11th IFAC Symposium on Dynamics and Control of

11th IFACSystems,
Process Symposium on Dynamics
including and Control of
Biosystems
11th
11th IFAC
Process
IFAC Symposium
Systems,
Symposium on
on Dynamics
including and
Biosystems
Dynamics and Control of
Controlonline
Available of at www.sciencedirect.com
June 6-8,Systems,
Process 2016. NTNU, Trondheim,
including Norway
Biosystems
June 6-8,Systems,
Process 2016. NTNU, Trondheim,
including Norway
Biosystems
June
June 6-8,
6-8, 2016.
2016. NTNU,
NTNU, Trondheim,
Trondheim, Norway
Norway
ScienceDirect
IFAC-PapersOnLine 49-7 (2016) 533–538
Vinyl
Vinyl Acetate
Acetate Monomer
Monomer (VAM)
(VAM) Plant
Plant Model:
Model:
Vinyl
Vinyl Acetate
Acetate
A New Benchmark Monomer
Monomer
Problem (VAM)
(VAM) Plant
for Control Plant Model:
Model: Study
and Operation
A
A New Benchmark Problem for Control and Operation Study
A New
New Benchmark
Benchmark Problem
Problem for
for Control
Control and
and Operation
Operation Study
Study
Yuta Machida* Shigeki Ootakara** Hiroya Seki*** Yoshihiro Hashimoto**** Manabu Kano†
Yuta Machida* Shigeki Ootakara**§ Hiroya Seki*** Yoshihiro Hashimoto**** # Manabu Kano # †
†
†
Yuta
Yuta Machida*
Yasuhiro
Machida* Shigeki
Miyake
Shigeki
‡
Ootakara**
Naoto Anzai
Ootakara** Hiroya
Masayoshi Seki***Sawai Yoshihiro
§
Takashi Hashimoto****
Katsuno Manabu
Toshiaki Omata Kano
Anzai§ Hiroya Seki*** Yoshihiro Hashimoto**** Manabu Kano
# # †
Yasuhiro Miyake ‡ Naoto
‡ §
Masayoshi Sawai § Takashi
§
Katsuno## Toshiaki Omata #
#
Yasuhiro Miyake ‡ Naoto Anzai§
‡ § Masayoshi Sawai§
Yasuhiro Miyake Naoto Anzai Masayoshi Sawai Takashi Katsuno Toshiaki Omata § Takashi Katsuno# Toshiaki Omata#
*Omega Simulation Co.,Ltd., Tokyo 169-0051, Japan (e-mail: machida@omegasim.co.jp)
*Omega Simulation Co.,Ltd., Tokyo 169-0051, Japan (e-mail: machida@omegasim.co.jp)
*Omega
*Omega Simulation ** Mitsui Tokyo
Co.,Ltd., Chemicals Inc., Osaka 592-8501, Japan
Simulation ** Mitsui Tokyo
Co.,Ltd., Chemicals 169-0051,
169-0051, Japan (e-mail:
(e-mail: machida@omegasim.co.jp)
Japan592-8501,
Inc., Osaka machida@omegasim.co.jp)
Japan
*** Chemical Resources ** Laboratory,
Mitsui
**Laboratory, Tokyo
Chemicals
Mitsui Chemicals Institute
Inc., of
Osaka Technology,
592-8501,
592-8501, Japan Yokohama 226-8503, Japan
*** Chemical Resources Tokyo Inc.,
InstituteOsaka
of Technology,
Technology, Japan
Yokohama 226-8503, Japan
*** ****Dept.
Chemical
*******Dept.
Chemical† of Computer
Resources
Resources Science,
Laboratory, Nagoya
Tokyo Institute of Nagoya
Yokohama 466-8555,
226-8503, Japan Japan
of ComputerLaboratory,
Science, Nagoya Tokyo Institute of Technology, Nagoya Yokohama 226-8503,
466-8555, Japan Japan
****Dept.
****Dept. † ofDept.
Computer
ofDept.
Computerof Systems
Science,
Science,Science,
Nagoya
Nagoya Kyoto University,
Institute
Institute of
of Kyoto
Technology,
Technology, 606-8501,
Nagoya
Nagoya Japan
466-8555,
466-8555, Japan
Japan
†
† of Systems Science, Kyoto University, Kyoto 606-8501, Japan
†Dept. of‡ Systems Science, Kyoto University, Kyoto
Dept. of‡ Ube Industries,
‡Systems Science,Ltd.,
Ube Industries, Kyoto
Ltd.,
Yamaguchi
University,
Yamaguchi Kyoto 606-8501,
755-8633,
755-8633,
Japan Japan
606-8501,
Japan Japan
‡ §
‡Ube Industries,
Ube §Zeon
Industries, Ltd.,
Corp., Yamaguchi
Tokyo 100-8246,755-8633,
Japan Japan
#
§
Zeon Corp., Tokyo 100-8246, Japan Japan
Ltd., Yamaguchi 755-8633,
## Yokogawa§ZeonElectric Corp., 100-8246,
Tokyo 169-0051Japan Japan
§
Zeon Corp.,
Corp., Tokyo
Tokyo 100-8246,
## Yokogawa Electric Corp., Tokyo 169-0051 Japan
Japan
Yokogawa
Yokogawa Electric Corp., Tokyo 169-0051 Japan
Abstract: A rigorous dynamic plant Electric
model Corp.,of a vinylTokyoacetate
169-0051 Japan (VAM) production was
monomer
Abstract: A rigorous dynamic plant model of a vinyl acetate monomer (VAM) production was
developed.
Abstract:
Abstract: This
A
A plant
rigorous
rigorous model
dynamic
dynamic enablesplant
plantthe users
model
model to ofexperience
aa vinyl realistic
vinyl acetate
ofexperience acetate plant
monomer
monomer operation,
(VAM)
(VAM) since it reflectswas
production
production the
