Escolar Documentos
Profissional Documentos
Cultura Documentos
h i g h l i g h t s
" Production of isophorone incorporating multiple reactions was modelled.
" A dynamic non-equilibrium model of reactive distillation was developed.
" The model was used to understand inuence of various parameters on a key mass ratio.
" Simulated results will be useful for improving performance of isophorone reactor.
a r t i c l e
i n f o
Article history:
Available online 15 July 2012
Keywords:
Isophorone
Reactive distillation
Model
Performance enhancement
a b s t r a c t
The production of isophorone in a reactive distillation column was modelled. Mass transfer and heat
transfer accompanied by chemical reactions were incorporated in a dynamic non-equilibrium multistage
model. A base case simulation was identied and established. The base case simulation showed the value
of key performance indicator that is mass ratio of acetone consumed to isophorone produced as 1.54. The
computational model was then used to understand and to evaluate inuence of various key design and
operating parameters on this key performance indicator. Inuence of reactor volume, mass transfer coefcient and feed composition, location and temperature was evaluated. The simulated results indicate that
the key performance indicator can be improved (reduce acetone consumed per isophorone produced) by
manipulating acetone content in the feed and location of the feed. The presented results will be useful for
realising performance enhancement of isophorone and other similar reactors.
2012 Elsevier B.V. All rights reserved.
1. Introduction
Isophorone is an important industrial chemical having applications in inks, paints, lacquers, adhesives, vinyl resins, copolymers,
coatings, nishes, and pesticides. It is also used as a chemical intermediate for organic synthesis. Isophorone is commercially produced [1] by alkali catalysed condensation of acetone at high
pressure and high temperature. Aldol condensation of acetone is a
complex reaction [28] involving the feedstock: acetone, water
and hydroxide catalyst; several intermediate products (diacetone
alcohol, mesityl oxide), byproducts (oxylitone and heavies) and
the desired product of the reactionisophorone. Acetone condensation for isophorone production can be commercially feasible, if the
specicity of the reaction is controlled to achieve a high selectivity
towards isophorone. One of the key indicators of the plant performance in isophorone production is the overall mass ratio of
acetone consumed to isophorone produced. In this work, a mathematical model of isophorone reactor is developed and used for
350
P.J. Darda, V.V. Ranade / Chemical Engineering Journal 207208 (2012) 349367
Nomenclature
Total and component molar balance
MLj
total number of liquid phase moles on stage j (kmol)
MVj
MLi;j
MVi;j
Hdiff
i
W diff
i
MW(i)
T Critical
i
Tr(i,j)
DL(i,j)
a
CE
holdupL(j)
S(j)
ULE(j)
ULS(j)
e
h
lLj
qLj
qVj
Abubbling
j
P.J. Darda, V.V. Ranade / Chemical Engineering Journal 207208 (2012) 349367
hj
lVi
1
AVj
V
T
T Interface
j
T LB
j
T VD
j
hj
V
Volj
eL
H
i;j
Weir ow calculations
heightow(j) height of the clear liquid owing over the weir (m)
Q Lj
liquid phase volumetric ow originating from stage j
(m3 h1)
lengthw(j) weir length (m)
g
acceleration due to gravity (m s2)
L
dVolumej
dt
Volj
351
eV
H
i;j
T
T Lj
T Vj
C Li
Energy balance
ELj
liquid phase energy content of stage j (kJ)
C Vi
EVj
ELj
PTotal
j
PSi;T
HLf
j
DHS
kAi
HVf
j
OPLj
vapour pressure
(mm Hg/bar)
Antoine constant
Antoine constant
Antoine constant
IPVj
EVj
QLostLj
QLostVj
ALj
1
1
A(i)
B(i)
C(i)
of component i at temperature T
A of component i ()
B of component i ()
C of component i ()
Abbreviations
NC
number of components
RatioA/IPH key mass ratio of acetone consumed to isophorone
produced
AvgVolL average liquid stage volume
with the solution containing water or even lower alkanols. Typically a temperature gradient exists in the column so that the temperature in the lower purication section (231 C) exceeds the
temperature in the median reaction section (210 C), while temperature in the reaction section is higher than the temperature in
the upper distillation section (200 C). A pressure of 27 atm is
maintained in the column while a pressure of 1516 atm is maintained in the head of the distillation column reactor. Alkali solution
and acetone condensation reaction products formed as a consequence of the reaction are collected in the sump. The alkali solution
collects as the lower aqueous phase while the condensation reaction products collect in the sump as the upper organic phase. Removal of a fraction of water from the aqueous phase and loss of
alkali due to neutralisation may be made up by fresh alkali addition to the recycle stream introduced at the boundary of the reaction and distillation zones. Acetone consumed in the reaction zone
is also made up by fresh acetone introduced at the boundary of the
reaction and purication zones. The acetone condensation products consist of isophorone and high boiling reaction products.
The objective of the design is to minimise the formation of undesired products consisting of mesityl oxide, phorone, b isophorone
and xylitones. Alkali in lower phase of the sump aids in hydrolysis,
yielding isophorone and minor portion of the higher boiling over
condensates. The condensation reaction products recovered from
352
P.J. Darda, V.V. Ranade / Chemical Engineering Journal 207208 (2012) 349367
Reflux
Condenser
Recycle to
Reaction Column
Reflux
Feed
Liquid
Flow
Height of Clear
Liquid Flowing
Over Weir
Weir
Weir
Gas
Liquid
Mass
Transfer
Gas
Flow
Hydrolysis
Column
(30 stages)
Reaction
Column
(11 stages)
Bottoms
Stage
Steam Feed
Product of
Reaction Column
Product of
Hydrolysis Column
Fig. 1. Schematic representation of the isophorone production process using a reactive distillation column (adapted from [1]).
the sump of the zone may be subject to vacuum distillation to recover pure isophorone.
Production of isophorone via alkali catalysed condensation of
acetone (for example see [1,9]) can lead to colour forming compounds and the unwanted isomer: b isophorone. Cane and Yeomans [10] address the issue of colour formation and claim a
process, which incorporates removal of a side-stream from hydrolysis section of the column. Colour forming compounds are concentrated at a point in the column where acetone content is stripped
to a relatively low level. The side stream is distilled to obtain acetone/water in the distillate product, while the residue is treated
with a strong mineral acid like toluene-p sulphonic acid, which
decomposes the contaminants into high boiling compounds. The
treated residue is further distilled to obtain isophorone in the distillate (containing 9% w/w of unwanted isomer: b isophorone).
Additionally the emulsied base product from the hydrolysis section is treated with a strong mineral acid like phosphoric acid to reduce the pH in the range of 6.57.5. Neutralisation of the crude
isophorone stream from the bottom of the hydrolysis section reduces the alkali metal ion content signicantly. The acid treated
stream is decanted and the isophorone containing phase is fed to
distillation column for water/crude isophorone separation. Crude
isophorone is further distilled to obtain pure isophorone in the distillate. The nal product contains about 0.9% w/w of b isophorone.
