chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 346–356

Contents lists available at ScienceDirect

Chemical Engineering Research and Design

journal homepage: www.elsevier.com/locate/cherd

Design and control of an improved acrylic acid

process

Xiaomeng Suo, Hao Zhang, Qing Ye ∗ , Xin Dai, Hao Yu, Rui Li
Jiangsu Key Laboratory of Advanced Catalytic Materials and Technology, Institute of Petrochemical Engineering,
Changzhou University, Changzhou, Jiangsu 213164, China

a r t i c l e i n f o a b s t r a c t

Article history: The acrylic acid (AA) process involves the partial oxidation of a flammable and explosive gas
Received 17 May 2015 medium (propylene), so considerable attention is paid to the concentration of the reactants.
Received in revised form 16 August Water and air is added to dilute the reactant and enhance the thermal stability of the reac-
2015 tion. Different compositions of water and air can lead to different methods to separate AA
Accepted 24 August 2015 product. A modified AA process is developed based on the AA process proposed by Turton,
Available online 1 September 2015 and consists of a tubular reactor, an absorber and only two distillation columns, one of which
is an azeotropic distillation column. This process is characterized by shortening and simpli-
Keywords: fying AA refining process so that equipment investment cost can be reduced. A plantwide
Acrylic acid control structure featured with cascade control provides effective control for the multiunit
Improved process process and insures safe operation of the two distillation columns. Dynamic results also
Dynamic control reveal that it is useful to apply temperature/temperature cascade control and composi-
tion/temperature cascade control to the distillation columns with maximum temperature
limitations.
© 2015 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

1. Introduction There are two safety concerns about the AA process based
on the reactions above. Firstly, the reactions involve par-
Acrylic acid (AA) is widely used as an intermediate of chem- tial oxidation of a flammable and explosive gas medium
icals and polymer in textile industry (Xu et al., 2006). There (propylene), so considerable attention must be paid to the
are several alternative processes to produce it, but the most compositions of propylene and oxygen in the feed stream to
common way nowadays is the partial oxidation of propylene the reactor.
(Lin, 2001). The mechanism of producing AA is that propylene The second safety concern is associated with the highly
is oxidized to acrolein first and then the acrolein is oxidized exothermic polymerization of AA (Cutie et al., 1997). AA of
to AA. However, several side reactions occur, resulting in the high concentration is dimerized at temperatures higher than
oxidation of reactants and products. The main reaction and 110 ◦ C, therefore the separation sequence must be operated
the typical side reactions are as follows: under vacuum to keep the bottom temperature in the distilla-
Main reaction: tion columns below this temperature.
[o] [o]
In Turton’s design (2008) (Fig. 1), a fluidized-bed reactor is
C3 H6 −→CH2 CHCHO−→CH2 CHCOOH used, which is operated at isothermal condition and ensures
safety and stability of the reaction with the addition of large
Side reactions:
amounts of water and air. The water inlet concentration of
[o]
C3 H6 −→CO2 + H2 O the reactor is quite large. As a result, liquid–liquid extraction
is a proper way to separate water from the product stream.
[o]
CH2 CHCHO−→CH3 COOH + CO2 AA and acetic acid (ACE) are extracted into the organic phase.