was
developed.
real plant This plant model
characteristics and enables
practical the users toon
problems the basis realistic
of plant operation,
experienced practitioners’since it reflects
opinions. the
More
developed.
developed. This
This plant
plant model
real plant characteristics model enables
enables the
and practical the users
users to
problems toonexperience
experience
the basis of realistic
realistic plant
plant operation,
experienced operation,
practitioners’since
since it
it reflects
reflects
opinions. the
the
More
importantly,
real
real plant
plant the plant model
characteristics
characteristics and
and provides problems
practical
practical a new benchmark
problems on
on the
the problem;
basis
basis of
of the users practitioners’
experienced
experienced can investigate
practitioners’ start-up/shut-
opinions.
opinions. More
More
importantly, the plant model provides a new benchmark problem; the users can investigate start-up/shut-
down operation,
importantly,
importantly, the plant-wide
the plant
plant modelprocess
model provides
provides control,
new fault
aa new detection
benchmark
benchmark and diagnosis,
problem;
problem; the usersand
the users canothers.
can Multiple
investigate
investigate scenarios
start-up/shut-
start-up/shut-
down operation, plant-wide process control, fault detection and diagnosis, and others. Multiple scenarios
prepared
down in
operation,
down operation, the developed
plant-wide
plant-widemodel model
process cannot
control,
processcannot
control, be simulated
fault detectionin conventional
and diagnosis, benchmark
and others. problems.
Multiple The plant
scenarios
prepared in the developed be fault detection
simulated and diagnosis,
in conventional and others.
benchmark Multiple The
problems. scenarios
plant
model
prepared
prepared can inbetheused also
developed for chemical
model cannotengineering
be education.
simulated in This
conventional advantageous
benchmark plant model
problems. is
Thereleased
plant
model caninbetheused
developed
also formodel chemicalcannot be simulated
engineering in conventional
education. benchmark
This advantageous problems.
plant model isThe plant
released
from
model Omega
modelOmega can be Simulation
used
can beSimulation also
used also for Co.,
for Ltd.
chemical with a free
engineering limited license
education. of
This Visual Modeler,
advantageous which
plant is
model a commercial
is released
from Co.,chemical
Ltd. withengineering
a free limited education.
license This advantageous
of Visual Modeler, plant whichmodel is released
is a commercial
dynamic
from
from Omega simulator and can
Simulation Co.,be
Co., linked
withwith MATLAB®. Thisof article
Visualaims to introduce isthe VAM plant
dynamic Omega Simulation
simulator and can beLtd.
Ltd. with
linked aa free
with free limited
limited license
MATLAB®. license Visual
Thisofarticle Modeler,
Modeler,
aims which
which is
to introduce theaa VAM
commercial
commercialplant
model,
dynamic the steady-state
simulator
dynamicthesimulator and balance,
can
and balance, be
can be linked various
linked disturbances
with MATLAB®. and malfunctions,
This article and
aims operation
to introducescenarios.
the
the VAM plant
VAM
model, steady-state variouswith MATLAB®.
disturbances This article aims
and malfunctions, to introduce
and operation scenarios. plant
model,
model, the steady-state balance, various disturbances and and malfunctions, and operation scenarios.
© 2016,the
Keywords: IFAC steady-state
Process balance,
(International
simulators, various
Federation
processofdisturbances
Automatic
models, Control) malfunctions,
dynamic Hosting
models, and
by Elsevieroperation
benchmark Ltd. All scenarios.
rights reserved.
examples, chemical
Keywords: Process simulators, process models, dynamic models, benchmark examples, chemical
industry,
Keywords:
Keywords: plant-wide
Process control.
Process simulators, process
simulators, process models,
models, dynamic dynamic models,
models, benchmark
benchmark examples,examples, chemical chemical
industry, plant-wide control.
industry,
industry, plant-wide
plant-wide control.
control.


