage
in the LPC is crucial to eciently achieve dynamic control of
the process. The inuences of alternative temperature control
stages 18 and 31 in the LPC on the control structures are
investigated when stage 33 is used as the control stage in the
HPC. As for the PSD process without heat integration, the
dynamic control loop with stage 18 used as the control stage in
the LPC can handle the disturbances of feed ow rate and feed
composition more eectively. To the PSD with partial heat
integration, the selection of control stage in LPC has little
inuence on the nal ecient control structure from the
controllability viewpoint. For the PSD with full heat
integration, it shows a better control performance with stage
31 of the LPC as the control stage the feed ow and feed
composition disturbances. On the basis of the steady-state
simulation, the PSD with heat integrated sequence can save
more energy than the conventional PSD process. In order to
handle the disturbances eectively, the heat integrated
distillation process should be added with the complicated
control structure due to its degrees of freedom. This work will
facilitate the selection of temperature control stages in PSD
processes without and with heat integration for separation of
other azeotropic systems, especially for the systems whose
temperature proles have two large temperature changes from
stage to stage.

ASSOCIATED CONTENT

S Supporting Information
*

Dynamic simulation details and additional gures (Figures S1

S16). This material is available free of charge via the Internet at
http://pubs.acs.org.

Figure 9. Dynamic responses for the improved pressurecompensated

temperature control structure for the PSD with full heat integration
when stage 31 in the LPC is selected as the control stage: feed ow
rate and feed composition disturbances.

B2 stream changes gently (Figure 9). Therefore, stage 31 as the

temperature control stage in the LPC performs better than
stage 18 in the PSD with full heat integration.
3.4. Discussion of Alternative Temperature Control
Stages in LPC. Stages 18 and 31 in the LPC can both act as
temperature control stages according to the slope criterion, and
it is dynamically feasible for them pairing with heat input. In the
LPC, stage 18 is located in the rectifying section, while stage 31
lies in the stripping section. On one hand, the temperature
signal from stage 18 to the temperature controller shows faster
response than that from stage 31 when the disturbances in ow
rate and composition occurred, because the feed stream is fed
at stage 20 in the LPC. On the other hand, the temperature of
stage 31 would vary quicker than that of stage 18 by
manipulating the reboiler duty. To combine the advantages of
these two alternative temperature control stages in the LPC, the
selection of the temperature control stages should be
considered in the PSD with dierent heat integration.

AUTHOR INFORMATION

Corresponding Author

*E-mail: yinglongw@126.com.
Notes

The authors declare no competing nancial interest.

ACKNOWLEDGMENTS
Comments and suggestions from two anonymous reviewers
and Professor William L. Luyben are gratefully acknowledged.
Financial support from National Natural Science Foundation of
China (Project 21306093) is gratefully acknowledged.

4. CONCLUSIONS
In this work, dynamic controls of PSD processes without and
with heat integration for separation of an azeotropic mixture of
75 mol % THF and 25 mol % methanol are simulated using
Aspen Plus Dynamics. The results show that the HPC can be
eectively controlled because its temperature prole exhibits a
1653

NOTATION
HPC = high pressure distillation column
Kc = gain of controller
LPC = low pressure distillation column
PSD = pressure swing distillation
PC1, PC2 = pressure controller of the HPC, LPC
QR = reboiler duty
QR/F = reboiler duty/molar ow rate of feed
RR = reex ratio
TC1, TC2 = temperature controller of the LPC, HPC
THF = tetrahydrofuran
I = integral time of controller
DOI: 10.1021/ie505024q
Ind. Eng. Chem. Res. 2015, 54, 16461655

