chemical engineering research and design 8 7 ( 2 0 0 9 ) 4760

Contents lists available at ScienceDirect

Chemical Engineering Research and Design

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

Retrotting conventional column systems to

dividing-Wall Columns
R. Premkumar, G.P. Rangaiah
Department of Chemical & Biomolecular Engineering, National University of Singapore, Engineering Drive 4,
Singapore 117576, Republic of Singapore

a b s t r a c t
Distillation, the most common separation process in chemical process industries, requires signicant energy inputs.
Dividing-Wall Column (DWC), which works on the basis of Fully Thermally Coupled Distillation System (FTCDS), is
chosen for this study due to its lower energy consumption compared to the conventional column system. The main
objective of this study is to investigate the potential of retrotting conventional 2-column (C2C) systems in operation
for separating ternary mixtures into three products, to DWCs. For this, six applications of industrial importance
are selected and conventional 2-column systems are designed, which are assumed to be currently in operation in
the plants. Then, retrotting these systems to DWC is studied. Results show that retrotting the existing 2-column
systems to DWCs is very attractive both economically and for its reduced energy requirements.
2008 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
Keywords: Dividing-Wall Column; Design; Retrotting; HYSYS; Fully Thermally Coupled Distillation System

1.

Introduction

Distillation columns are used for about 95% of liquid separations and the energy use from this process accounts for an
estimated 3% of worlds energy consumption (Hewitt et al.,
1999). Motivated by this large energy requirement for distillation, researchers have developed various column arrangements that can bring in savings in both energy and capital
cost. Any reduction of energy consumption will not only bring
economical benets but also environmental benets in terms
of reduced usage of fossil fuels and its associated emissions.
Column congurations vary from simple to complex congurations. Simple column congurations refer to direct
sequence, indirect sequence and distributed sequence (Shah,
2002), and are the conventional column systems operated
worldwide today. Complex column arrangements refer to recycling the vapor and/or liquid, heat integration, etc. Thermally
coupled arrangements are realized by setting up two-way
vapor/liquid ows between different columns of the simple column congurations. Reported studies reveal that Fully
Thermally Coupled Distillation System (FTCDS, also called
Petyluk System) provides the maximum energy reduction in
columns. In most cases, this is implemented in the form of

a Dividing-Wall Column (DWC) in which both columns are

housed in a single shell. This reduces not only the energy
consumption but also space and investment requirements
compared to the conventional column system.
The DWC has been known for several decades since its
patent by Wright (1949). Petyluk et al. (1965) introduced
the thermal coupling for separating ternary mixtures and
presented a fully thermally coupled congurationPetyluk
column. Amminudin et al. (2001) noted the industrial acceptance and commercialization of DWC by organizations such
as BASF AG, M.W. Kellogg (together with BP, later known as
BP Amoco), and Sumitomo Heavy Industries Co. together with
Kyowa Yuka. Linde AG constructed the worlds largest DWC
for Sasol, an estimated 107 m tall and 5 m in diameter (Schultz
et al., 2002). Recently, Adrian et al. (2004) reported that BASF
operates about 30 DWCs worldwide in their plants.
Fidkowski and Krolikowski (1987) have established that the
Petyluk system requires the least energy among all the options
for a three-product system. This key advantage is also applicable for a DWC, which has the minimum vapor ow rate for the
particular separation compared to the conventional system.
Triantafyllou and Smith (1992), and Agrawal and Fidkowski
(1998) have shown that, for a mixture of three components,

Corresponding author. Fax: +65 6779 1936.

E-mail address: chegpr@nus.edu.sg (G.P. Rangaiah).
Received 29 February 2008; Received in revised form 25 June 2008; Accepted 27 June 2008
0263-8762/$ see front matter 2008 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.cherd.2008.06.013