In contrast to multiphase (vapour/liquid) processes for isophorone production in reactive distillation columns, a process for
isophorone production in vapour phase via aldol condensation of
acetone has been claimed by Grebinoski et al. [11]. The objective
of their invention is to improve efciency in product yield and energy consumption. The process consists of feeding fresh acetone
mixed with various process recycle streams made up of acetone,
mesityl oxide and isophorone. Recycled by product streams
either equilibrate or convert to the desired productisophorone
thereby increasing the overall process yield of isophorone from
acetone. Vaporized and superheated stream is fed to the reactor
at 225325 C. The superheated feed reacts in vapour phase
converting 1035% of acetone. Initial reaction product consists of
420% of desired product: a isophorone. The catalyst used for the
process has been described by Schutz [12] and Schutz and Cullo
[13]. Unreacted acetone is separated from reaction products and
recycled to the reactor. Several distillation columns are used to
separate acetone/organic components/water, mesityloxide/water/
reaction products, phorone/b isophorone/mesitylene/a isophorone
and a isophorone/heavies.
The objective of improving process efciency in isophorone
production has also been investigated by Cook [14]. Cook [14]
highlights the strategies of reduce/reuse/recycle applied to the
isophorone production process. According to Cook [14], the
P.J. Darda, V.V. Ranade / Chemical Engineering Journal 207208 (2012) 349367
353
354
P.J. Darda, V.V. Ranade / Chemical Engineering Journal 207208 (2012) 349367
tion occurring in the liquid phase, and overall mass and enthalpy
balances. Vapourliquid interface mole fractions are related via
UNIQUAC model dened activity coefcients and distribution coefcients. Accumulation terms were introduced in the liquid and vapour phase component balances and the resulting system of
differential algebraic equations (DAEs) was solved using the solver:
LIMEX. Sundmacher and Hoffmann [19] also describe a steady
state non-equilibrium model for packed reactive distillation column. The model consists of molar balances for the condenser, column and the reboiler. Column balances consist of component and
overall vapour and liquid molar balances. Mass transfer in the liquid and vapour phase is driven by liquid and vapour phase mole
fraction differential. Equilibrium conditions were dened using
activity coefcient based distribution coefcients, with the activity
coefcients dened using UNIQUAC method. The model system
was solved using a relaxation method which gave rise to a system
of differential algebraic equations (DAEs). Resulting DAEs were
solved using extrapolation solver: LIMEX.
2.2. Other models
Amongst other models, the general rate based model applicable
for a broad spectrum of NOx absorption operations described by
Hupen and Kenig [21] consists of balances in the bulk liquid/gas
phases, liquid/gas lms, and in the gas area. The bulk liquid balance takes into account axial change along with mass transfer
and chemical reactions. The differential heat balance of the liquid
bulk takes into account heat ux and cooling effects. Balances
are also described for the gas bulk. A differential component balance in the liquid lm accounts for mass transfer with chemical
reactions resulting in an effective diffusivity description of the
component uxes. The differential energy balance for the lm accounts for heat ux including the conductive and convective terms.
In addition to liquid/gas bulk and lm balances, special gas area
equations describing the change in gas composition, reactions,
and energy balances are also incorporated in the model. The system of equations is discretised in axial and normal co-ordinates
to yield an algebraic system of equation which were implemented
in Aspen Custom Modeler.
Contrary to non-equilibrium models summarised in Section 2.1,
Alejski and Duprat [20] describe an equilibrium stage model, which
includes the component and total balances for condenser/partial
condenser, plates and reboiler. Total energy balance along with vapourliquid equilibrium and plate efciency complete the description of the model. The model was solved using Gears method for
stiff systems with automatically chosen order and length of the
integration step. In addition to the aforementioned equilibrium
model [20], the critical issues of approximating the mixing effects
on the distillation plate by a cascade of stirred cells and application
of minimisation method for solution of the reactive distillation
system were addressed by Alejski [22] and Alejski et al. [23]
respectively.
Relevant models of reactive distillation have also been described by Qi et al. [24] and Gruner et al. [25]. Qi et al. [24] describe
a steady state model for packed reactive distillation column, while
the problem of control in reactive distillation is addressed by Gruner et al. [25].
2.3. Model development
In this work, the dynamic non-equilibrium (NEQ) model was
developed for a multi-stage reactive distillation system including
multiple components, multiple chemical reactions and appropriate
heat transfer and mass transfer terms.
Although the interest of the present work was to study steady
state performance of reactive distillation system, a dynamic model
355
P.J. Darda, V.V. Ranade / Chemical Engineering Journal 207208 (2012) 349367
was chosen to have the possibility of carrying out transient simulations and also for numerical reasons. Solution of non-linear algebraic equations describing a steady state model, using a solver like
Sundials-2.4.0, requires good initial guesses for stable region of
convergence. Convergence issues can be better handled by dening
the model system as a time dependentdynamic model, which can
be solved using a suitable ordinary differential equation solver or a
solver for differential/algebraic equations in Sundials-2.4.0.
The reactive distillation model consisted of the following balances (Appendix A): (1) overall molar balances, (2) component molar balances, (3) component mole fractions, (4) liquid/vapour phase
volumetric hold up on each stage and (5) overall energy balances.
Additional balances equations were also formulated for: (1) equilibrium and forward reaction rates, (2) multi-component mass
and heat transfer rates across gasliquid interface using a lm
model, (3) dew point, bubble point and interface temperatures
and (4) temperature of the bulk liquid and vapour phases along
various stages using stage enthalpies. The various components
modelled in the reactive distillation system included: water, acetone, isophorone, hydroxide (catalyst), mesityl oxide, isoxylitone
and diacetone alcohol (7 components).
The input to the model consisted of the following parameters:
(1) liquid and vapour phase molar volumes of various components,
(2) initial liquid and vapour phase mole fractions of various components, (3) overall liquid side mass transfer coefcients, (4) liquid
and vapour phase input mole fractions, (5) pressure, (6) ratio of liquid and vapour phase capacity volume to initial volume, (7) liquid
output/accumulation ratio I, vapour output/accumulation ratio, liquid output/accumulation ratio II and the time instance at which
the ratio of liquid output/accumulation is switched from ratio I
to ratio II, (8) initial liquid and vapour phase moles on each stage,
(10) input liquid and vapour phase molar ow to the feed stage, (9)
location of feed stage and reux ratio, (10) Antoine constants A, B
and C for calculating vapourliquid distribution coefcients, dew
point and bubble point temperatures, (11) liquid heat capacity, vapour heat capacity, heat of vapourisation, heat of solution, input liquid and vapour phase enthalpies, reference temperature and
liquid and vapour phase heat transfer coefcients and heat transfer
area. Additionally the following inputs were set in the main program in order to complete the denition of the reactive distillation
system studied: (1) number of equations, (2) number of species, (3)
number of reactions, (4) number of stages, (5) overall simulation
time, (6) rate constants of various reactions involved in isophorone
production and (7) activity coefcients of various components.
The values of various inputs used in the simulation of the reactive distillation model are dened in the base case simulation (Tables 3a and 3b). Specic data and information were not readily
available for estimating many of these signicant parameters.