∗
Corresponding author. Tel.: +86 519 86330355; fax: +86 519 86330355.
E-mail address: huagonglou508@126.com (Q. Ye).
http://dx.doi.org/10.1016/j.cherd.2015.08.022
0263-8762/© 2015 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
 chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 346–356
Nomenclature
AA acrylic acid
ACE acetic acid
TOL toluene
HX heat exchanger
Ns1 total stages of the azeotrope column
Ns2 total stages of De-ACE column
Ntubes the number of reactor tubes
RR reflux ratio
TAC total annual cost
VLL vapor–liquid–liquid equilibrium
LLE liquid–liquid equilibrium
PI proportional and integral settings
Meanwhile, water, AA and a small amount of extracting agent
(diisopropyl ether) get into the aqueous phase. Because of the
introduction of the extractant, two distillation columns are
required for the separation of the extractant from the organic
phase and the aqueous phase. Then AA and ACE are separated
to obtain the final AA product of high purity.
In this proposed design, the inlet water concentration of
the reactor is cut down and air composition is significantly
increased so that water composition in the out-let stream of
the reactor is significantly reduced and a different way based
on the azeotropic principle for dehydration can be considered.
The separation process is much more simplified compared
with Turton’s flowsheet. It does not need extraction column
and the recycled system of extracting agent, and as a result,
capital investment is reduced. Water is separated from the AA
and ACE by using only one azeotrope column and the entrainer
(toluene) is not carried into product stream.
Plantwide control should be noticed, too. That how to con-
struct a plantwide control system using Aspen Dynamics has
been well illuminated in Luyben’s book (2006). The vital and
primary safety consideration in this process is keeping the bot-
tom temperatures in the columns below 110 ◦ C to prevent the
highly exothermic polymerization of AA (Chen and McAvoy,
Fig. 1 – Turton’s flowsheet.
347
2003). The bottom temperatures can be tightly controlled by
manipulating reboiler duty, but the purity specifications of the
bottoms cannot be held easily at the same time. When the feed
flowrate disturbances are introduced, the reactor performance
changes, and as a result, the compositions of the AA and ACE
in the feed stream to the two distillation columns change. The
purity specifications of the bottoms cannot be held when bot-
tom temperature is set at a constant value. In this work, an
effective control structure featured with cascade control is put
forward to solve this problem.
2. Process description
The modified process shown in Fig. 2 produces 51,094 ton per
year of 99.5% by mole AA product. The conditions in the figure
are the optimal results which are analyzed in the later section
of this paper. The operation hour is taken to be 8000 h/year.
The reaction unit is somewhat simplified because it is
assumed that two reactions take place in a single reactor to
produce AA and other byproducts. The products of the reac-
tion are cooled to 70 ◦ C immediately via a heat exchanger (HX)
to avoided further oxidation reactions. Then the stream leav-
ing HX is fed to a flash drum. The gas steam leaving the flash
drum is fed to an absorber column in which AA and ACE are
recovered by a steam coming from the decanter of azeotrope
column. The gas from the top of the absorber is composed of
N2 , O2 , water and a small amount of propylene.
The liquid streams from the bottoms of the flash tank
and the absorber are fed to azeotrope column, the function
of which is to separate water from the AA and ACE by the
formation of water–toluene azeotrope using toluene as the
entrainer. The vapor withdrawn from the top is condensed
and fed to a decanter in which organic and aqueous phases
are formed. The organic phase is recycled to azeotrope col-
umn while part of the aqueous phase is fed to the absorber
to recover AA and ACE. Azeotrope column is operated at
0.265 bar. The top stage temperature is 50.52 ◦ C, which ensures
the use of cooling water in the overhead of the condenser. The
separation of water is achieved using a 14-stage column. It
should be noted that the bottom temperature of the column
348 chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 346–356

Fig. 2 – Improved flowsheet.