Article

Industrial & Engineering Chemistry Research

(23) Luyben, W. L. New control structure for feed-effluent heat

exchanger/reactor systems. Ind. Eng. Chem. Res. 2012, 51, 85668574.
(24) Luyben, W. L. Decanter Anomaly. AIChE J. 2013, 59, 2088
2095.
(25) Luyben, W. L. Use of mass or molar reflux-to-feed ratios in
distillation single-end control structures. Ind. Eng. Chem. Res. 2013, 52,
1588315895.
(26) Shamsuzzoha, M. Closed-loop PI/PID controller tuning for
stable and integrating process with time delay. Ind. Eng. Chem. Res.
2013, 52, 1297312992.
(27) Jones, D.; Bhattacharyya, D.; Turton, R.; Zitney, S. E. Plant-wide
control system design: Primary controlled variable selection. Comput.
Chem. Eng. 2014, 71, 220234.
(28) Jones, D.; Bhattacharyya, D.; Turton, R.; Zitney, S. E. Plant-wide
control system design: secondary controlled variable selection.
Comput. Chem. Eng. 2014, 71, 253262.
(29) Luyben, W. L. Control of a multiunit heterogeneous azeotropic
distillation process. AIChE J. 2006, 52, 623637.
(30) Luyben, W. L. Control of a column/pervaporation process for
separating the ethanol/water azeotrope. Ind. Eng. Chem. Res. 2009, 48,
34843495.
(31) Ling, H.; Luyben, W. L. New control structure for divided-wall
columns. Ind. Eng. Chem. Res. 2009, 48, 60346049.
(32) Ling, H.; Luyben, W. L. Temperature control of the BTX
divided-wall column. Ind. Eng. Chem. Res. 2009, 49, 189203.
(33) Kiss, A. A.; Rewagad, R. R. Energy efficient control of a BTX
dividing-wall column. Comput. Chem. Eng. 2011, 35, 28962904.
(34) Ling, H.; Cai, Z.; Wu, H.; Wang, J.; Shen, B. Remixing control
for divided-wall columns. Ind. Eng. Chem. Res. 2011, 50, 1269412705.
(35) Zhang, H.; Ye, Q.; Qin, J.; Xu, H.; Li, N. Design and control of
extractive dividing-wall column for separating ethyl acetateisopropyl
alcohol mixture. Ind. Eng. Chem. Res. 2013, 53, 11891205.
(36) Wang, Y.; Cui, P.; Ma, Y.; Zhang, Z. Extractive distillation and
pressure-swing distillation for THF/ethanol separation. J. Chem.
Technol. Biotechnol. 2014, DOI: 10.1002/jctb.4452.
(37) Li, W.; Shi, L.; Yu, B.; Xia, M.; Luo, J.; Shi, H.; Xu, C. New
pressure-swing distillation for separating pressure-insensitive maximum
boiling azeotrope via introducing a heavy entrainer: design and
control. Ind. Eng. Chem. Res. 2013, 52, 78367853.
(38) Luyben, W. L. Design and control of a fully heat-integrated
pressure-swing azeotropic distillation system. Ind. Eng. Chem. Res.
2008, 47, 26812695.
(39) Yu, B.; Wang, Q.; Xu, C. Design and control of distillation
system for methylal/methanol separation. part 2: pressure swing
distillation with full heat integration. Ind. Eng. Chem. Res. 2012, 51,
12931310.
(40) Navarro, M. A.; Javaloyes, J.; Caballero, J. A.; Grossmann, I. E.
Strategies for the robust simulation of thermally coupled distillation
sequences. Comput. Chem. Eng. 2012, 36, 149159.
(41) Vazquez-Ojeda, M.; Segovia-Hernandez, J. G.; Hernandez, S.;
Hernandez-Aguirre, A.; Maya-Yescas, R. Optimization and controllability analysis of thermally coupled reactive distillation arrangements
with minimum use of reboilers. Ind. Eng. Chem. Res. 2012, 51, 5856
5865.
(42) Halvorsen, I. J.; Dejanovic, I.; Skogestad, S.; Olujic, Z . Internal
configurations for a multi-product dividing wall column. Chem. Eng.
Res. Des. 2013, 91, 19541965.
(43) Delgado-Delgado, R.; Barroso-Mun oz, F. O.; SegoviaHernandez, J. G.; Hernandez-Escoto, H.; Castro-Montoya, A. J.;
Rico-Ramrez, V.; Hernandez, S. Multiple steady states in thermally
coupled distillation sequences: revisiting the design, energy
optimization, and control. Ind. Eng. Chem. Res. 2014, 53, 17515
17521.
(44) Dejanovic, I.; Halvorsen, I. J.; Skogestad, S.; Jansen, H.; Olujic,
Z . Hydraulic design, technical challenges and comparison of alternative
configurations of a four-product dividing wall column. Chem. Eng.
Process.: Process Intensification 2014, 84, 7181.
(45) Mulia-Soto, J. F.; Flores-Tlacuahuac, A. Modeling, simulation
and control of an internally heat integrated pressure-swing distillation