Since the objective of the present study was to evolve guidelines
for performance enhancement, suitable approximations were
made to estimate these parameters. The published information
on reaction mechanism, stoichiometry, order, forward rate constants and equilibrium constants and effect of temperature/catalyst concentration on the rate constants was used. The VLE
parameters of the considered system including activity coefcients
were estimated using UNIFAC group contribution method for various components [30]. Activity coefcients of acetone and water
obtained by UNIFAC group contribution method were tted for
various values of acetone/water liquid phase mole fractions. Azeotrope formation at higher acetone mole fraction was avoided in
simulations by operating in the lower range of acetone mole fraction. Heat and mass transfer coefcients were estimated from the
literature, while physicochemical properties of various components were obtained from the literature and from plant data. The
key output variables (results described in Section 3) of the model
included: (1) liquid and vapour phase mole fractions of various
components on each stage, (2) liquid and vapour phase molar ows
originating from each stage, (3) liquid and vapour phase volume on
each stage, (4) dew point and bubble point temperatures corresponding to the stage composition on each stage and (5) bulk liquid and vapour phase enthalpies and temperatures on each
stage. The issue of dening output liquid ows originating from
various stages of the reactive distillation column is discussed next.
2.3.1. Volume and liquid outow from a stage
One of the key issues in model development for reactive distillation is the description of the output liquid and vapour molar
ows originating from various stages. The output ows originating
from a stage also inuence the steady state volume attained at
each individual stage. In this work, since a dynamic and non-equilibrium model of reactive distillation was developed, it was essential to account for variable volume in conjunction with output
ows in the denition of the model equations (Appendix A: Eqs
A.2, A.3, A.11, A.12, A.19 and A.20). Various authors in the literature
have outlined different methodologies for addressing the issue of
characterising liquid and vapour ow in a distillation column and
these are described in this section.
Dalaouti and Seferlis [31] have proposed a modelling approach
that combines the rigorous non-equilibrium (NEQ) rate based
equations with the model order reduction properties of orthogonal
collocation on nite elements (OCFEs) approximation techniques.
These techniques can be employed to the optimal design, operation
optimisation and dynamic simulation of complex staged reactive
separation processes. The main feature of the OCFE formulation
is that material and energy balances resulting from the NEQ rate
based equations are satised exactly only at the chosen collocation
points the roots of the discrete Hahn family of orthogonal polynomials. Such a selection ensures that the collocation points coincide with the location of the actual stages in the limiting case that
the order of the polynomial becomes equal to the number of actual
stages for any given column section. Consequently, the complete
model can be fully recovered. Lagrange interpolation polynomials
were used within each nite element to approximate the liquid
and vapour component ow rates, as well as the liquid and vapour
stream enthalpies.
The non ideal ow behaviour of liquid phase has also been characterised by Noeres et al. [32] by implementing one point injection/ one point detection method to study liquid phase residence
time distribution for the catalytic structured packing MULTIPAK.
In their study, conductivity measurements were done under different liquid and gas ow rates. Measurements were evaluated using
the one dimensional axial dispersion (ADM) model and partial differential equation (PDE) model with axial dispersion and mass exchange between the dynamic and the stagnant liquid zone.
Wijn [33] and Jhon and Lee [34] cite the application of the Francis Weir equation for calculation of output liquid volumetric ow
from a stage, using the height of the clear liquid owing over the
weir and the weir length. Wijn [33] indicates that considerable
deviations from the Francis Weir equation can be expected,
depending on the ow regime, gas/liquid properties, operating
conditions and tray geometry. We use the Francis Weir equation
to have an order of magnitude comparison of the height of the clear
liquid owing over the weir (heightow) in simulations with the
experimental data of Wijn [33]. The Francis Weir equation used
for validating liquid ows is shown in Eq. (W1).
heightowj 1:43
Q Lj =3600 lengthwj 2
!1=3
g
Q Lj
6593:7572 lengthwj
!2=3
W1
356
P.J. Darda, V.V. Ranade / Chemical Engineering Journal 207208 (2012) 349367
Q Lj
Output=Accumulation
L
ft
LRatioj
dVolumej
!
W2
dt
heightowj
L
ft
LRatioj
!2=3
dVolumej
6593:7572 lengthwj
dt
!2=3
W3
be seen from Fig. 2 that liquid stage volume is higher around the reux/ liquid feed region of the column and decreases with increasing
stage number towards the bottoms part of the column. Results of
Fig. 2 correspond to the denition of the liquid phase volumetric
ows depicted in Eq. (W2).
The dening expression for Eq. (W3) Francis Weir Eq. (W1) is
based on the premise that heightow is a function of the volumetric
liquid ow originating from a stage and the weir length. Simulation
results (base case simulation) indicate that volumetric liquid ow
originating from various stages increases with increase in the simulation time and decreases as the stage number increases from 1 to
41. This increase/decrease in the liquid volumetric ow originating
from a stage translates to corresponding increase/decrease in the
value of implied heightowin the simulation results. The Francis
Weir Eq. (W1) can also be applied to the isophorone distillation
column reactor plant data to obtain the values of the implied heightow. When the values of the output liquid volumetric ows of reaction and hydrolysis columns (Fig. 1) are used, the values of implied
heightow obtained are: reaction column output: 1.67 102 m and
hydrolysis column output: 8.88 103 m. Besides the volumetric
liquid ow, the values of the overall liquid side mass transfer coefcient also inuence the values of implied heightow. Steady state
results of the simulations with increasing values of overall liquid
side mass transfer coefcients including 1.25 (base case), 1.75,
2.25 and 3.50 h1 show that the value of implied heightow for various stages increases with increasing values of the overall liquid
side mass transfer coefcient. It is also important to note here that
the simulation values of the implied heightow (7.02 103
1.42 102 m) and plant data implied heightow values of
1.67 102 m (for reaction column output) and 8.88 103 m
(for hydrolysis column output) are in line with the experimental
results cited by Wijn [33] for sieve and valve trays indicating consistency in order of magnitude for implied heightow in simulations
with experimentally validated tray hydraulics. Contrary to volumetric accumulation rate dependent output liquid ow originating
from various stages, the description of the kinetics and reaction
rates in the model (Appendix A) is substantially rigorous and this
requires precise denition of multiple reaction rate parameters.
Various reactions involved in the aldol condensation of acetone
for isophorone production are described next along with the reaction rate parameters obtained from the literature and from the
simulation results.