48

chemical engineering research and design 8 7 ( 2 0 0 9 ) 4760

the DWC reduces the total vapor ow by 1050% compared

to conventional systems using direct and indirect sequences.
Reduction in vapor ow contributes to lower duties of reboiler
and condenser, and consequently capital and operating costs.
Furthermore, DWC uses only 1 reboiler and 1 condenser when
compared to 2 reboilers and 2 condensers for a conventional 2column (C2C) system. This would add to the savings in capital
as well as operating cost.
Despite the economic edge and lower energy requirements
of DWCs, lack of reliable design methods and concerns on
their operation have prevented their commercial application
(Schultz et al., 2002). Abdul Mutalib and Smith (1998a) noted
that it is impractical to manipulate the vapor split in a DWC
although it is easier to manipulate the liquid split via special
design of the liquid distributor. However, liquid split is also
usually left uncontrolled, and the operating values of both liquid and vapor splits result from the natural balancing of ow
resistances inside the column. Simulation and experimental
studies of Abdul Mutalib and Smith (1998a,b), demonstrate
that the DWC can be operated successfully.
Kolbe and Wenzel (2004) have stated the following average
savings gained by using a DWC against a conventional column system: 25% in investment cost, 35% in operating cost
and 40% in space requirement. The savings in space requirements are due to reduction in number of reboilers, condensers
and associated equipments such as pumps, their supports,
etc. This will be of particular interest in locations such as Singapore where land is limited. Thus, DWC is attractive to many
chemical and related industries in the current scenario of
competition and environmental concerns, in order to reduce
energy usage for distillation. One way is to make use of the
existing equipments and operate the plant more efciently,
possibly with minor modications and small investment. An
investment with payback period of less than 3 years is favorably considered by the management. Retrotting an existing
conventional column system to a DWC has potential for reducing energy and an acceptable payback period.
In this investigation, retrotting conventional columns
to DWCs is analyzed by considering six industrial applications. First, conventional 2-column systems are designed for
these applications. Then, both new and retrotted DWCs
are designed; the latter uses the existing equipment in the
conventional 2-column system as much as possible. Finally,
detailed techno-economic analysis of retrotting for the six
applications is carried out. The process simulator, HYSYS is
employed for simulation and optimization of all columns for
the six applications. A step-by-step procedure for DWC design
using HYSYS is described. Results show that retrotting conventional 2-column systems to DWCs reduce operating cost
by more than 30% with an acceptable payback period, for all
the six applications studied.

2.

Design procedure for DWC using HYSYS

Design of DWCs has been studied in the open literature using

basic equations (e.g., Triantafyllou and Smith, 1992) and using
commercial simulators (e.g., Kim, 2006). Investigators generally use FenskeUnderwoodGilliland (FUG) equations as
part of shortcut method for initialization followed by rigorous simulation with simulators such as HYSYS and Aspen
Plus. Chapter 4 in Seider et al. (2003) provides a good overview
to process simulation using these simulators. The FUG equations are also used in the shortcut method in simulators. Some

Fig. 1 HYSYS ow sheet with three shortcut columns for

simulating a FTCDS.
studies (e.g., Triantafyllou and Smith, 1992; Muralikrishna et
al., 2002; Dunnebier and Pantelides, 1999; Amminudin et al.,
2001) used mathematical models to design the FTCDS/DWC
and then a simulator to conrm the design.
The steps involved in the design and implementation of a
DWC is similar to a conventional column. The feed conditions,
product specications, column operating conditions are available from the plant or the given examples depending upon the
situation. The thermodynamic model to be used for predicting
physical properties may be available from the process team in
the plant. If not, this shall be selected depending on the components in and operating conditions for the application. The
main steps in the design of a FTCDS/DWC are: (a) shortcut distillation for nding initial estimates of variables required for
rigorous simulation, (b) rigorous simulation of the FTCDS, (c)
optimization of the system, and (d) design and sizing of the
DWC; these are outlined below.

2.1.

Shortcut distillation

As noted by Amminudin et al. (2001), three shortcut columns

shown in Fig. 1 are nearly equivalent to a FTCDS in Fig. 2. In this
representation, recycle streams in a FTCDS are not explicit but
are internal; rectifying liquid and stripping vapor in the shortcut column for the prefractionator respectively correspond
to the recycled liquid and vapor streams to column T-100
(Liquid101-Out and Vapor101-Out in Fig. 2). We use the threecolumn model in Fig. 1 for shortcut calculations to obtain the
necessary estimates for rigorous simulation. The column T100 is equivalent to the prefractionator of the FTCDS, and the
columns T-101A and T-101B together are equivalent to the
main column of the FTCDS (see Figs. 2 and 3). Importantly,
the stream Side 1 and Side 2 shall have the same purity so
that their ows can be added; the combined Side stream represents the side stream (side draw in the main column of
FTCDS, see Fig. 2). In the shortcut column T-100 in Fig. 1, the
condenser is a partial condenser, and Distillate 1 is a vapor
stream.
First column T-100 is simulated assuming a percentage of
say 1% lights in Bottom 1 and 1% heavies in Distillate 1. Then,
column T-101A is simulated by specifying the middle component (heavies in Distillate) and top component (lights in Side
1) as per the purity requirements provided in the example or
the process data from the plant. Similarly, T-101B is simulated
with Side 2 (heavies in top) and Bottom (lights, which is the
middle component, in Bottom) specications. The specications of lights in Bottom 1 and/or heavies in Distillate 1 of

chemical engineering research and design 8 7 ( 2 0 0 9 ) 4760

49

Fig. 2 Final (converged) FTCDS by rigorous simulation.

column T-100 are varied by trial and error to arrive at nearly
equal composition of Side 1 and Side 2.

2.2.