 simulator attempts to keep the fidelity of the rigorous
1. INTRODUCTION simulator attempts to keep the fidelity of the rigorous
1. INTRODUCTION nonlinear
simulator model
simulator attempts
attempts as high
to as possible.
to keep
keep the
the fidelityThe dynamic
fidelity of the simulator
of the rigorous
rigorous
1. INTRODUCTION
1. INTRODUCTION nonlinear
makes a good model as high
compromise as between
possible. Themodel
the dynamic simulator
accuracy and
A test problem which checks the industrial relevance of new nonlinear nonlinear model
model as
as high
high as
as possible.
possible. The
The dynamic
dynamic simulator
simulator
A test problem which checks the industrial relevance of new makes a good compromise between the model accuracy and
solvability. We improved the VAM plant model developed
ideas
A test and
A test technical
problem whichdevelopments
which checks
checks the is of great
the industrial importance
relevance of
of new in makes aa good Wecompromise between
VAM the model accuracy and
ideas problem
and technical developments industrial
is of great relevance
importance new in makessolvability.
by
good
Yumoto We
solvability.
compromise
improved the
et al. (2010)the
improved
between
with
VAM
the
plant
reference
plant
model
modelaccuracy
developed
to practitioners’
model developed
and
the
ideasstudy
and of process
technical control
developments and process
is of data
great analysis.
importance A
in solvability. We improved the VAM plant model developed
ideasstudy
the and technical
of processdevelopments
control and is of great
process data importance
analysis. in A by Yumoto et al. (2010) with reference to practitioners’
opinions to etmake
great
the
the study
great
number
study
numberof of researchers
of process
process control
control and
of researchers
haveprocess
and utilizeddata
haveprocess
utilizeddata
the analysis.
Tennessee
the analysis.
Tennessee A by
by Yumoto
A opinions Yumoto to etmakeal. realistic
al. (2010)
(2010) with
realistic
operation
with reference
reference
operation
scenarios.
to
scenarios.
The new
to practitioners’
practitioners’
The new
Eastman
great challenge
number of problem
researchers in various
have fields
utilized including
the plant-
Tennessee VAM
opinions
opinions plantto
to model
make
make has the
realistic
realistic following
operation
operation features,
scenarios.
scenarios. which
The
The new the
new
great number
Eastman challengeof researchers have utilized
problem in various the Tennessee
fields including plant- VAM VAM plant model has the following features, which the
wide
Eastman control, production
challenge problem efficiency
in various optimization,
fields soft-sensor
including plant- Tennessee
VAM plant
plant Eastman
model
model process
has
has the
the model cannot
following
following cope with:
features,
features, which
which the
the
Eastman challenge problem in various fields including plant- Tennessee
wide control, production efficiency optimization, soft-sensor Tennessee Eastman process model cannot cope with: Eastman process model cannot cope with:
design,
wide fault
wide control, detection
control, production and
production efficiency diagnosis,
efficiency optimization, etc., since
optimization, soft-sensor it was
soft-sensor -Tennessee It can Eastman
simulate process
transientmodel cannot cope
operations with: start-up
including
design,
introduced fault
more detection
than 20 yearsand diagnosis,
ago (Downsetc., since it1993).
and Vogel, was - It can simulate transient operations including start-up
design,
design, fault
fault detection
detection and
and diagnosis,
ago (Downsetc.,
diagnosis, etc., since
since it was
it1993).
was -- It andcan shut-down.
simulate
introduced
Even now,
more
the
than 20 years
Tennessee Eastman
and Vogel,
problem is the most It
andcan simulate transient
shut-down. transient operations
operations includingincluding start-up
start-up
introduced
introduced
Even more
now,more than
the than 20 years ago
20 yearsEastman
Tennessee (Downs
ago (Downs and Vogel,
and Vogel,
problem 1993).
is the1993). most -- and It can generate
shut-down. noises and disturbances with such
popular problem. Later, Luyben and Tyreus (1998) and
It shut-down.
can generate noises and disturbances with such
Even
Even now,
popular now,
problem.the
the Tennessee
Tennessee
Later, Luyben Eastman
Eastman
and Tyreus problem
problem
(1998) is developed
is the
the most
most -- It
developed realistic
It can
can sources asnoises
generate
generate measurement
noises and
and noise of sensors,
disturbances
disturbances with
with non-
such
such
apopular
model problem.
popular of a vinylLater,acetate monomer (VAM)(1998) plant, which is a realistic sources as measurement noise of sensors, non-
a model problem.
of a vinylLater,acetate Luyben
Luyben
monomer and
and Tyreus
Tyreus
(VAM)(1998) developed
developed
plant, which is a ideality
realistic
realistic
ideality
in actuators
sources
sources
in as
as
actuators
such
measurement
measurement
such
as
as
valve
noise
noise
valve
dead
of
of
dead
band,
sensors,
sensors,
band,
daily
non-
non-
daily
alarger
modelsystem
a model of containing
of aa vinyl
vinyl acetate standard
acetate monomerchemical
monomer (VAM)
(VAM) plant, unit operations
plant, which
which is is aa temperature
ideality in fluctuations,
actuators etc.
such as valve dead band, daily
larger
for system containing standard chemical unit operations ideality
temperature in actuators
fluctuations, such
etc. as valve dead band, daily
largerreal
larger
for system
system
real
chemical
containing
containing
chemical
components.
standard
standard chemical
components.
The process
chemical
The processunit design - temperature
unit operations
operations
design - temperature It can simulate various failure
fluctuations,
fluctuations, etc. modes,
etc. which are set on
background
for real is well-established