1.20
Liquid
Volume:
Stage 1
1.00
0.80
Liquid
Volume:
Stage 11
Liquid Stage
0.60
Volume
[m3]
Liquid
Volume:
Stage 21
0.40
Liquid
Volume:
Stage 31
0.20
Liquid
Volume:
Stage 41
0.00
1.00E+00
1.00E+01
1.00E+02
1.00E+03
1.00E+04
357
P.J. Darda, V.V. Ranade / Chemical Engineering Journal 207208 (2012) 349367
2A
DAA
k2
k3
DAA
MO H2 O
k4
R1 ; R2
R3 ; R4
A MO
PH
k6
k7
2MO ! IPH A
k8
PH
IPH H2 O
k9
R5 ; R6
R7
R8 ; R9
A MO
IPH H2 O
k6
R5 ; R6
Cross condensation between various ketones: acetone and isophorone leads to the formation of decomposition product of isophorone:isoxylitone (IX) (reversible reactions: (R10,R11)). Also, two
molecules of mesityl oxide can further condense to form xylitones
(reversible reactions: (R12,R13)). This condensation pathway may
continue further to form heavier ketones.
k10
IPH A
IX H2 O
k11
k12
2MO
IX H2 O
k13
R10 ; R11
R12 ; R13
kk 0 for k 2; 4; 6 and 13
M1
The resulting 7 reactions: R1, R3, R5, R7, R10, R11 and R12 considered
in mechanism (M1) dene the reaction system used in the
2
6
6
6
6
6
6
6
6
6
6
6
6
6
4
Components=Reactions
R1
R3
R5
R7
Water
1 1
Acetone
2
1
1
1 1
1
0
0
MesitylOxide
1 1 2
1 7
7
7
0 7
7
0 7
7
7
0 7
7
2 7
7
7
1 5
1
1
Isoxylitone
R12
1 1
Hydroxide
1 1
R11
Isophorone
DiacetoneAlcohol
R10
1
0
I1
2
6
6
6
6
6
6
6
6
6
6
6
6
6
4
Components=Reactions
R1
R3
R5
R7
R10
R11
R12
Water
0 7
7
7
0 7
7
0 7
7
7
0 7
7
2 7
7
7
0 5
Acetone
Isophorone
Hydroxide
0
0
0
0
0
0
0
0
1
0
0
0
MesitylOxide
Isoxylitone
DiacetoneAlcohol
1 2
0
I2
Reactions
BaseCoefficient
R1
R3
R5
R7
R10
R11
R12
1 1 1 1
I3
Reactions
BaseComponent
R1
R3
DAA DAA
R5
R7
R10
R11
R12
IPH
IPH
IX
IX
I4
358
P.J. Darda, V.V. Ranade / Chemical Engineering Journal 207208 (2012) 349367
Table 1
Pre-exponential factors and activation energies for various reactions involved in the aldol condensation of acetone.
Reaction
Rate constant/
equilibrium
constant
Overall
reaction
order
Inferred pre-exponential
factor k0(k)/equilibrium
constantb
Inferred/assumed
activation energy E(k)
(kJ kmol1 K1)
Reference
k1
k1
k1
k1
k1
k2
k2
K1,2
2
1
2
2
2
1
1
1,1
4.08 102
4.17 109
6.67 109
1.93 103
9.71 106
1.95 1017
3.37 109
2.26 102
14,650
34,400
71,057
34,400a
34,400a
97,540
34,400a
k3
k3
k3
k3
k4
K3,4
1
1
0
1
1
1,1
7.19 1015
9.68 1011
2.55 1012
4.89 1015
5.98 108
5.36 106
90,000
78,523
90,000a
90,000a
53,038
25,322
0
2
5.35 102
3.19 1017
34,400a
93,259
4.52 1014
90,000a
kLi;j
DLi;j C E U LEj 0:5
2
pSj
K1
Table 2
Parameters used for estimating mass transfer coefcients.
Mass transfer parameter
Mass
L
Flowj ;
Mass
V
Flowj
Description
Numerical values
Liquid phase:
2.341.15 kg h1
Vapour phase:
1.890.68 kg h1
Liquid phase:
587.0
746.0 kg m3
Vapour phase:
29.64
18.70 kg m3
1.0 m
1.50 m2
qLj ; qVj
lengthw(j)
Weir length
Bubbling area of the tray
Abubbling
j
heightw(j)
0.64 m
0.26 m
uSV(j)
r(j)
pitch(j)
Ahole
j
L
height j
0.20 m
0.29 m2
4.25
2.42 102 m s1
2.04 102 N m1
In the above expression (K1), liquid phase effective velocity is calculated using the following equation:
U LEj
U LSj
e holdupLj sin h
K2
The diffusion coefcient in the Eq. (K1) can be computed using the
correlation provided by Umezawa and Nagashima [41] for acetone,
benzene and alkanes in supercritical carbon dioxie using the Taylor
Dispersion method. The dependence of diffusion coefcient on
359
P.J. Darda, V.V. Ranade / Chemical Engineering Journal 207208 (2012) 349367
DLi;j
Adiff
i
exp
Bdiff
i
T ri;j
Hdiff
T ri;j 12
i
0:5 exp
W diff
i
W diff
i
!
K3
T ri;j
T Lj
Critical
T i
K5
Numerical values of various parameters including those of structured packing [16] required for calculation of liquid side mass transfer coefcient (Eq. (K1)), liquid phase effective velocity (Eq. (K2)) and
for liquid diffusion coefcient (Eq. (K3)) are: (1) ULS(j) = 1.84
103 m s1, (2) holdupL(j) = 0.40, (3) e = 0.90 and 0.622; h = 60 and
45; a = 492.0 m2 m3 and 128.20 m2 m3 for Sulzer BX and KATAPAK structured packings respectively, (4) S(j) = 1.11 m, and (5)
CE = 0.90.
On substituting the numerical values of various parameters in
Eqs. K1, K2 and K3, the following values of the liquid phase diffusion coefcient and liquid side mass transfer coefcient for Sulzer
BX structured packing are obtained: DL(i,j) = 2.90 108 m2 s1,
kL(i,j) = 1.32 105 m s1. Using the value of packing area
(=492.0 m2 m3) for Sulzer BX structured packing, the value of
overall liquid side mass transfer coefcient including specic area
is: kLa(i,j) = 6.49 103 s1 = 23.36 h1. Similar calculations for the
structured packing: KATAPAK yield the values of liquid side mass
transfer coefcients: kL(i,j) = 1.75 105 m s1 and kL a(i,j) = 2.25
103 s1 = 8.10 h1.
Mass transfer coefcient values obtained for structured packings were followed by calculations for sieve trays in order to have
a better understanding of the mass transfer parameters used in the
model. Svandova et al. [39] have chosen four empirical correlations
(including correlations of [40]) for the estimation of the number of
transfer units or relations for the estimation of binary mass transfer coefcients for a sieve tray. In this work we estimate the mass
transfer coefcient and interfacial area based on the correlations of
Zuiderweg [40] shown in Eqs. (K6)(K14). These correlations are
applicable to sieve trays. We assume that the value of interfacial
area obtained from these correlations can be used in the present
study. The value of liquid side mass transfer coefcient can be calculated using the viscosity of liquid phase (1.61 104 Pa s) in the
lLj
0:25
m s1
K6
Eq. (K6) yields a value of 2.31 105 m s1/8.31 102 m h1 for
the mass transfer coefcient. Apart from the mass transfer coefcient, interfacial area is also required, to complete the denition
of the overall liquid side mass transfer coefcient. The interfacial
area per unit volume of the liquid phase was expressed as the
parameter ahf(j): interfacial area per unit bubbling area. Estimation
of this parameter depends on the regime of operation in the distillation column (spray regime/froth emulsion regime). Expressions
relating the interfacial area per unit bubbling area and various process parameters are shown in Eqs. (K6)(K14):
K4
It is possible to dene the dependence of diffusion coefcient on
molecular weight (Eqs. (K3) and (K4)) and conduct an extensive
analysis with different diffusion coefcients and mass transfer coefcients for various components. However in this study, acetone was
assumed to be the primary component for calculation of liquid diffusion coefcient using Eqs. (K3) and (K4).