must be kept below 110 ◦ C for safety concern. If so many stages

Table 1 – Reaction kinetics.a
are used, the bottom temperature will exceed the maximum
i Ei (kcal/kmol) ko,i
temperature limitation at the condenser pressure of 0.265 bar
with the pressure drop of 0.006895 bar through the trays. 1 15,000 1.59 × 105
The bottom product of azeotrope column is fed to the 2 20,000 8.83 × 105
second distillation column (De-ACE column), the function of 3 25,000 1.81 × 108
which is to separate AA and ACE apart. The separation is a
Overall reaction rates have the units of kmol/m3 reactor h/(kPa)2 .
achieved by using a 34-stage column and a reflux ratio of 79.6.
This column is operated at 0.1 bar. The top stage tempera-
ture is 54.99 ◦ C, so inexpensive cooling water can be used in Partial pressures are in kPa. The preexponential terms and
the condenser. And the bottom temperature is 108.2 ◦ C, so the the activation energies for reaction (1)–(3) are given in Table 1.
maximum temperature limitation is not violated, too.
3.2. Phase equilibrium
3. Kinetics and phase equilibrium
The NRTL-HOC model phase equilibrium parameters, as is
shown in Table 2, are used in the simulations of reac-
3.1. Kinetics
tor, absorber, azeotrope column and De-ACE distillation
column. The ternary diagrams of AA–water–toluene and
The kinetic equations given by Turton are easily implemented
ACE–water–toluene is shown in Fig. 3(a) and (b). The main rea-
in Aspen Plus because they are in the convenient power-law
son for toluene as the favorable entrainer can be explained
form. The reactions taking place are kinetically controlled
under the conditions used in the process. The reaction kinetics
for the catalyst are as follows. Table 2 – The NRTL binary parametersa of the system.
Component i Aij (J/mol) Aji (J/mol) ˛ij
3
C3 H6 + O2 → CH2 CHCOOH + 3H2 O (1) H2 O + AA 919.46 −293.65 0.3
2
H2 O + ACE −723.89 609.89 0.3
5 AA + ACE 42.57 283.02 0.3
C3 H6 + O2 → CH3 COOH + 3H2 O (2) H2 O + TOL −27,269.36 14,759.76 0.2
2
ACE + TOL 0.00 0.00 0.3

9 Values
C3 H6 + O2 → 3CO2 + 3H2 O (3)
a
are calculated as follows:
n  n 
2 ji Gji xj n x Gij xk kj Gkj
ln i = j=1
n + j=1
nj ij − k=1
n
G x G x G x
k=1 ki k j=1 kj k k=1 kj k
 E  Where
Aij −Aji
i
−ri = ko,i exp − p(C3 H6 )p(O2 ) ij = RT , Gij = exp(−˛ij ij ), Aij (J/mol).
RT
 chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 346–356 349

Fig. 3 – The ternary diagram for (a) AA–water–toluene and (b) ACE–water–toluene at 0.265 bar.

by the residue curve maps and the LLE of this system.
The azeotropic compositions of toluene are better because
this mixture contains much more water, which means that
toluene, as the entrainer, is more capable of carrying water
to the top of the column. Thus, less entrainer of toluene is
needed to be inside of the column. The distance in Fig. 3 for
the points between organic phase composition and azeotropic
composition can be better and further apart. So toluene is a
good choice as the entrainer for this system.
The phase equilibrium in the AA process is quite nonlinear
with a heterogeneous azeotrope formed in the binary system
of toluene and water, as is shown in the T–xy diagram given
in Fig. 4. The valid phases in azeotrope column are VLL. The
pressure is 0.265 bar, which gives a heterogeneous azeotrope
temperature of 38.8 ◦ C closed to the decanter temperature. The
bottom stream from the column contains 0.14 wt% water. The
composition of the vapor from the top of the column is close
to the composition of the toluene/water azeotrope.
The separation of ACE and AA in the De-ACE column is not
easy. Fig. 5 gives the T–xy diagram of the AA/ACE system at
a pressure of 0.1 bar generated by Aspen plus. Because of a
small amount of light component (ACE) entering in the feed,
the distillate rate of the De-ACE column is small. As a result,
a high reflux ratio is required. A small amount of impurities
(ACE, in this work) can be removed by distillation, however all
these distillation columns exhibit high reflux ratios.
4. Effect of important design variables
4.1. Reactor
The feed vapor stream is feed to a tubular reactor contain-
ing 16,000 tubes, 0.026 m in diameter and 10 m in length. The
exothermic heat is removed by heat conduction oil cooling at
290 ◦ C. The cooling duty is 24.98 MW. The void fraction of the
catalyst in the tubular reactor is 0.8, and the solid density is
2000 kg/m3 . The heat-transfer coefficient in the tubular reactor
is 1 kW m−2 K−1 (Guan et al., 2002).
Fig. 6 shows the effect of the tube numbers (Ntubes) and
coolant temperature on the production of AA and ACE. The

Fig. 4 – T–xy diagram for toluene/water at 0.265 bar.