chemical components. in the The original paper. Several
process design It can simulate various failure modes, which are set on
for real chemical components. The
background is well-established in the original paper. Several -- It process design the
It can basis
can simulateof the
simulate variousadvice of
various failure experienced
failure modes,
modes, which engineers
which are set
are setand
on
on
studies
background
background on control
is
is system design
well-established
well-established in
in fororiginal
the
the this plant
original paper.
paper.have been
Several
Several the
plant basis of the
operating advice of experienced
personnel. engineers and
studies on control system design for this plant have been the
the basis
basis of
of the
the advice
advice of
of experienced
experienced engineers
engineers and
and
conducted
studies
studies on (Chen
on control
control andsystem
McAvoy, 2003;
design forOlsen
for this et
this al., 2005;
plant have
have beenSeki - plant operating personnel.
It provides an operator interface, which mimics a control
conducted (Chen andsystem
McAvoy, design
2003; Olsen plant
et al., 2005; been
Seki - plant plant operating
operating
It provides personnel.
personnel.
an operator interface, which mimics a control
et al.,
conducted
conducted2010; Tu
(Chen et al.,
and 2013;
McAvoy, Psaltis
2003; et al.,
Olsen2014).
et al., 2005; Seki room of real plants, so that thewhich usersmimics
can experience
et al., 2010;(ChenTu etand al., McAvoy,
2013; Psaltis 2003; Olsen
et al., et al., 2005; Seki -- It
2014). provides
It provides
room of an operator
an operator
real plants, sointerface,
interface,
that thewhich usersmimics aa control
control
can experience
et
et al.,
al., 2010;
2010; Tu
Tu et
et al.,
al., 2013;
2013; Psaltis
Psaltis et
et al.,
al., 2014).
2014). “real”
room plant
of operations.
real plants, so that the users can experience
This paper introduces the advanced VAM plant model, which room
“real” of
plantreal plants, so that the users can experience
operations.
This paper introduces the advanced VAM plant model, which In the “real”
next plant operations.
section, the VAM plant is described. In section 3,
is implemented
This
This paper
paper introduces on the the
introduces commercial
advanceddynamic
advanced VAM
VAM plant simulator,
model, Visual
model, which
which In “real”
the next plant operations.
section, the VAM plant is described. In section 3,
is implemented on the thecommercial dynamic plant
simulator, Visual details of the dynamic model
Modeler
is (Omegaon
is implemented
implemented
Modeler
Simulation
the
the commercial
(OmegaonSimulation commercialCo., Ltd.),
dynamic
dynamic
Co., Ltd.),
as asimulator,
new benchmark
as asimulator,
new benchmark Visual
Visual details
In the next
In the next section,
of section,
the dynamicthe
the VAM
VAM plant implemented
plant
model is
is described.
described. In
implemented Inon Visual
onsection
section
Visual3,
3,
problem
Modeler in lieu
(Omega of the
SimulationTennessee
Co., Eastman
Ltd.), as a problem.
new Visual
benchmark Modeler
details
details are
of
of explained,
the
the dynamic
dynamic including
model
model static balance
implemented
implemented and
on
on process
Visual
Visual
Modeler (Omega
problem in lieu ofSimulation
the Tennessee Co., Ltd.),
Eastman as a problem.
new benchmark Visual Modeler Modeler are explained, including static balance and process
Modeler
problem haslieufound themany applications in problem.
industry Visualas an flow diagram.
are Section 4including
are explained, shows operation
static
static balancescenariosandand
and also
process
Modeler in
problem in
haslieu of
of the
found manyTennessee
Tennessee Eastman
Eastman
applications in problem.
industry Visualas an flow Modelerdiagram.
demonstrates
flow diagram.
explained,
Section 4including
an example
Section 4
shows operation
of
shows start-up
operation
balance
scenarios
operation. In
scenarios
process
and
section
and
also
5,
also
operator
Modeler
Modeler training
has
has found
foundsimulator.
many
many It is capable
applications
applications of
in
in handling
industry
industry various
as
as an
an flow diagram.
demonstrates Section
an example 4 shows
of operation
start-up scenarios
operation. In and
section also
5,
operator training simulator. It is capable of handling various demonstrates summary of supporting
an example ofenvironment
start-up for
operation. the
In users
section is
5,
modes
operator
operator of plant
training
training operations
simulator.
simulator. including
It
It is
is capable
capablestart-up/shut-down
of
of handling
handling as
various
various a demonstrates
summary of an example
supporting of start-up
environment operation.
for In
the section
users 5,
is
modes of plant operations including start-up/shut-down as a described. summary Finally conclusions
of supporting
supporting are drawn. for
environment for the
the users
users is is
prerequisite
modes of
modes of plant of
plant the operator
operations training
including system,
start-up/shut-downwhereas asthea summary of environment
prerequisite of operations
the operator including
trainingstart-up/shut-down
system, whereas asthea described. described.
Finally conclusions are drawn.
Finally conclusions are drawn.
prerequisite of the operator training
prerequisite of the operator training system, whereas the system, whereas the described. Finally conclusions are drawn.
2405-8963 ©
Copyright © 2016, IFAC (International Federation of Automatic Control)
2016 IFAC 533 Hosting by Elsevier Ltd. All rights reserved.
Copyright
Peer review©under
2016 responsibility
IFAC 533Control.
of International Federation of Automatic
Copyright
Copyright ©
© 2016
2016 IFAC
IFAC
10.1016/j.ifacol.2016.07.397 533
533
IFAC DYCOPS-CAB, 2016
June 6-8, 2016. NTNU, Trondheim, Norway Yuta Machida et al. / IFAC-PapersOnLine 49-7 (2016) 533–538
534