The reduced temperature used in Eq. (K3) was calculated using
the average column temperature of 473.0 K and the critical temperature of carbon dioxide (=304.1 K). On substituting these values
of average column temperature and reduced temperature in Eq.
(K5), the value of reduced temperature for Eq. (K3) turns out to
be 1.56.
FPj
bj
K7
V
L
Mass Flowj qj
lengthwj
K8
Abubbling
j
L
0:5
!0:5
0:25
pitchj FPj
bj
Ahole
j
bubbling
Aj
K10
L
K9
L
bj height j
K11
K12
Interfacial area per unit bubbling area (m2 m2) is given by:
Spray regime:
ahfj
40
b0:3
!0:37
2
L
uSVj qVj height j FPj
rj
K13
ahfj
43
b0:3
!0:53
2
L
uSVj qVj height j FPj
rj
K14
Table 2 indicates the parameters used in the calculation of the interfacial area per unit bubbling area in Eqs. (K6)(K14). While using the
L
parameters of Table 2, the value of FPj 3:0 bj heightj turns out to
0.24, indicating presence of spray regime in the reactive distillation
column. Calculations of the interfacial area yield a value of
27.45 m2 m2 for ahfj . When the interfacial area per unit bubbling
area (1.50 m2) is adjusted for base case average liquid stage volume
(0.39 m3), the value of the interfacial area per unit liquid volume
takes a value of 106.11 m2 m3. This value of interfacial area per unit
volume of liquid results in the value of kL a(i,j) = 106.11
m2 m3 2.31 105 m s1 = 2.45 103 s1 = 8.81 h1.
Simulations of reactive distillation for isophorone production
indicate that steady state convergence is attained when the values
of kL a(i,j) used were in the range of 1.253.50 h1. Hence, for analysis purposes, a value of kLa(i,j) = 1.25 h1 was used as base case,
which is described in the Section 3. The maximum value of kLa(i,j)
set to 3.50 h1 in various simulations, was 43.20% and 39.71% of
the corresponding values for the KATAPAK structured packing
and sieve trays respectively. For purposes of sensitivity analysis
the base case kLa(i,j) was varied in the range of 1.253.50 h1 and
simulations were carried out to study the inuence of the overall
liquid side mass transfer coefcient on the mass ratio of acetone
consumed to isophorone produced in the reactive distillation
column.
360
P.J. Darda, V.V. Ranade / Chemical Engineering Journal 207208 (2012) 349367
Table 3a
Base case simulation parameters.
Reaction
k1 (2)
k3 (1)
k5 (2)
k7 (2)
k10 (2)
k11 (2)
k12 (2)
Water
Acetone
Isophorone
Hydroxide catalyst
Mesityl oxide
Isoxylitone
Diacetone alcohol
0.72
2.78 101
1.20 103
1.30 103
0.0
0.0
0.0
1.0
0.0
0.0
0.0
0.0
0.0
0.0
P Total
15 bar
j
54:19 (kmol h
1
Component/Antoine constants
A(i)
B(i)
C(i)
Water
Acetone
Isophorone
Hydroxide catalyst
Mesityl oxide
Isoxylitone
Diacetone alcohol
8.14
7.63
8.59
0.0
10.98
0.0
7.91
1810.94
1566.69
2602.97
0.0
2991.21
0.0
2036.93
244.49
273.42
265.0
0.0
265.0
0.0
236.18
361
P.J. Darda, V.V. Ranade / Chemical Engineering Journal 207208 (2012) 349367
Table 3b
Simulation parameters dening the liquid and vapour phase energy transfer rates for base case simulation.
Liquid phase heat
capacity: C Li
capacity: C Vj
Water
Acetone
Isophorone
312.46
1007.31
2396.99
104.41
374.27
890.60
Hydroxide
catalyst
693.61
257.71
Component
Heat of
vapourisation:
Numerical value
4.17 104
6.08 104
1.45 105
19.69
900.0
41,267.93
4.19 104
66,468.60
(kJ kmol1)
Liquid/vapour side heat transfer
coefcient and heat transfer area:
10,000.0/21,600.0
Mesityl
oxide
1702.09
632.41
1.03 105
Isoxylitone
2396.99
890.60
1.45 105
Diacetone
alcohol
2014.62
748.53
1.22 105
HLf
1
(kJ kmol1)
1
6
Water
11
16
Distillation
Column
21
Stages
[]
26
Isophorone
Acetone
Isoxylitone
Mesityl Oxide
Hydroxide
31
36
41
Fig. 3. Variation of liquid phase composition along the column at steady state.
(4) multicomponent vapour liquid equilibrium (VLE) of the reaction system under consideration.
A sensitivity analysis was carried out to study the inuence of
overall liquid side mass transfer coefcient and average liquid
stage volume on the performance of the isophorone reactive distillation column. Appropriate range of diffusion coefcient and overall liquid side mass transfer coefcient were identied (Section
2.3.3) and varied in the sensitivity analysis. Simulations were
carried out by increasing values of the overall liquid side mass
transfer coefcients in the range of 1.25 (base case value)
3.50 h1. This change in the overall liquid side mass transfer coefcient amounted to a maximum increase of +180% in the base case
value. In addition to the overall liquid side mass transfer coefcient, the average liquid volume of various stages was increased/
decreased by changing the ratio of stage liquid input/output molar
ows to corresponding liquid molar accumulation rates (Appendix
A: Eqs. (A.2) and (A.3)). The base case average liquid stage volume
of 0.39 m3 was consequently varied in the range of 32.88% and
+68.04%, and the resulting average liquid stage volume values in
the range of 0.26 m3 and 0.65 m3 were used in the sensitivity analysis. The liquid stage volumes for various stages were used in conjunction with the volumetric ows for the corresponding stages in
order to obtain stage wise and overall residence time values in the
distillation column reactor. Fig. 6 shows that the key mass ratio of
acetone consumed to isophorone produced increases with the
overall residence time in the distillation column reactor and with
overall liquid side mass transfer coefcient. Increase in the overall
liquid side mass transfer coefcient increases the average liquid
stage volume and overall consumption of acetone while leading
to higher value of the key mass ratio. Analysis of simulations
(including base case simulation) with similar volumetric ow
parameters indicates that an increase in the value of overall liquid
side mass transfer coefcient in the range of 1.253.50 h1 successively reduces the average relative values of liquid phase mole
fractions of acetone and mesityl oxide. This reduction causes
reduction in selectivity towards isophorone. In addition, increase
in mass transfer coefcient also increases liquid ow within the
column (at set value of reux ratio, R) and therefore increases
the average liquid stage volume. The combined effect of these
two factors results in increase in the mass of acetone consumed
per unit mass of isophorone produced with increase in the overall
liquid side mass transfer coefcient.