350 chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 346–356

Fig. 5 – T–xy diagram for AA and ACE at 0.1 bar.

production of AA and ACE increase as more tubes are used but
the increase is at an increasing smaller rate. Here, 16,000 tubes
are used to realize the same convention in Turton’s process.
The coolant temperature is set at 290 ◦ C, where there exists
a peak value of the production of AA. The production of ACE
increases as coolant temperature increases. So, the reaction
temperature should be well controlled to ensure the largest
production of AA. Fig. 7 gives the temperature profile of the
tubular reactor. The maximum temperature is 300.7 ◦ C and
located at about 0.33 m down the 10-m reactor, which should
be controlled when the reactor is in operation.
4.2. Azeotrope column
The function of this column is to separate water from the AA
and ACE by the formation of water–toluene azeotrope which
is withdrawn from the overhead. The composition of water in
the bottom stream is specified below 0.14 wt%, which can be
achieved by increasing the number of theoretical plates (Ns1).
However, the bottom temperature increases at the same time.
Fig. 8 shows the effects of Ns1 bottom temperature and the
total annual cost (TAC) of this column (Douglas, 1998). TAC
decreases as more trays are added while the bottom tem-
perature increases. So Ns1 is set at 14 so that the bottom
temperature doesn’t exceed 110 ◦ C, and the cost of azeotrope
column is minimized at the same time.
4.3. De-ACE column
The function of column is to separate AA and ACE. It is a
conventional distillation column, but it has maximum tem-
perature limitation in its bottom too. Fig. 9 shows the effects

Fig. 6 – Effect of Ntubes and coolant temperature of AA and ACE production.

 chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 346–356
Fig. 7 – Reactor temperature profile.
Fig. 8 – Effect of Ns1 on bottom temperature and TAC.
Fig. 9 – Effect of Ns2 on bottom temperature and TAC.
351
of Ns2 on TAC and bottom temperature. TAC of the De-ACE
column decreases with the increasing of Ns2. However, the
bottom temperature increases with more trays being used.
Therefore the total stages of De-ACE column is configured at
34 so that the bottom temperature is not violated and TAC of
the De-ACE column is minimized at the same time.
5. Plantwide control
In this part, dynamic control of this process is studied based
on the steady-state design which has been determined above
and shown in Fig. 2. It is necessary to determine the volumes of
all vessels before it is converted to a pressure-driven dynamic
simulation. The units with dynamics include the tubular reac-
tor, the flash tank, the absorber, and two distillation columns.
The size of the tubular reactor has been determined above.
The flash tank, reflux drums and column bases of the two dis-
tillation columns and the absorber base are sized to provide
5 min of hold-up when at half full. The decanter is sized to
provide 10 min of holdup when at the 50% level. The reactor is
simulated as a tubular reactor with constant coolant temper-
ature. It is a 30-lump model and runs without no difficulty in
Aspen Dynamics. The absorber and two distillation columns
are simulated with radfrac model in Aspen. The flash tank is
simulated with flash2 model in Aspen.
For the controls of the two distillation columns, optimal
control trays are selected to hold the composition profile by
adjusting manipulated variables, and it is often possible to
achieve effective control. Several methods about how to select
the best control trays are well summarized in Luyben’s book
and only steady-state information is needed. Here, a sensitiv-
ity criterion is used to find the tray on which there is the largest
change in temperature for a change in the manipulated vari-
able. A small change is made in the reboiler duty of the two
distillation columns and reflux ratio of De-ACE column, and
the resulting change in temperature of all trays is obtained.
The tray with the largest temperature difference is considered
to be the most “sensitive”. Fig. 10 shows the temperature pro-
files of azeotrope column and temperature differences of every
stage by changing its reboiler duty. As a result, the 11th tray
of azeotrope column is selected as temperature control stage
when using reboiler duty as the manipulated variable.
It also should be noticed that there exists maximum tem-
perature limitation in the bottom of both distillation columns.
Therefore, bottom temperatures of azeotrope column must
be tightly controlled. Reboiler duty has the fastest and the
most direct effect on bottom temperatures, so bottom tem-
perature is controlled by manipulating reboiler duty. However,
it precludes the direct use of the control stage when bottom
temperature is used to manipulating reboiler duty, because
reboiler duty cannot be manipulated in two independent tem-
perature control loops at the same time. This dilemma is
resolved by using 11th tray temperature/bottom temperature
cascade control structure.
This structure is to change the set point of bottom tempera-
ture of azeotrope column and De-ACE column to control water
concentration in its bottom. Because bottom temperature is
not held constant, the set point of the bottom temperature
controller is required. A temperature controller is added to
control the temperature of the 11th tray by manipulating the
set point of bottom temperature controller. The output signals
generated by the tray temperature controller are controlled
within a range from 106 ◦ C to 110 ◦ C so that the maximum
352 chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 346–356