Rigorous simulation

Once shortcut estimates are completed, FTCDS has to be simulated rigorously using HYSYS (Fig. 2) or some other way, to
get accurate results. The approaches for rigorous simulation
involve less number of assumptions and are more realistic
than FUG equations; they are: equilibrium-stage models and
rate-based models. The former uses mass, equilibrium, summation of mole fractions and enthalpy (MESH) equations for
each stage. The resulting set of large number of nonlinear
equations is solved using a suitable numerical method. The
rate-based models use mass transfer rates. Equilibrium-stage
models are more common in simulators and are generally adequate for nearly ideal distillation systems (Seider et al., 2003).
Rigorous simulation used in the present study is based on the
equilibrium-stage model.
This section presents the steps involved in rigorous simulation. First, open a new case/le in the process simulator such
as HYSYS, add the components involved in the process, choose
the thermodynamic models (for predicting required physical
properties including phase behaviour) and proceed to the simulation environment. The FTCDS has two recycle loops and
hence requires two recycle blocks for its simulation as shown
in Fig. 2. The placement of recycle blocks is guided by the availability of initial estimates for Vapor101-In and Liquid101-In

Fig. 3 Similarity of vapor ows in FTCDS and DWC; PF is

the prefractionator.

(shown in Fig. 2) from shortcut calculation. Note that estimates

for Liquid101-Out and Vapor101-Out are not available; hence,
T-100 cannot be solved since these 2 streams are the additional
feeds to it.
The prefractionator does not have a reboiler and condenser,
and can be simulated by an absorber column. Hence, select
and add the absorber column into the ow sheet as T-100
(Fig. 2). Then, add distillation unit for T-101, and the material streams Feed, Vapor100-Out, Liquid100-Out, Vapor101-Out
and Liquid101-Out for column T-100, and Distillate, Side and
Bottom for column T-101 as shown in Fig. 3. To initialize
the recycle stream, a recycle block has to be added to the
ow sheet: one RCY-1 for Vapor100-Out and another RCY2 for Liquid100-Out. Vapor100-Out and Vapor101-In shall be
connected as inlet and outlet to RCY-1 respectively. Similarly
Liquid100-Out and Liquid101-In shall be connected as inlet
and outlet to RCY-2 respectively.
Provide the data from shortcut estimates to the Vapor101In and Liquid101-In. Dene the columns by the connecting
the necessary streams, providing the number of stages, stage
location for Vapor101-In, Liquid101-In, Side, Liquid101-Out
and Vapor101-Out. The stage number for Liquid101-Out is the
same as Vapor101-In entry stage. Similarly, Vapor101-Out has
the same stage number as Liquid101-In. The shortcut calculations give an estimate for the number of stages for T-101A
and T-101B, which are equivalent to the main column T-101 in
rigorous simulation. The feed stage of Distillate 1 and Bottom
1 (both from shortcut estimation) is taken respectively as the
feed stage for Vapor101-In and Liquid101-In in the rigorous
simulation. Similarly, Side draw location was selected based
on shortcut calculations. The complete the ow sheet is as
shown in Fig. 2.
Column T-101 has 5 degrees of freedom (DOF) and hence
5 specications shall be provided. This is in the context of
simulation given conditions of all feeds, stage/tray pressures
in the column, number of stages, feed stage and condenser.
A typical column with a total condenser, reboiler, distillate
and bottoms products has 2 DOF (e.g., see Seider et al., 2003);
each side product adds 1 to DOF. Since T-101 has 3 side products (Vapor101-Out, Side and Liquid101-Out), its DOF 5 (=2 + 3),
which can be used for specifying product purities and/or ow
rates.
Provide the product purity in each stream: namely, light
component purity in Distillate, middle component purity in
Side, heavy component purity in Bottom, as 3 specications.
The ow rate of Liquid101-Out and Vapor101-Out shall be
added as the remaining 2 specications. Activate simulation
calculations and converge T-101. Now, streams Liquid101-Out
and Vapor101-Out are calculated and available for further cal-

50

chemical engineering research and design 8 7 ( 2 0 0 9 ) 4760

culations. Column T-100 has no degree of freedom and so no

specication to be given. This column is analogous to T-100
simulated via shortcut calculations. Hence, provide the number of stages, feed location and pressure prole, etc. from the
initial estimates obtained from shortcut estimation. The number of stages in T-100 (i.e., prefractionator) is assumed to be the
same as the number of stages along the dividing wall, and is
estimated based on the sum of the number of stages below the
feed stage in T-101A and the number of stages above the feed
stage in T-101B (see Fig. 1). The complete ow sheet should
converge after several iterations due to the two recycle blocks.

2.3.