In addition to the design parameters of volume (residence time)
and overall liquid side mass transfer coefcient, vapour liquid
362
P.J. Darda, V.V. Ranade / Chemical Engineering Journal 207208 (2012) 349367
1
6
Water
11
Isophorone
16
Distillation
Column
Stages
[]
Acetone
21
Mesityl Oxide
26
31
36
41
Fig. 4. Variation of vapour phase composition along the column at steady state.
1
6
11
Liquid Phase
Temperature
16
Vapour Phase
Temperature
Distillation
Column 21
Stages
[]
Bubble Point
Temperature
Dew Point
Temperature
26
31
36
41
Fig. 5. Temperature proles along the column.
Apart from the key issues of mass transfer, average liquid stage
volume (residence time) and specic VLE parameters, addressing
the conguration and operational aspects of the process can also
help gain insights into making improvements in the performance
of the reactive distillation column reactor. The reactive distillation
modelling framework in this study was modied to study the inuence of conguration and operational aspects including (1) feed
composition, (2) feed location, and (3) liquid enthalpy introduced
in the column.
Figs. 7a and 7b show the inuence of variation in the feed composition and feed location on the relative values of the mass ratio
of acetone consumed to isophorone produced and average liquid
stage volume compared to the base case simulation.
When the feed acetone mole fraction is increased in the range
of: base case -20%, and equal water/acetone mole fraction of
363
P.J. Darda, V.V. Ranade / Chemical Engineering Journal 207208 (2012) 349367
1.30
2.20
1.25
1.20
1.15
Relative values
of the ratio of acetone
1.10
consumed to
isophorone produced
and average liquid stage 1.05
volume compared to
base case
1.00
[ ]
Ratio of
2.00
Acetone
Consumed to
Isophorone
Produced 1.80
[ ]
0.95
1.60
0.90
1
1.40
1.50
FC = 20% xA
1.75
2.00
2.25
2.50
FC = +0% xA
2.75
FC = + 20% x A
FC =Equal x A, xW
1.04
1.02
1.00
Relative values
of the ratio of acetone
consumed to
isophorone produced
and average liquid stage
volume compared to
base case
[]
0.98
0.96
0.94
0.92
0.90
1
FL= 2
2
FL= 4
3
FL= 6
4
FL= 9
364
P.J. Darda, V.V. Ranade / Chemical Engineering Journal 207208 (2012) 349367
1
6
Input Liquid
Enthalpy = 90%
of Base Case
11
16
Distillation
Column
21
Stages
[]
Input Liquid
Enthalpy = 110%
of Base Case
26
31
Input Liquid
Enthalpy = 120%
of Base Case
36
41
Fig. 8. Liquid phase temperatures proles along the column for four cases: (1) 90% of the base case feed liquid enthalpy, (2) base case, (3) 110% of the base case feed liquid
enthalpy and (4) 120% of the base case feed liquid enthalpy.
ide as a catalyst was modelled using liquid and vapour phase molar
and energy balances. The resulting system of differential algebraic
equations (DAEs) was solved using the DAE solver: sundials 2.4.0.
Reactive distillation model simulations were carried out for a 43
stage column with 7 components including: (1) water, (2) acetone,
(3) isophorone, (4) hydroxide catalyst, (5) mesityl oxide, (6) isoxylitone and (7) diacetone alcohol. The results of the simulations:
liquid/vapour phase mole fractions of various species and liquid/
vapour phase temperatures along various stages of the reactive
distillation column were plotted and analyzed. Mass and energy
balances including liquid and vapour ows for all the components
on various stages were attributed to the model balances. Numerical solution while using heat transfer, mass transfer and chemical
reactions on each stage indicated that acetone was consumed in
various stages and, products: isophorone, isoxylitone, mesityl
oxide, diacetone alcohol and water were formed as a result of 7
reactions incorporated in the reaction scheme. Simulations also
indicated that steady state was reached for the liquid phase volumetric holdup while using the stage input/output molar ow
parameters described in Eqs. (A.2) and (A.3).
In this work, a base case simulation was dened with net acetone consumption of 800 kg h1. The base case simulation mass
ratio of acetone consumed to isophorone produced was 1.54. The
model was then used to carry out sensitivity analysis by varying:
(1) average liquid stage volume/residence time (via stage input/
output molar ow parameters) and overall liquid side mass transfer coefcient, (2) feed composition, (3) feed location, and (4) liquid enthalpy fed to the column.
Sensitivity analysis results showed that: (1) the key mass ratio
of acetone consumed to isophorone produced (RatioA/IPH) increases
with increase in the overall reactive distillation column residence
time and overall liquid side mass transfer coefcient, (2) increase
in the mole fraction of acetone in the feed signicantly increases
the values of average liquid stage volume (AvgVolL), while the RatioA/IPH increases marginally, (3) an increase in the feed location
from 29 decreases the value of RatioA/IPH, while AvgVolL is independent of the feed location, and (4) increase in the liquid phase
feed enthalpy is reected in terms of increase in the stage depen-
dent liquid phase temperature and the average liquid phase column temperature.
The reactive distillation model presented here can be used to
provide design recommendations and suggest improvements
while providing appropriate details of the isophorone production
process along various stages of the column. The model presented
here will also be useful starting point for further extensions to include more rened approaches including multi-zonal models.
Appendix A. Model equations for reactive distillation
The model for reactive distillation system consists of a set of 10
differential algebraic equations (DAEs):
1. 6 ordinary differential equations (ODEs) and,
2. 4 algebraic equations (AEs).
The system of equations consists of the following balances: (1)
overall liquid molar balance (ODE), (2) overall vapour molar balance (ODE), (3) component liquid molar balance (ODE), (4) component vapour molar balance (ODE), (5) component liquid mole
fraction (AE), (6) component vapour mole fraction (AE), (7) overall
liquid volume (AE), (8) overall vapour volume (AE), (9) overall liquid phase energy balance (ODE) and (10) overall vapour phase energy balance (ODE). These individual molar balances are outlined
next along with dening equations for bulk liquid/vapour phase
temperatures and dew point/bubble point temperatures.