Part of VAM removed from the top of the absorber is

2. VAM PLANT
Figure 1 shows the overall process flow diagram of the VAM
plant, which firstly generates VAM product in the reactor,
then separates the product from the unreacted materials and
by-products by using the distillation column and other units.
2.1 Process Overview
Eight materials appear in the VAM plant as shown in Table 1.
Three raw materials are introduced to the process. Ethylene
(C2H4) and oxygen (O2) are fed in the gas phase; acetic acid
(AcOH) is fed in the liquid phase and vaporized with
superheated steam at the vaporizer. These three materials are
mixed and introduced to the reactor, in which the following
gas phase reactions are taking place.
[Main reaction]
C2H4 + CH3COOH + 1/2O2
→
[Side reaction]
C2H4 + 3 O2
→
The main reaction generates VAM (CH2=CHOCOCH3)
product and by-product water (H2O) from C2H4, AcOH
(CH3COOH), and O2. The side reaction generates by-
product carbon dioxide (CO2) and H2O from C2H4 and O2.
Both reactions are exothermic, thus reaction heat is removed
by boiler feed water (BFW) circulation and steam is
generated at the shell side of the reactor. The reactor outlet
gas containing about 5mol% VAM product is cooled down to
37degC with two coolers. Unreacted AcOH, H2O, and VAM
are condensed as liquid VAM crude at the separator. On the
other hand, separated gas leaving from the separator includes
unreacted C2H4, O2, by-product CO2, inert ethane (C2H6),
and a small amount of uncondensed VAM. This separated
gas is compressed by the compressor to circulate recycle gas
flow, then introduced to the absorber. The uncondensed
VAM is absorbed by the cold AcOH which is fed from the
top of the absorber. The mixture of VAM and AcOH is
discharged from the bottom of the absorber, and mixed with
the VAM crude at the intermediate buffer tank.
Table 1. Materials in the VAM plant
Material Description
Ethylene (C2H4) Raw material of VAM
Oxygen (O2) Raw material of VAM
Acetic Acid (AcOH) Raw material of VAM
VAM Product
Water (H2O) By-product
Carbon Dioxide (CO2) By-product
Ethane (C2H6) Accompanying gas of Ethylene
Nitrogen (N2) Inert gas
recycled to the inlet of the process. The remaining part of gas
is introduced to the CO2 remover and the gas purge system,
which keeps concentration of CO2 around 5~10mol% and
C2H6 around 5mol% in the gas recycle line. The VAM crude
at the intermediate buffer tank is fed to the azeotropic
distillation column. VAM-H2O mixture discharged from the
top of the column is condensed at the condenser and
separated at the decanter. VAM forms the organic phase and
H2O goes to the aqueous phase; VAM product is discharged
as organic product from the decanter. Unreacted AcOH is
discharged from the bottom and recycled to both the
vaporizer and the absorber.
2.2 Process constraints
The VAM process must be operated under the following
constraints, which come from safety, equipment protection,
and product quality requirements.
- The O2 concentration must not exceed 8mol%
anywhere in the gas pipeline to avoid an explosion.
- Pressure in the gas pipeline must not exceed 862kPaG
because of the mechanical construction limit.
- The peak reactor temperature must remain below
200degC to prevent damage to the catalyst.
- The reactor inlet temperature must remain above
130degC to avoid dew point of AcOH.
- The AcOH concentration in the VAM product must
remain below 150ppm as a product specification.
- The VAM concentration in the bottom of the distillation
column must remain below 100ppm to prevent
polymerization of VAM.
- The compressor must run while O2 is remaining in the
gas pipeline to avoid forming of O2 hot spot and
explosion.
3. IMPLEMENTATION OF THE VAM PLANT MODEL
The VAM plant model was implemented with Visual
Modeler (Omega Simulation Co., Ltd.), which is the
commercial dynamic simulation software package, with
reference to the process configuration and the static balance
proposed by Luyben and Tyreus (1998). In addition, the
VAM plant model was arranged with reference to actual
issues and opinions from the fields of Japanese chemical
industry to make the model and operation scenarios more
realistic than the original as follows.
- Tuning the steady-state balance
- Adding pipelines and units for start-up operation
- Adding disturbances and malfunctions for abnormal
situation operation