Optimization

The design obtained by shortcut calculations and rigorous

simulation is generally not the optimum, and hence it needs
to be optimized. The studies of Lek et al. (2004) and Rangaiah
et al. (2006) on distillation columns show that operating cost
is the major part of the total annual cost and that minimizing reboiler/condenser duties (or equivalently reux ratio)
minimizes operating cost leading to nearly optimal design
of columns. Hence, the FTCDS design is optimized by minimizing the reboiler duty. The design variables affecting this
duty are: (1) Liquid101-Out draw rate, (2) Vapor101-Out draw
rate, (3) Feed location of Vapor101-In to T-101, (4) Feed location of Liquid101-In to T-101, (5) Side draw location, (6) Feed
location of T-100, (7) No. of stages in column T-100 and (8)
No. of stages in column T-101. Optimization of the FTCDS
is carried out via a four-step procedure outlined below, by
varying each of the rst six variables systematically and
sequentially. The last two variables (number of stages in T100 and T-101 given by shortcut calculations) are not varied
as their effect is expected to be marginal based on the study
of Rangaiah et al. (2006). Hence, the resulting design after the
four steps is close to the optimum but may not be the global
optimum.
Step 1Optimize feed location of Vapor101-In to column T-101:
Recall that the feed stage for Vapor101-In is the same as
that for drawing out Liquid101-Out; similarly, the feed stage
Liquid101-In is the same as that for drawing out Vapor101Out. Feed stage for Vapor101-In is varied in suitable steps to
nd the optimum feed location to column T-101. For each
chosen Vapor101-In feed stage, ow rates of Vapor101-Out
and Liquid101-Out are varied and optimized for the lowest
reboiler duty.
Step 2Optimize feed location of Liquid101-In to column T-101:
Starting from the optimal solution in step 1, feed location
of Liquid101-In is varied; for each selected feed location, the
draw rates of Vapor101-Out and Liquid101-Out are varied to
achieve the lowest reboiler duty. The data corresponding to
the lowest reboiler duty will be taken as the optimum at the
end of this step.
Step 3Optimize side draw location in column T-101: Next, draw
location of Side stream is varied. For each selected draw location, the draw rate of Vapor101-Out and Liquid101-Out has

been varied, and the solution corresponding to the lowest

reboiler duty is taken as the optimum and carried forward to
the next step.
Step 4Optimize feed location of column T-100: The feed location
of column T-100 is varied, and, for each chosen feed location,
the draw rates of Vapor101-Out and Liquid101-Out in T-101
are varied to nd the design with the lowest reboiler duty.
The effect of the optimization steps on reboiler duty is
shown in Table 1 for three applications discussed later. It can
be seen from this table that there is a considerable reduction in
the reboiler duty during the rst step. Subsequent steps have
minimal effect which indicates that the reboiler duty is close
to the minimum after the 4 steps.

2.4.

Sizing of column, condenser and reboiler for DWC

The next step after the optimization is to size the column,

reboiler and condenser for the DWC. The simulation results
for the optimized FTCDS shall be used for DWC as the concentration prole, heat duties and performance data for both of
them are identical. The sizing calculations for the columns in
a DWC are similar to the conventional 2-column system. However, there is slight difference in sizing the DWC as it houses
the prefractionator as well, and is described below.
In general, column diameter should be sufcient to handle
the respective maximum vapor and liquid ow rates in the
column, and depends on mainly vapor ow rate. This will also
ensure that the pressure drop is in the acceptable range. Column diameter is determined by the ooding condition of the
column that xes the upper limit on vapor velocity. The operating velocity is normally between 70 and 90% of the ooding
velocity (Sinnott, 2005). In this study, 80% of the ooding velocity was used as the operating vapor velocity, Vact . The ooding
vapor velocity, Vmax can be estimated from the correlation:


Vmax = K1

L v
v

(1)

where K1 is a constant which is taken to be 0.07 m/s. The operating velocity is

Vact = 0.8 Vmax

(2)

The column diameter, D shall be calculated by:


D=

4G
vap Vact

(3)

where G is the vapor ow rate in kg/s, and vap is the vapor

density in kg/m3 . The column is usually fabricated in increments of 0.5 ft; so, the calculated diameter is rounded up to
the nearest 0.5 ft. This results in a lower vapor velocity and
hence forms a more conservative estimate. The factor K1 in
Eq. (1) is for sieve trays. Other internals are downcomers, feed
distributors and liquid/vapor draw-off pipes. Cost of these is