A.1. Overall liquid molar balance
Overall liquid phase molar balance for stage j:
L
dMj
dt
Lj1 Lj F Lj
0
NC
X
NGL
i;j
i1
1
iPNC Oi;k
i1
L
M
Ek
Y Oi;k C
j
Bzi;k k0k e RT
xi;j C
@ bk
A
h iPNC Oi;k 1
i1
i1 k1
L
V j
NC X
NR B
X
A:1
365
P.J. Darda, V.V. Ranade / Chemical Engineering Journal 207208 (2012) 349367
If :
VmaxLj
<
! Lj 0
If : Volj P VmaxLj ! Lj
dMi;j
dt
If :
L
Volj1
If :
L
Volj1
<
VmaxLj1
A:2
Output=Accumulation dM j
L
ft
LRatioj
dt
! Lj1 0
L
P VmaxLj1 ! Lj1
Output=Accumulation
L
ft
LRatioj1
dM j1
dt
A:3
MaxVol=InitialVol
ILj
NC
X
lLi xinitial
i;j
i1
Output=Accumulation
A:4
dMi;j
R
R1
dt
V
dM 1
A:5
dt
A:6
xi;j
xInterface
xi;j M Lj
i;j
MLi;j
A:17
M Lj
A:7
xInterface
i;j
A:16
F Lj F Lj LRj
NGL
i;j kL ai;j
A:15
VmaxLj LRatioj
NR B
MLj
Ek
X
Y Oi;k C
Bzi;k k0k e RT
xi;j C
@ bk
A
h iPNC Oi;k 1
i1
k1
L
V j
yInterface
i;j
K Distribution
i;j
A:8
yInterface
yBulk
i;j
i;j
A:9
yi;j
M Vi;j
A:18
MVj
dM j
dt
V j1 V j
F Vj
NC
X
NGL
i;j
A:10
L
Volj
i1
NC
X
lLi MLi;j
A:19
i1
If :
V
Volj
VmaxVj
! V j
Output=Accumulation
V
ft
VRatioj
dMj
A:11
dt
If :
V
Volj1
<
VmaxVj1
Volj
NC
X
lVi MVi;j
A:20
i1
! V j1 0
V
Output=Accumulation
V
VRatioj1
If : Volj1 P VmaxVj1 ! V j1 ft
dMj1
dt
A:12
MaxVol=InitialVol
VmaxVj VRatioj
IVj
lVi yinitial
i;j
dt
Lj1
ELj1
M Lj1
Lj
ELj
MLj
L
L
L
F Lj HLf
j Ej OP j QLost j
A:13
A:21
i1
A:22
366
P.J. Darda, V.V. Ranade / Chemical Engineering Journal 207208 (2012) 349367
S
L
PTotal
j yi;j ci;j xi;j P i;T
log10 PSi;T
dt
EVj1
V j1 V
Mj1
EVj
V j V
M j
V
V
V
F Vj HVf
j Ej IP j QLost j
yi;j cLi;j
A:23
j T j
i;j H i;j
A:24
i1
ELj
L
ALj hj
NC
X
Interface
eV
T VD
NGL
j T j
i;j H i;j
A:25
i1
L
T Interface
j
PNC
ALj hj
L
ALj hj
GL e V
i1 N i;j H i;j
L
ALj hj
T LB
j
V
AVj hj
PNC
AVj hj
L
ALj hj
GL e L
i1 N i;j H i;j
V
AVj hj
A:26
e L C L T LB T 0 DHS xi;j
H
i
i;j
j
e V kA C V T VD T 0
H
i
i;j
i
j
A:27
A:28
T Lj
T Vj
M Lj
DH s
T0
L
x
C
i1 i;j i
P
EVj
A
NC
i1 yi;j ki
M Vj
PNC
T0
V
y
C
i1 i;j i
PNC
A:29
A:30
xi;j K Distribution
xi;j
i;j
PSi;200
C
A:33
!
A:34
PTotal
j
A:35
i1
xi;j yi;j
PTotal
j
A:36
cLi;j PSi;T VD
yi;j
S
L
i;j P i;T VD
A:37
PTotal
j
NC
X
yi;j 1
A:38
i1
yi;j xi;j
T VD
j
V
AVj hj
PTotal
j
A:32
NC
X
xi;j 1
i1
i1
V
AVj hj
PSi;T
cLi;j
K Distribution
i;j
NC
X
NC
X
V
eL
T Interface
T LB
NGL
j
i;j H i;j Ej
j
A:31
Bi
Ai
C i T
cLi;j PSi;T LB
A:39
PTotal
j
NC
X
xi;j cLi;j PSi;T LB PTotal
j
A:40
i1
The 10 sets of differential algebraic equations described in this section for reactive distillation column constitute to 4NC + 6 balances
(Table A.1) for each stage. Additional stage equations for liquid/vapour molar outow, gasliquid mass transfer, reaction terms and
calculations for energy balances are also required to complete the
description of the model (Table A.2). Thus for a reactive distillation
model consisting of NSTAGES, the total number of equations
dened as input to the solver amounts to NSTAGES x (4NC + 6). In
case of a 43 stage column (41 reaction/hydrolysis stages + 1 condenser + 1 bottoms stage) with 7 components, total number of
equations becomes 43 34 = 1462. This system of 1462 equations
was solved using the differential algebraic equation solver: Sundials
- 2.4.0 (https://computation.llnl.gov/casc/sundials/main.html). The
solution of the differential algebraic equations is obtained as time
dependent/stage wise liquid and vapour phase mole fraction and
temperature prole. The solver was initialized using the base case
parameters dened in Table 3a and solved in order to obtain steady
state in simulations. Additionally, the non linear equations: (A.37)
and (A.40) dening the dew point and bubble points of the vapour
and liquid phase mixtures were solved using the bisection method
for purposes of numerical solution (http://numericalmethods.eng.
usf.edu).
Table A.1
Stage balances used as input to the differential algebraic equation (DAE) solver for the reactive distillation model.
1
2
3
4
5
6
7
8
9
10
Stage balances
Number of balances
Type of balances
1
1
NC
NC
NC
NC
1
1
1
1
4NC + 6
ODE
ODE
ODEs
ODEs
AEs
AEs
AE
AE
ODE
ODE
Eq.
Eq.
Eq.
Eq.
Eq.
Eq.
Eq.
Eq.
Eq.
Eq.
(A.1)
(A.10)
(A.15)
(A.16)
(A.17)
(A.18)
(A.19)
(A.20)
(A.21)
(A.23)
367
P.J. Darda, V.V. Ranade / Chemical Engineering Journal 207208 (2012) 349367
Table A.2
Additional stage equations for the reactive distillation model.
1
2
3
4
5
6
7
8
9
10
Number of
equations
Type of
equation
Equation number in
Appendix A
1
1
NC
NC NR
1
1
1
AE
AE
AEs
AEs
Implicit AE
Implicit AE
AE
AEs
AE
AEs
References
[1] J.R. Walton, B. Yeomans, Isophorone Production Using a Potassium Hydroxide
Catalyst. US Patent 3,981,918, 1976.
[2] G.S. Salvapati, K.V. Ramanamurty, M. Janardanarao, Selective catalytic selfcondensation of acetone, J. Mol. Catal. 54 (1989) 930.
[3] A. Chaubal, Simulation of Reactive Distillation (Case Study: Isophorone Plant),
M.Tech Project Report, Indian Institute of Technology, Mumbai, 1998.
[4] N.O. Lemcoff, R.E. Cunningham, Kinetics of diacetone alcohol conversion to
mesityl oxide catalyzed by ion exchange resin, J. Catal. 23 (1971) 8192.
[5] R. Bakker, G. Hangx, G. Kwant, H. Maessen, P. Markusse, Reaction and Catalyst
Deactivation Kinetics of the Aldol Condensation of Acetone, DSM Technical
Report, Intelligent Column Internals for Reactive Separations (INTINT) Project
No. GRD1 CT1999 10596, 2001.