534
 IFAC DYCOPS-CAB, 2016
June 6-8, 2016. NTNU, Trondheim, Norway Yuta Machida et al. / IFAC-PapersOnLine 49-7 (2016) 533–538 535

Fig. 1. Process flow diagram of the VAM plant

Table 2. Representative steady-state stream data. New balance / Original balance (Luyben and Tyreus, 1998)
C2H4 feed AcOH feed O2 feed Reactor inlet Reactor outlet
Stream number 1 2 3 4 5
Flow [kmol/h] 59.4 / 49.9 55.5 / 47.1 35.4 / 31.2 1143 / 1159 1116 / 1137
Temperature [degC] 30.0 / 30.0 30.0 / 30.0 30.0 / 30.0 148.5 / 148.0 156.8 / 154.9
Pressure [kPaG] 933.0 / 931.0 1299 / 931.0 933.0 / 931.0 783.1 / 783.1 525.5 / 518.2
O2 [mol%] 0.000 / 0.000 0.000 / 0.000 1.000 / 1.000 0.060 / 0.074 0.029 / 0.051
CO2 [mol%] 0.000 / 0.000 0.000 / 0.000 0.000 / 0.000 0.058 / 0.007 0.064 / 0.011
C2H4 [mol%] 0.999 / 0.999 0.000 / 0.000 0.000 / 0.000 0.709 / 0.584 0.674 / 0.553
C2H6 [mol%] 0.001 / 0.001 0.000 / 0.000 0.000 / 0.000 0.051 / 0.214 0.052 / 0.218
VAM [mol%] 0.000 / 0.000 0.000 / 0.000 0.000 / 0.000 0.001 / 0.002 0.052 / 0.042
AcOH [mol%] 0.000 / 0.000 1.000 / 1.000 0.000 / 0.000 0.114 / 0.110 0.066 / 0.072
H2O [mol%] 0.000 / 0.000 0.000 / 0.000 0.000 / 0.000 0.007 / 0.009 0.062 / 0.053
N2 [mol%] 0.000 / 0.000 0.000 / 0.000 0.000 / 0.000 0.000 / 0.000 0.000 / 0.000

CO2 purge C2H6 purge Column feed VAM product H2O product
Stream number 10 11 13 15 16
Flow [kmol/h] 4.78 / 5.10 0.93 / 0.18 240.7 / 229.2 58.5 / 49.6 58.7 / 49.9
Temperature [degC] 38.0 / 40.0 38.0 / 40.0 39.3 / 42.5 41.0 / 40.0 41.0 / 40.0
Pressure [kPaG] 815.4 / 781.3 815.4 / 781.3 59.1 / 477.8 22.7 / 22.8 22.7 / 22.8
O2 [mol%] 0.000 / 0.000 0.036 / 0.059 0.000 / 0.000 0.000 / 0.000 0.000 / 0.000
CO2 [mol%] 1.000 / 1.000 0.072 / 68ppm 0.000 / 0.000 0.000 / 0.000 0.000 / 0.000
C2H4 [mol%] 0.000 / 0.000 0.821 / 0.667 0.000 / 0.000 0.000 / 0.000 0.000 / 0.000
C2H6 [mol%] 0.000 / 0.000 0.064 / 0.266 0.000 / 0.000 0.000 / 0.000 0.000 / 0.000
VAM [mol%] 0.000 / 0.000 0.002 / 0.002 0.234 / 0.206 0.963 / 0.950 0.002 / 0.002
AcOH [mol%] 0.000 / 0.000 0.004 / 0.005 0.466 / 0.513 23.0ppm/370 ppm 16.1ppm/370ppm
H2O [mol%] 0.000 / 0.000 0.001 / 0.001 0.300 / 0.281 0.037 / 0.050 0.998 / 0.998
N2 [mol%] 0.000 / 0.000 0.000 / 0.000 0.000 / 0.000 0.000 / 0.000 0.000 / 0.000

535
IFAC DYCOPS-CAB, 2016
June 6-8, 2016. NTNU, Trondheim, Norway Yuta Machida et al. / IFAC-PapersOnLine 49-7 (2016) 533–538
536

Table 3. Modification of the steady-state balance

Changing points Affected streams Backgrounds
Reduce C2H6 4, 5, 6, 7, 9, 11, C2H6 concentration is remained 2 to 5mol% in the actual process.
concentration 12 In the past model, C2H6 concentration was more than 21% and it was
decreasing efficiency of reaction. C2H6 concentration is remained
around 5mol% in the new balance.
Increase CO2 4, 5, 6, 7, 9, 10, CO2 concentration is remained around 5 to 10mol% in the actual
concentration 12 process. In the past model, CO2 was removed perfectly with the CO2
remover, but it was not reasonable balance. CO2 concentration is
remained around 6.0mol% in the new balance.
Increase C2H4 1, 2, 3, 4, 5, 6, 7, VAM reaction efficiency is improved due to changing C2H6 and CO2
concentration; 9, 10, 12, 13, 14, concentration as above. It enables to produce more VAM product with
Decrease O2 15, 16, 17 lower O2 feed concentration. As a result of modification, VAM
concentration production is increased 20% and raw material feed flow balance is
changed. This improvement also affect to feed composition and operation
condition of the distillation column.
Tighten specification 15 AcOH concentration in VAM product is tightened from 600ppm to
of VAM product 150ppm by reference to actual product specification.