Table 1 Reboiler duty at different steps in the optimization of the DWC

Example
BTX separation
Depropanizer/debutanizer
Alkanes separation

Prior to optimization
1,531
33,660
1,053

After step 1

After step 2

After step 3

After step 4

1367
9436
952

1357
9435
934

1355
9409
934

1355
9409
934

chemical engineering research and design 8 7 ( 2 0 0 9 ) 4760

minor compared to that of column shell and trays, and hence

it is neglected.
Fig. 3 shows the similarity of the vapor ow rates between
FTCDS and DWC. Both the main column of the FTCDS and
the DWC is divided into three sections: top, middle and bottom sections. In the FTCDS, vapor ow rate Vt represents the
vapor load in the top section, Vb the vapor load in the bottom section, V1 the vapor ow to the prefractionator, V2 the
vapor ow to the main column middle section, V3 the vapor
ow from the prefractionator to the main column, V4 vapor
ow from the main column middle section to its top section.
When the prefractionator is housed within a single shell to
form the DWC, the corresponding vapor ow rates apply.
It is obvious from Fig. 3 that V1 + V2 = Vb and V3 + V4 = Vt .
Diameter of the prefractionator (D1 ) and of the main column
(D2 ) of the FTCDS shall be arrived as per the conventional column sizing. Cross-sectional area of the middle section of the
DWC is then the sum of the area of the prefractionator and
the main column, which can be used to estimate the diameter of the middle section of the DWC. Nevertheless, the DWC
should be sized for the maximum vapor load in each of the top,
middle and bottoms sections, available from rigorous simulation. Diameter of each of the three sections shall be calculated
based on vapor rate similar to the conventional column. The
middle and bottom sections will be usually the same as or
bigger than the top section. Depending upon the change in
diameter, one can choose either to use a single diameter or
step diameter (top section with a smaller diameter). In the former case, the larger diameter of all the three sections should
be used.
The cost correlations for column, trays, reboiler and condenser are taken from Turton et al. (2003). The costing consists
of estimating capital and operating costs. The capital cost for
the distillation system consists of total module cost of column(s), reboiler and condenser. The costing for number of
trays is taken with respect to the higher number of stages
assuming 70% efciency. The cost correlations should be
updated with the current cost index; for this study, the Chemical Engineering Plant Cost Index (CEPCI) of 560 was used. The
total module cost is 18% more than the bare module cost of
the column, trays, reboiler and condenser, which accounts for
all the cost for the associated instruments, piping, structures,
etc. (Turton et al., 2003). Operating cost is the sum of the cost
of steam (for reboilers) and cooling medium (for condensers).
The electricity, which is required for circulation pumps, is
small and neglected. An Excel program was prepared to perform the sizing of the columns, trays, reboilers and condensers
followed by costing for the equipments and utilities.

3.

Results and discussion

A search of the open literature had revealed that several applications of DWCs were successfully studied as new design. A
few other applications involve a side draw stream in a single
column; these can be tried in a DWC for efcient separation. All these examples are summarized in Table 2. Out of
these applications and from our own industry source, 6 were
selected to investigate and discuss the techno-economic evaluation of retrotting conventional 2-column systems to DWCs.
Design conditions and specications for these applications are
listed in Table 3, which form the basis for the study.
In this study, three columnsC2C, new DWC and
retrotted DWC (R-DWC) are designed for each selected

51

application. C2C has the nearly optimal design for a new

conventional column; however, in the actual industrial application, it will be the existing columns in the plant, identied
for retrotting evaluation. The conguration can either be
direct or indirect sequence depending upon the product specications. For retrotting the C2C to a DWC (R-DWC), the
existing column is checked for the design parameters and
vapor ow rate of DWC; this includes checking to see if the
ratio of the actual vapor velocity to maximum/ooding vapor
velocity is in the acceptable range. This is the deciding factor
for retrotting one of the two columns in the C2C as a DWC.
Similarly, existing reboilers and condensers should be checked
for re-use with minimal modications.
The selected, existing column is modied with the addition
of the middle section, which houses the dividing wall. It will be
cut at site into 2 portions to form the top and bottom sections
of R-DWC. Middle section of R-DWC is shop fabricated and
shall be available before the actual retrotting at site starts.
This section is rst welded to the top of the bottom section
and then the removed top portion will be welded on top of
the middle section to complete the column. In this exercise,
number of stages in the existing column shall be sufcient
to meet the required stages in top and bottom sections of the
DWC. In essence, the existing column is used fully leaving with
the column modication of adding the middle section only.
During the shutdown of the existing plant, the piping and
other equipments around the columns are disconnected from
the column. Subsequently, the trays installation would be carried out at site. As per the industrial practice, trays and column
internals are installed in the column in vertical position (i.e.,
after installation of the column at site). However, in the recent
past, trays have been installed in shop or at site (i.e., column in
horizontal position) before installation of the column. This is
due to the time constraint, which industries are facing; extra
time for site installation means late startup and associated
production loss. Before the column installation, trays should
be well secured to avoid dislocation during transport and/or
installation of the column. The necessary testing will be carried out to satisfy the safety and local statutory regulation
requirements in force. In case there is no place around the
column to effect the modications, then the main column
will be removed from the existing location to another location
where the necessary modications and testing can be done.
The reboilers and condensers are re-used and/or new ones
installed to complete the retrotting. Practical considerations
in retrotting are outlined in Appendix A.