Fig. 10 – Temperature profile of azeotrope column and

temperature differences of every stage.

temperature limitation is not violated. Another advantage of

this control structure we proposed is that the reboiler duty
is manipulated to maximum extent under the condition that
maximum temperature limitation is not violated.
Fig. 11 shows temperature profiles of De-AC column and
temperature differences of every stage generated by changing
reboiler duty and reflux ratio by 0.1%. The 3rd tray is the most
sensitive to both reboiler duty and reflux ratio. However, it is
not good to use the 3rd tray temperature-bottom temperature Fig. 11 – Temperature profile of De-ACE column and
cascade loop like azeotrope column. Conventional distillation temperature differences of every stage.
control wisdom suggests that a tray near the bottom of the
column should be controlled to achieve effective inferential
Reflux rate cannot be used to control the tray temperature
composition control of the bottom product by using a tray
directly. Changing reflux ratio influences the reflux rate by
temperature. The 3rd tray is far from the bottom and the strip
holding reflux level constant, and the 3rd tray temperature
section of the column, and large deadtime or lag is introduced
is controlled indirectly.
and poor dynamic performance cannot be prevented when the
The purity of the bottom product is controlled by using
control loop is used.
a tray temperature, a tray near the bottom or using a
Here, a dual-end control scheme is applied to hold the
specifications of AA and ACE. The reason why dual-end con-
trol structure is necessary can be explained as follows. There
is a methodology to decide whether dual-end control struc-
ture is required. A steady-state simulation is run and AA and
ACE purities are held constant with feed composition chang-
ing around the design value. If reflux ratio shows significant
change to ACE feed composition, a dual control structure (two
compositions, two temperatures, or one of each) is required
to handle feed disturbances. Plot of reflux ratio to ACE feed
composition is presented in Fig. 12, and the design value of
reflux ratio is indicated by yellow line. Reflux ratio changes
significantly with ACE feed composition, so dual-end control
structure is necessary.
The 3rd tray temperature is controlled by manipulating
reflux ratio. Because reflux ratio is quite large, the level of Fig. 12 – The effect of feed composition of ACE to reflux
reflux drum must be controlled by manipulating reflux rate. ratio of De-ACE column.
 chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 346–356 353

Table 3 – Parameters of temperature controller.