3.1 Tuning steady-state balance

Table 2 shows the steady-state material balance of both the
new VAM plant model and the original balance proposed by
Luyben and Tyreus (1998). Table 3 shows representative
modifications of the steady-state, which is the important
target profile for the start-up operation.
3.2 Adding pipelines and units for start-up operations
To make operation scenarios more realistic, the following
pipelines and units are added to the plant model.
- [Configuration of the AcOH vaporizer]
The AcOH vaporizer was expressed with two unit
models, i.e., a heat exchange unit and a vapor-liquid
separation unit, in the original model (Yumoto et al.,
2010). In the new plant model, the vaporizer is expressed
with a single unit model that can calculate heat exchange
and separation simultaneously.
- [Start-up/shut-down lines]
Several start-up/shut-down lines were implemented to
realize realistic start-up/shut-down operation procedures.
In the distillation column, for example, an initial product
charge line and an emergency deliquoring line were
added to the decanter vessel.
- [Intermediate buffer tank before the distillation column]
An intermediate buffer tank is equipped between the
reaction section and the separation section in many
chemical processes to alleviate disturbances from the
reaction section to the separation section. A new
intermediate buffer tank was added before the distillation
column as the same manner of the real process. This
intermediate tank can be bypassed by switching lines
around the vessel on demand.
3.3 Adding disturbances and malfunctions for abnormal
situation operation
Tables 4 and 5 show disturbances and malfunctions, which
were implemented on the new VAM plant model. The users
can activate any single or multiple events at any time, not
only during steady-state operation but also during start-
up/shut-down operation. The users can flexibly change
parameters of disturbances and malfunctions such as
magnitude, time constant, and timing of recovering. The
disturbances are coming from outside of the VAM process,
such as changes in temperature and pressure of utilities. Thus,
the control system should be designed to suppress these
disturbances. On the other hand, process failure caused by
malfunctions such as pump trip cannot be suppressed by
control, so fault detection and diagnosis are needed. The
users are able to investigate various issues related to process
control, process monitoring, and others by using these
malfunctions and disturbances.
4. OPERATION SCENARIOS
The VAM plant model can provide realistic operation
scenarios, which cannot be investigated in conventional
benchmark problems. The users can experience various real
operation of the VAM plant as if they are operators of the
real plant.
4.1 Initial state
The VAM plant model has initial state files which include
information of all units and streams. The users need to load
following initial state files before starting operation.
- [Before start-up]
This initial state is prepared for start-up operation. All
process units remain shut-down condition at this state.
- [Steady-state]
This initial state is the steady-state described in Table 2.

536
 IFAC DYCOPS-CAB, 2016
June 6-8, 2016. NTNU, Trondheim, Norway Yuta Machida et al. / IFAC-PapersOnLine 49-7 (2016) 533–538 537

Table 4. Disturbance list

No Description Pattern
1 Change C2H4 feed composition Step/Ramp
2 Change AcOH feed composition Step/Ramp
3 Change C2H4 feed pressure Step/Ramp
4 Change O2 feed pressure Step/Ramp
5 Heavy rain Step/Ramp
6 Change cooling water temperature Step/Ramp
7 Change cooling water feed pressure Step/Ramp
8 Change 1.5MPa steam feed pressure Step/Ramp
9 Change 5.0MPa steam feed pressure Step/Ramp
10 Change 1.5MPa steam temperature Step/Ramp
11 Change 5.0MPa steam temperature Step/Ramp

Table 5. Malfunction list

No Description Pattern
1 Reduce reaction activity Step/Ramp
2 Reduce heat transmission in vaporiser Step/Ramp
3 Accident error of temperature sensors Step/Ramp
4 Stop absorber circulation pump Step
5 Stop column feed pump Step
6 Stop column bottom pump Step
7 Stop column reflux pump Step
8 Stop compressor Step
9 Stick separator bottom valve Step
10 Stick absorber bottom valve Step
11 Stop analysing sensors Step
12 Failure of gas feed pressure sensor Step
13 Failure of column reflux flow sensor Step
14 Failure of column bottom level sensor Step
15 Crack weir of decanter Step Fig. 2. Start-up operation procedure of the VAM plant
16 Increase differential pressure of column Step/Ramp
17 Decrease efficiency of CO2 removal Step/Ramp
18 Measurement error of column bottom Step/Ramp
level sensor
19 Measurement error of vaporiser Step/Ramp
level sensor