3.1.

Benzene, toluene, xylene (BTX) application

Table 4 provides the comparison of C2C, new DWC and R-DWC

for the BTX application. The C2C system has 2 columns of
diameter 0.9 and 0.8 m, with 31 and 40 stages respectively. The
DWC system has only one column with diameter of 0.9 m. In
R-DWC, the existing column 1 (C-100) from the plant is used
since the diameter requirement is met. The number of stages
in the existing column is not sufcient to satisfy the total
stages in the top and bottom section of the DWC. Hence, a
column section with 2 trays is shop fabricated and added to
the top section during the retrotting. The middle section is
separately shop fabricated and assembled at site. The existing
condenser and reboiler can be re-used. Hence, costing shows
the cost for additional middle section and trays. Such an addition of the section can take place in a short duration of 510
days depending upon the site condition. From Table 4, it can

52

Table 2 Reported and potential applications of DWC in the open literature

Column conguration (s)

Kaibel (1987)

DWC and a conventional

column conguration

Triantafyllou and Smith (1992)

FTCDS and DWC

Fidkowski et al. (1993)

Abdul Mutalib and Smith (1998a,b)

Conventional column
congurationdirect
sequence
DWC

Agrawal and Fidkowski (1998)

FTCDS

Dunnebier and Pantelides (1999)

Hairston et al. (1999)

DWC

Amminudin et al. (2001)

FTCDS and DWC

Rev et al. (2001)

FTCDS and conventional

column system

Shah (2002)

FTCDS and DWC

Muralikrishna et al. (2002)

Schultz et al. (2002)

DWC
DWC

Blancarte-Palacios et al. (2003),

Harlvorsen and Skogestad
(2003), and Jimenez et al. (2003)

FTCDS

Adrian et al. (2004)

Bruggemann and Marquardt (2004)
Kolbe and Wenzel (2004)

DWC
FTCDS
DWC

Kim (2006)

FTCDS and DWC

Rangaiah et al. (2006)

Conventional column
with a side product

Description/details
Application studied is separating a mixture of n-hexane, n-heptane and
n-octane. DWC is compared with the conventional column conguration of
direct sequence.
Application used is separation of close-boiling mixtures of C4sseparate
1-butene from the feed mixture containing i-butane,1-butene, n-butane,
trans-2-butene and cis-2-butene.
Applications cited as examples are separation of (1)
acetaldehydemethanolwater, (2) acetonechloroformbenzene and (3)
ethanolwaterethylene glycol system
Application in the rst paper is the separation of methanol, iso-propanol and
butanol. The second paper deals with simulation and pilot plant studies for the
same application.
Application considered is the separation of air to produce argon for FTCDS.

In addition to the application used by Triantafyllou and Smith (1992), a case

study on separation of alkane mixtures containing 2-methylbutane, pentane,
hexane and heptane with DWC is discussed.
Application referred to is the separation of pure ethyl acetate from the mixture
of lower and higher boiling impurities developed by Sumitomo for Kyowa Yuka
Ltd., Japan.
Case study is the replacement of a conventional depropanizer and debutanizer
by a DWC.
Case study is the separation of a mixture of ethanol, n-propanol and n-butanol
for comparing energy consumption and cost advantages of FTCDS system with
the conventional column system with or without heat integration.
Separation of LPG from mixed pentanes and a heavy end in a light-end renery
distillation train.
Separation of equimolar mixture of benzene, toluene and o-xylene.
One application suggested is the linear alkyl benzene (LAB) complex and
another is pre-fractionation of kerosene within a LAB complex.
Blancarte-Palacios et al. (2003) considered a mixture of n-pentane, n-hexane,
n-heptane and n-octane using a FTCDS. Harlvorsen and Skogestad (2003), and
Jiminez et al. (2003) considered a mixture n-pentane, n-hexane and n-heptane
only.
Separation of a mixture with butanol (15%), pentanol (70%) and hexanol (15%).
Separation of benzene, toluene and ethyl benzene (BTE) mixture.
Reduction of benzene in gasoline and separation of benzene, toluene and xylene
(BTX) mixture.
Two applications referred here. One is the fractionation of BTX and the other is
the gas concentration process to produce gas products from a gas mixture
containing methane, ethane, propene (in lights), propane (as middle
component), i-butane, i-butene, n-butane, i-pentane and n-pentane in heavies.
Three applications: (1) mixture of benzene, toluene and biphenyl, (2) mixture of
alkanes, and (3) mixture of benzene, toulene and p-xylene.

Remarks

Concluded that DWC will have capital cost savings

in addition to all the savings realized by FTCDS.
These applications are potential for using and
evaluating DWC.