Controllers TC1 TC2 TC3 TC4 TC5 TC6
◦
SP ( C) 300.7 70 108.7 89.08 108.2 60.16
OP range 250–300 ◦ C −0.00253–0 MW/mol 0–7.8 MW 105–110 ◦ C 0–6.8 MW 0–0.02186 D/R ratio
Kc (gain) 0.3185 0.0601 0.1722 8.796 0.1401 212.2

 I (min) 3.96 5.28 5.28 13.2 5.28 66

Fig. 13 – Plantwide control structure.

composition controller. However, according to the simula-
tion we have run, AA composition cannot be held when
the temperature of the trays is held near the bottom or
the bottom temperature is held constant by manipulating
reboiler duty. So composition control is considered, and
composition–temperature cascade control structure is used in (5)
the bottom of De-ACE column. This control structure com-
bines the advantages of composition control strategy with
temperature control strategy and avoids the defects of both (6)
strategies. Temperature control is fast but it may not hold
purity specifications. Composition control is slow but it can (7)
drive AA purity to the set point. The composition controller
(reserve acting) is used to determine AA composition and
generates set point signals for bottom temperature controller (8)
(reserve acting). The set point ranges from 106 ◦ C to 110 ◦ C, so
maximum temperature limitation is not violated. (9)
The other various control loops are listed as follows:
(10)
(1) Fresh feed of the mixture of propylene, water and air is
flow controlled (reverse acting).
(11)
(2) Peak temperature of the tubular reactor is controlled
by manipulating the temperature of heat conduction oil
(12)
(reverse acting).
(3) The pressure of the reactor is controlled by manipulating
the discharge valve of the reactor (direct acting).
(4) The condenser duty of HX is ratioed to the feed flow
(13)
rate of HX, and the ratio (condenser duty/mass feed flow
rate of HX) is manipulated by the temperature of the
stream leaving HX (direct acting). This feed forward con-
trol structure works well to control temperature changes
of the flow leaving HX caused by fresh feed flow rate
disturbances, and it performs better than the direct tem-
perature control.
The pressure in the flash tank is controlled by manip-
ulating the gas leaving the top of the vessel (direct
acting).
The liquid level of the flash tank is controlled by manip-
ulating the stream leaving the bottom (direct acting).
The base levels of the absorber and two distillation
columns are controlled by manipulating the flow rate of
the stream leaving the bottom (direct acting).
The pressure in the absorber is controlled by manipulat-
ing the flow rate of the off-gas (direct acting).
The flow rate of the wash stream to the absorber is
rationed to the gas feed coming from the flash tank.
The pressure in the azeotrope column is controlled by
manipulating the flowrate of the vapor stream from the
top (direct acting).
Organic reflux from the decanter is rationed to the
azeotrope column feed.
The bottom temperatures of the azeotrope column and
De-ACE column are controlled by manipulating the
reboiler heat inputs of the two columns, respectively
(reverse acting).
The temperature of the decanter is controlled by manip-
ulating the condenser duty (reverse acting).
354 chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 346–356

Fig. 14 – Dynamic response of the control structure to 20% feed flowrate disturbances.

(14) The decanter pressure is controlled by manipulating the
flow rate of the vent stream (direct acting).
(15) Aqueous level is controlled by manipulating the flow rate
of the aqueous stream (direct acting).
(16) Organic level is controlled by manipulating the flow rate
of toluene makeup (reverse acting).
(17) Liquid level in the reflux drum of De-ACE column is con-
trolled by manipulating reflux flowrate because the reflux
ratio is quite large (direct acting) (Luyben, 2009).
(18) Distillate flowrate of De-ACE column is rationed to reflux
flowrate.
(19) The top pressure of De-ACE column is controlled by
manipulating the condenser duty (reverse acting).
(20) A deadtime of 1 min is inserted to all temperature control
loops to fit measurement lags while a deadtime of 3 min
is necessary for composition control loops.
Conventional proportional and integral (PI) settings are
used in all control loops (Luyben, 1996, 2000). The level controls
are only proportional with gain (Kc ) = 2. The pressure controls
of flash tank, decanter and two distillation columns are PI with
Kc = 20 and  I = 12 min, which is the Aspen default turning. The
 chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 346–356 355

Fig. 15 – Dynamic response of the control structure to 10% feed composition disturbances.