[6] G.G. Podrebarac, F.T.T. Ng, G.L. Rempel, A kinetic study of the aldol
condensation of acetone using an anion exchange resin catalyst, Chem. Eng.
Sci. 5217 (1997) 29913002.
[7] C. Huang, L. Yang, F.T.T. Ng, G.L. Rempel, Application of catalytic distillation for
the aldol condensation of acetone: a rate-based model in simulating the
catalytic distillation performance under steady-state operations, Chem. Eng.
Sci. 5319 (1998) 34893499.
[8] P. Kustrowski, D. Sulkowska, L. Chmielarz, A. RafalskaLasocha, B. Dudek, R.
Dziembaj, Inuence of thermal treatment conditions on the activity of
hydrotalcitederived MgAl oxides in the aldol condensation of acetone,
Micropor. Mesopor. Mat. 78 (2005) 1122.
[9] K. Schmitt, J. Disteldorf, G. Baron, Production of Isophorone, US Patent
3,337,633, 1967.
[10] C. Cane, B. Yeomans, Process for the Production of Isophorone, US Patent
4,059,632, 1977.
[11] M.C. Grebinoski, D. Glassman, C.L. Elias, A.A. Schutz, Isophorone Process, US
Patent 5,352,839, 1994.
[12] A.A. Schutz, Basic Mixed Oxide, US Patent 4,970,191, 1990.
[13] A.A. Schutz, L.A. Cullo, Process for making Efcient Anionic Clay Catalyst,
Catalysts made thereby, and method of making Isophorone, US Patent
5,153,156, 1992.
[14] J.B. Cook, Industrial Application of Reduce Reuse Recycle Strategies in the
Production of Isophorone, MS Thesis, University of South Alabama, Mobile,
2006.
[15] R. Taylor, R. Krishna, Modeling of homogeneous and heterogeneous reactive
distillation processes, in: K. Sundmacher, A. Kienle (Eds.), Reactive
Distillation: Status and Future Directions, Wiley-VCH Verlag GmbH & Co.
KGaA, 2002, pp. 217240.
[16] A. Katariya, R. Kamath, K. Moudgalya, S. Mahajani, Non-equilibrium stage
modeling and non-linear dynamic effects in the synthesis of TAME by reactive
distillation, Comput. Chem. Eng. 32 (2008) 22432255.
[17] Y. Xu, Y. Zheng, F.T.T. Ng, G.L. Rempel, A three-phase nonequilibrium dynamic
model for catalytic distillation, Chem. Eng. Sci. 60 (2005) 56375647.
[18] K. Sundmacher, U. Hoffmann, Development of a new catalytic distillation
process for fuel ethers via a detailed nonequilibrium model, Chem. Eng. Sci.
5110 (1996) 23592368.
[19] K. Sundmacher, U. Hoffmann, Oscillatory vapourliquid transport phenomena
in a packed reactive distillation column for a fuel ether production, Chem. Eng.
J. 57 (1995) 219228.
[20] K. Alejski, F. Duprat, Dynamic simulation of the multicomponent reactive
distillation, Chem. Eng. Sci. 5118 (1996) 42374252.
[21] B. Hupen, E.Y. Kenig, Rigorous modelling of NOx absorption in tray and packed
columns, Chem. Eng. Sci. 60 (2005) 64626471.
[22] K. Alejski, Computation of the reacting distillation column using a liquid
mixing model on the plates, Comput. Chem. Eng. 155 (1991) 313
323.
[23] K. Alejski, J. Szymanowski, M. Bogacki, The application of a minimization
method for solving reactingdistillation problems, Comput. Chem. Eng. 128
(1988) 833839.
[24] Z. Qi, K. Sundmacher, E. Stein, A. Kienle, A. Kolah, Reactive separation of
isobutene from C4 crack fractions by catalytic distillation processes, Sep. Purif.
Technol. 26 (2002) 147163.
[25] S. Gruner, K.D. Mohl, A. Kienle, E.D. Gilles, G. Fernholz, M. Friedrich, Nonlinear
control of a reactive distillation column, Control Eng. Pract. 11 (2003) 915
925.
[26] R. Krishnamurthy, R. Taylor, A nonequilibrium stage model of multicomponent
separation processes. Part I: model description and method of solution, AIChE
J. 31 (1985) 449456.
[27] R. Taylor, R. Krishna, Modeling reactive distillation, Chem. Eng. Sci. 55 (2000)
51835229.
[28] R. Baur, Modeling Reactive Distillation Dynamics, PhD Thesis, Clarkson
University, Potsdam, 2000.
[29] R. Baur, A.P. Higler, R. Taylor, R. Krishna, Comparison of equilibrium stage and
nonequilibrium stage models for reactive distillation, Chem. Eng. J. 76 (2000)
3347.
[30] S. Sandler, Chemical, Biochemical and Engineering Thermodynamics, fourth
ed., Wiley, New Delhi, 2010.
[31] N. Dalaouti, P. Seferlis, A unied modeling framework for the optimal design
and dynamic simulation of staged reactive separation processes, Comput.
Chem. Eng. 30 (2006) 12641277.
[32] C. Noeres, A. Hoffmann, A. Gorak, Reactive distillation: non-ideal ow
behaviour of the liquid phase in structured catalytic packings, Chem. Eng.
Sci. 57 (2002) 15451549.
[33] E.F. Wijn, Weir ow and liquid height on sieve and valve trays, Chem. Eng. J. 73
(1999) 191204.
[34] Y.H. Jhon, T. Lee, Dynamic simulation for reactive distillation with ETBE
synthesis, Sep. Purif. Technol. 31 (2003) 301317.
[35] M. Banaji, P. Donnell, S. Baigent, P Matrix Properties, Injectivity and Stability in
Chemical Reaction Systems, University College London.
[36] S. Thotla, V. Agarwal, S.M. Mahajani, Simultaneous production of diacetone
alcohol and mesityl oxide from acetone using reactive distillation, Chem. Eng.
Sci. 62 (2007) 55675574.
[37] Y.K. Kim, J.D. Hateld, Kinetics and equilibrium data of the dehydration
hydration reaction between diacetone alcohol and mesityl oxide in phosphoric
acid, J. Chem. Eng. Data 30 (1985) 149153.
[38] H. Watanabe, J. Seto, Catalytic effect of barium ferrite magnetic particles on
aldol condensation of acetone, Appl. Catal. A-Gen. 134 (1996) 217224.
[39] Z. Svandova, J. Markos and L. Jelemensky, Impact of mass transfer on modelling
and simulation of reactive distillation columns, in: Mohamed ElAmin (Eds.),
Mass Transfer in Multiphase Systems and its Applications, InTech, 2011, pp.
649676.
[40] F.J. Zuiderweg, Sieve trays: a view on the state of the art, Chem. Eng. Sci. 3710
(1982) 14411464.
[41] S. Umezawa, A. Nagashima, Measurement of the diffusion coefcients of
acetone, benzene, and alkane in supercritical CO2 by the taylor dispersion
method, J. Supercrit. Fluid. 5 (1992) 242250.