Concluded that partial thermal coupling is

advantageous compared to FTCDS for cryogenic
applications.

Application commercialized with a DWC.

Discussed retrotting an existing column and

benets achieved.
BTX production is one of the largest energy
consumers in a cracker plant.

Potential applications for DWC.

chemical engineering research and design 8 7 ( 2 0 0 9 ) 4760

Reference

Table 3 Characteristics of the applications studied

No.

Components

BTX separation
Benzene

Specications of top, side and

bottom products (mole%)

Other conditions

Reference

100 kg mol/h

Benzene: 99.5%

Column pressure: 10 atm

0.33
0.34

10 atm
Saturated liquid

Toluene: 91%
p-Xylene: 92%

Total condenser
PengRobinson Model

BTE separation
Benzene
Toluene
Ethyl Benzene

0.33
0.33
0.34

100 kg mol/h
1.75 bar
Saturated liquid

Benzene: 99.5%
Toluene: 96%
Ethyl benzene: 96%

Column pressure: 1.75 bar

Total condenser
PengRobinson Model

From our industry source

Depropanizer/debutanizer
Ethylene
Propene
Propane
i-Butane
i-Butene
n-Butane
i-Pentane
n-Pentane
n-Hexane

0.0128
0.076
0.2312
0.1443
0.2683
0.0409
0.094
0.1008
0.0317

1600 kg mol/h
14.9 bar
90% liquid

Recovery of
n-Propane: 94%
Butane products: 95%
Pentane products: 97%

Column pressure: 14.0 bar

Total condenser
MP steam at 5 bar
Cooling water at 25 C
SRK Model

Amminudin et al. (2001)

EWE separation
Ethanol
Water
Ethyl glycol

0.33
0.33
0.34

100 kg mol/h
1.5 bar
Saturated liquid

Ethanol: 81.8%
Water: 99.6%
EG: 84.6%

Column pressure: 1 atm

Total condenser
NRTL Model

From our industry source

EPB separation
Ethanol
1-Propanol
1-Butanol

0.1
0.8
0.1

300 kg mol/h
1 atm
Saturated liquid

Ethanol: 99%
1-Propanol: 99%
1-Butanol: 99%

Column pressure: 1 atm

Total condenser
NRTL Model

Emtir et al. (2001)

Alkanes separation
n-Pentane

0.34

100 kg mol/h

n-Pentane: 99.5%

Column pressure: 500 kPa

Bek-Pedersen and Gani

(2004), Rangaiah et al. (2006)

0.33
0.33

510 kPa
Saturated liquid

n-Hexane: 87.5%
n-Heptane: 88%

Total condenser
PengRobinson Model

n-Hexane
n-Heptane

Bek-Pedersen and Gani

(2004), Rangaiah et al. (2006)
chemical engineering research and design 8 7 ( 2 0 0 9 ) 4760

Feed conditions

0.33

Toluene
p-Xylene
2

Feed composition
(mole fraction)

53

54

chemical engineering research and design 8 7 ( 2 0 0 9 ) 4760

Table 4 Comparison of conventional 2-columns system, new DWC and retrotted DWC for BTX application
Sizing/cost details

Conventional 2-column system

(with units in brackets)

C-100

C-101

No. of stages
31
No. of stages (top, middle and bottom)
Diameter of column (m)
0.9
Height of column (m)
24
Condenser duty (kW)
1,476
Reboiler duty (kW)
1,508
0.8
Ratio of Vact /Vmax

0.8
30
876
890
0.8

Bare module cost

Column (US$)
Sieve trays (US$)
Condenser (US$)
Reboiler (US$)

223,746
74,092
26,344
64,010

203,501
90,127
22,545
48,511

Capital cost
Total module cost (US$)

480,739

Operating cost
Cooling water (US$/year)
Steam (US$/year)
Total (US$/year)

95,009
333,873
428,882

Total

40

T-101

DWC (retro)
R-101

68
17, 35, 16
0.9
51
1,306
1,355
0.8

68
17, 35, 16
0.9
51
1,306
1,355
0.8

427,247
164,219
48,889
112,521

395,453
162,525
26,403
60,378

300,260
111,224
0
0

450,903

931,642

795,154

504,562

56,394
197,047
253,441

151,403
530,920
682,323

84,066
299,999
384,065

84,066
299,999
384,065

868,652

543,096

484,977

Total annual cost (US$/year)

be seen there is savings of about 43% in the operating cost by

implementing a new DWC in addition to savings of about 14%
in the capital cost, compared to the C2C system. This gives a
clear direction to go for a DWC against the C2C system for new
plants. For existing plants, the C2C system can be retrotted to
a DWC; from Table 4, the resulting operating cost savings are
US$ 298,258/year and the payback period is only 20 months.
This attractive payback period is due to the re-use of most of
the equipments.