REFERENCES

(1) Repke, J.-U.; Klein, A.; Bogle, D.; Wozny, G. Pressure swing
batch distillation for homogeneous azeotropic separation. Chem. Eng.
Res. Des. 2007, 85, 492501.
(2) Luyben, W. L. Pressure-swing distillation for minimum- and
maximum-boiling homogeneous azeotropes. Ind. Eng. Chem. Res. 2012,
51, 1088110886.
(3) Luyben, W. L. Comparison of extractive distillation and pressureswing distillation for acetone/chloroform separation. Comput. Chem.
Eng. 2013, 50, 17.
(4) Wang, Y.; Cui, P.; Zhang, Z. Heat-integrated pressure-swingdistillation process for separation of tetrahydrofuran/methanol with
different feed compositions. Ind. Eng. Chem. Res. 2014, 53, 7186
7194.
(5) Skogestad, S. Control structure design for complete chemical
plants. Comput. Chem. Eng. 2004, 28, 219234.
(6) 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, 790803.
(7) Luyben, W. L. Control of the heterogeneous azeotropic nbutanol/water distillation system. Energy Fuels 2008, 22, 42494258.
(8) Hsu, K.-Y.; Hsiao, Y.-C.; Chien, I.-L. Design and control of
dimethyl carbonatemethanol separation via extractive distillation in
the dimethyl carbonate reactive-distillation process. Ind. Eng. Chem.
Res. 2009, 49, 735749.
(9) 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, 1056410576.
(10) Diemer, R. B.; Luyben, W. L. Design and control of a methyl
acetate process using carbonylation of dimethyl ether. Ind. Eng. Chem.
Res. 2010, 49, 1222412241.
(11) Luyben, W. L. Design and control of the acetone process via
dehydrogenation of 2-propanol. Ind. Eng. Chem. Res. 2010, 50, 1206
1218.
(12) Wang, S.-J.; Yu, C.-C.; Huang, H.-P. Plant-wide design and
control of DMC synthesis process via reactive distillation and
thermally coupled extractive distillation. Comput. Chem. Eng. 2010,
34, 361373.
(13) Luyben, W. L.; Chien, I.-L. Design and control of distillation
systems for separating azeotropes; Wiley: New York, 2011.
(14) Wu, Y. C.; Hsu, C.; Huang, H.-P.; Chien, I.-L. Design and
control of a methyl methacrylate separation process with a middle
decanter. Ind. Eng. Chem. Res. 2011, 50, 45954607.
(15) Luyben, W. L. Distillation design and control using Aspen
simulation; Wiley: New York, 2013.
(16) Qin, J.; Ye, Q.; Xiong, X.; Li, N. Control of benzene
cyclohexane separation system via extractive distillation using sulfolane
as entrainer. Ind. Eng. Chem. Res. 2013, 52, 1075410766.
(17) Wei, H.-M.; Wang, F.; Zhang, J.-L.; Liao, B.; Zhao, N.; Xiao, F.k.; Wei, W.; Sun, Y.-H. Design and control of dimethyl carbonate
methanol separation via pressure-swing distillation. Ind. Eng. Chem.
Res. 2013, 52, 1146311478.
(18) Wu, Y.-C.; Lee, H.-Y.; Lee, C.-H.; Huang, H.-P.; Chien, I.-L.
Design and control of thermally-coupled reactive distillation system for
esterification of an alcohol mixture containing n-amyl alcohol and nhexanol. Ind. Eng. Chem. Res. 2013, 52, 1718417197.
(19) Luyben, W. L. Effect of feed composition on the selection of
control structures for high-purity binary distillation. Ind. Eng. Chem.
Res. 2005, 44, 78007813.
(20) Luyben, W. L. Guides for the selection of control structures for
ternary distillation columns. Ind. Eng. Chem. Res. 2005, 44, 7113
7119.
(21) Hori, E. S.; Skogestad, S. Selection of control structure and
temperature location for two-product distillation columns. Chem. Eng.
Res. Des. 2007, 85, 293306.
(22) Wang, S.-J.; Wong, D. S.; Yu, S.-W. Effect of entrainer loss on
plant-wide design and control of an isopropanol dehydration process.
Ind. Eng. Chem. Res. 2008, 47, 66726684.
1654

DOI: 10.1021/ie505024q
Ind. Eng. Chem. Res. 2015, 54, 16461655

Article

Industrial & Engineering Chemistry Research

process for bioethanol separation. Comput. Chem. Eng. 2011, 35,
15321546.
(46) Luyben, W. L. Control of an extractive distillation system for the
separation of CO2 and ethane in enhanced oil recovery processes. Ind.
Eng. Chem. Res. 2013, 52, 1078010787.
(47) Hosgor, E.; Kucuk, T.; Oksal, I. N.; Kaymak, D. B. Design and
control of distillation processes for methanolchloroform separation.
Comput. Chem. Eng. 2014, 67, 166177.
(48) Buckley, P. S.; Luyben, W. L.; Shunta, J. P. Design of Distillation
Column Control Systems; Instrument Society of America: Research
Triangle Park, NC, 1985.

1655

DOI: 10.1021/ie505024q
Ind. Eng. Chem. Res. 2015, 54, 16461655