Fig. 3. Time-series data of start-up operation

537
IFAC DYCOPS-CAB, 2016
June 6-8, 2016. NTNU, Trondheim, Norway Yuta Machida et al. / IFAC-PapersOnLine 49-7 (2016) 533–538
538
4.2 Start-up operation
Figure 2 shows the outline of start-up operation procedure of
the VAM plant model. Total time to finish this operation
procedure is almost 20 hours, but the users can accelerate the
speed of simulation calculations up to 200 fold in the case of
time-consuming procedure, such as liquid filling of vessels
and unit warming up. This operation procedure must be
executed under process constraints described in section 2.
Figure 3 shows representative time-series process data during
the start-up operation.
4.3 Production load-up/load-down operation
The users can execute load-up/load-down operation from the
steady-state. Production load of the VAM process can be
increased by raising the O2 feed concentration and the reactor
temperature. Load-down operation is totally opposite. After
changing operation condition of the reactor, the distillation
column condition should be adjusted in conformity to reactor
load. This operation scenario will be a good example to
investigate the plant-wide control loop configuration.
4.4 Abnormal situation operation
The users can start a simulation of the VAM plant model
from the steady-state initial condition and experience
abnormal situation operation by triggering disturbances and
malfunctions at anytime. The users can study dynamic
behavior caused by such disturbances and malfunctions, and
also they can configure control logic and fault detection logic
on the custom calculation blocks which are supported on the
VAM plant model.
5. SUPPORTING ENVIRONMENT FOR THE USERS
The following elements are available on the Omega
Simulation Co., web site for free.
- The VAM plant model with a free limited license of
Visual Modeler.
- Time-series data sets for data analysis study including
fault detection and diagnosis.
- Explanation documents to use the VAM plant model,
which include process description, operation procedure,
system manual, and so on.
We are also developing interface which links the VAM plant
model and MATLAB®. This interface will be helpful to test
control logics developed on MATLAB® as if it is applied to
the real plant.
6. CONCLUSIONS
A rigorous plant model of vinyl acetate monomer (VAM)
process was developed. The users can get feelings of the real
plant operation through this model. It can also be used as a
test environment for various studies in the process control
field, such as plant-wide control, process monitoring, optimal
operation, and so on. The developed VAM plant model with
a free limited license of Visual Modeler is already available
through the web site of Omega Simulation Co., Ltd.:
http://www.omegasim.co.jp/contents_e/product/vm/cnt4/
(Official site : http://www.omegasim.co.jp/index_e.htm)
538
ACKNOWLEDGMENTS
This research has been conducted as a part of the projects of
the Workshop No. 31 (Process Control Technology) of the
Japan Society for the Promotion of Science (JSPS) 143rd
committee on Process Systems Engineering. Discussions
with the workshop members are greatly appreciated.
REFERENCES
Alexandre, C. D. and Costin, S. B. (2011). Chemical Process
Design, chapter 10, Wiley.
Chen, R. and McAvoy, T. J. (2003). Plantwide control
system design: Methodology and application to a vinyl
acetate process. Ind. Eng. Chem. Res., 42, 4753-4771.
Downs, J. and Vogel, E. (1993). A plantwide industrial
process control problems. Computers Chem. Engng, 17,
245-255.
Luyben, M. L. and Tyreus, B. D. (1998). An industrial
design/control study for the vinyl acetate monomer
process. Computers Chem. Engng, 22, 867-877.
Luyben. M. L. et al. (1998). An industrial design/control
study for the vinyl acetate monomer process, Computers
and Chemical Engineering. Vol. 22, No.7-8, pp.867-877.
Machida, Y,. Nakaya, M., Omata, T. (2014). Process
Education and Development Infrastructure Building by
Virtual Plant in DCS Manufacturing Company, The
International Federation of Automatic Control, Cape
Town (IFAC 2014), pp.12237-12242.
Olsen, D. G., Svrcek, W. Y., and Young, B. R. (2005).
Plantwide control study of a vinyl acetate monomer
process design. Chemical Engineering Communication,
192, 1243-1257.
Psaltis, A., Kookos, I. K., and Kravaris, C. (2014). Plantwide
control structure selection methodology for the
benchmark vinyl acetate monomer plant. Computers
Chem. Engng, 62, 108-116.
Rangaiah, G. P. and Kariwala, V. (2012). Plantwide control,
chapter 9, Wiley.
Seki, H., Ogawa, M., Itoh, T., Ootakara, S., Murata, H.,
Hashimoto, Y., Kano, M. (2010). Plantwide control
system design of the benchmark vinyl acetate monomer
production plant, Computers Chem. Engng, 34, pp.1282-
1295.
Tu, T., Matthew, E., and Chistofides, P. D. (2013). Model
predictive control of a nonlinear large-scale process
network used in the production of vinyl acetate. Ind. Eng.
Chem. Res., 52, 12463-12481.
Yumoto, T., Ootakara, S., Seki, H., Hashimoto, Y., Murata,
H., Kano, M., Yamashita, Y. (2010). Rigorous Dynamic
Simulator for Control Study of the Large-scale
Benchmark Chemical Plant, Dynamics and Control of
Process Systems (DYCOPS 2010), pp.49-54.