flow control is PI with Kc = 0.5 and  1 = 0.3 min. The tempera-
ture and composition control loops, which consist of deadtime
and PI parameters, are obtained by running relay-feedback
tests and using Tyreus-Luyben tuning rules. PI parameters of
temperature controllers are listed in Table 3. The composition
controller is PI with Kc = 1.52 and  I = 50.16.
The plantwide control structure we proposed is shown in
Fig. 13. Now, dynamic performance of this control structure is
evaluated by introducing feed flowrate disturbances and feed
composition disturbances. Fig. 14 gives the dynamic results
for 20% step changes in set point of the fresh feed flow con-
troller with orange lines for the increase and green lines for
the decrease. Fig. 15 shows the response of the control struc-
ture to 10% feed composition disturbances with varying of
air/propylene feed ratio, in which orange lines for the increase
and green lines for the decrease. Both of disturbances are
introduced at 0.5 h, and the whole process reaches a new
steady-state within 8 h. The peak temperature of the reactor is
356 chemical engineering research and design 1 0 4 ( 2 0 1 5 ) 346–356
well controlled. Temperature transient deviations of the flow
leaving HX are controlled within 1 ◦ C. Bottom temperatures
of the two distillation columns are controlled below 110 ◦ C by
using cascade control structure, and the purities of AA and
ACE are held close to the set points.
6. Conclusion
A modified AA process is developed based on Turton’s flow-
sheet. The component of water in the feed for reaction is
significantly reduced, and a method for separating water from
the product based on azeotropic principle is introduced. Com-
pared with that of Turton’s design, the modified process is
much more simplified so that equipment investment cost can
be reduced.
The effective dynamic control of the multiunit process
has been achieved by adding various PI controllers. Close
control of the high-purity AA product is insured by using tem-
perature/temperature and composition/temperature cascade
controls. Large feed disturbances are effectively handled with
bottom temperatures of two distillation columns and product
purities holding close to the set values.
Conflict of interest
The authors declare no financial interest.
Acknowledgments
We are thankful for support from the project fund of the China
Petroleum & Chemical Corp. (411024) and assistance from the
staff at the Institute of Petrochemical Technology (Changzhou
University).
References
Chen, R., McAvoy, T., 2003. Plantwide control system design:
methodology and application to a vinyl acetate process. Ind.
Eng. Chem. Res. 42, 4753–4771.
Cutie, S.S., Smith, P.B., Henton, D.E., Staples, T.L., Powell, C., 1997.
Acrylic acid polymerization kinetics. J. Polym. Sci. Pol. Chem.
B 35, 2029–2047.
Douglas, J.M., 1998. Conceptual Design of Chemical Processes,
first ed. McGraw Hill, New York.
Guan, T., Liu, X.Y., Pen, Z.G., 2002. Calculation for convective heat
transfer coefficient of organic heat transfer material forced
flowing in tube. Ind. Boiler Mag. 6, 9–10.
Lin, M.M., 2001. Selective oxidation of propane to acrylic acid
with molecular oxygen. Appl. Catal. A: Gen. 207, 1–16.
Luyben, W.L., 1996. Tuning proportional-integral-derivative
controllers for integrator/deadtime processes. Ind. Eng. Chem.
Res. 35, 3480–3483.
Luyben, W.L., 2000. Tuning proportional-integral controllers for
processes with both inverse response and deadtime. Ind. Eng.
Chem. 39, 973–976.
Luyben, W.L., 2006. Distillation Design and Control using Aspen
Simulation, first ed. Wiley, New York.
Luyben, W.L., 2009. Unusual control structure for high reflux ratio
distillation columns. Ind. Eng. Chem. Res. 48, 11048–11059.
Turton, R., Bailie, R.C., Whiting, W.B., Shaeiwitz, J.A., 2008.
Analysis, Synthesis and Design of Chemical Processes, third
ed. Pearson Education, Inc., New York, pp. 997–1008.
Xu, X.B., Lin, J.P., Cen, P.L., 2006. Advances in the research and
development of acrylic acid production from biomass. Chin. J.
Chem. Eng. 14, 419–427.