3.2.

DWC (new)

Benzene, toluene, ethyl benzene (BTE) application

Table 5 provides the comparison of the three cases for the BTE
separation. The C2C system has 2 columns of diameter 1.1 and
1.5 m, with 51 and 35 stages respectively. The DWC system has

2,352
2,398

only one column with 1.2 m diameter. In R-DWC, the existing column C-100 from the plant is used despite the slightly
smaller diameter. In this case, the ratio of actual vapor velocity to the maximum velocity is 0.9, which is slightly higher
than 0.8 used for other columns but still within acceptable
range of 0.70.9 (Sinnott, 2005). Another option is to use C-101
with a higher diameter which will give a much lower vapor
velocity. However the cost of investment is high for the additional/middle section due to the larger diameter (1.5 m versus
1.2 m), and the payback period works out to be 41 months.
Hence, using C-100 satises technical requirements and also
gives a lower payback period. The additional column section
with 37 trays has to be added as the middle section. The
number of stages in C-100 is more than sufcient to satisfy
the total stages in the top and bottom sections of R-DWC.

Table 5 Comparison of conventional 2-columns system, new DWC and retrotted DWC for BTE application
Sizing/cost details

Conventional 2-column system

(with units in brackets)

C-100

C-101

No. of stages
No. of stages (top, middle and bottom)
Diameter of column (m)
Height of column (m)
Condenser duty (kW)
Reboiler duty (kW)
Ratio of Vact /Vmax

51

35

1.1
41
1,262
1,282
0.8

1.5
31
753
776
0.8

Bare module cost

Column (US$)
Sieve trays (US$)
Cost of condenser (US$)
Cost of reboiler (US$)

331,968
140,837
33,748
58,605

446,086
134,749
29,657
45,348

Capital cost
Total module cost (US$)

697,840

Operating cost
Cooling medium cost (US$)
Steam cost (US$)
Total (US$/year)

81,234
283,837
365,071

Total annual cost (US$/year)

Total

DWC (new)

DWC (retro)

T-101

T-101

65
15, 37, 13
1.2
52
1,293
1,335
0.8

65
15, 37, 13
1.1
52
1,293
1,335
0.9

778,054
275,586
63,405
103,953

479,111
194,421
38,461
59,895

249,658
102,452
0
0

808,125

1,505,965

951,648

431,757

48,489
171,896
220,385

129,723
455,733
585,456

83,229
295,571
378,800

83,229
295,571
378,800

886,649

569,130

465,152

2,015
2,058

55

chemical engineering research and design 8 7 ( 2 0 0 9 ) 4760

Table 6 Comparison of conventional 2-columns system, new DWC and retrotted DWC for depropanizer/debutanizer
application
Sizing/cost details
(with units in brackets)

Conventional 2-column system

C-100

C-101

DWC (new)
Total

32

DWC (retro)

T-101

R-101

68
10, 40, 18
2.4
67
9,685
9,409
0.8

68
10, 40, 18
2.3
67
9,685
9,409
0.9

No. of stages
32
No. of stages (top, middle and bottom)
Diameter of column (m)
2.3
Height of column (m)
33
Condenser duty (kW)
8,058
Reboiler duty (kW)
7,376
0.8
Ratio of Vact /Vmax

2.3
33
7,756
8,064
0.8

Bare module cost

Column (US$)
Sieve trays (US$)
Condenser (US$)
Reboiler (US$)

3,084,705
241,941
352,127
175,706

3,019,246
241,941
118,133
186,877

6,103,951
483,882
470,260
362,583

6,981,134
557,481
764,654
209,211

3,459,106
306,963
764,654
0

Capital cost
Total module cost (US$)

4,754,046

4,388,871

9,142,917

10,489,082

5,595,675

Operating cost
Cooling water (US$/year)
Steam (US$/year)
Total (US$/year)

1,037,374
1,633,056
2,670,430

998,495
1,785,381
2,783,876

2,035,869
3,418,437
5,454,306

1,246,831
2,083,165
3,329,996

1,246,831
2,083,165
3,329,996

7,282,889

5,427,813

1,119,135

15,814
15,440

Total annual cost (US$/year)

The existing condenser and reboiler can be re-used. Hence,

the costing shows the cost for additional middle section and
trays. From Table 5, it can be seen there is savings of about
35% in the operating cost by implementing a DWC in addition to savings of about 36% in capital cost, compared to the
C2C system. This gives a clear direction to go for a DWC
against the C2C for new plants. For the existing plants, the
C2C system can be retrotted to a R-DWC; from Table 5, the
resulting operating cost savings are US$ 206,656/year and the
payback period is 25 months, which is attractive for the BTE
application.

3.3.

Depropaniser/debutaniser application

The C2C for this depropanizer/debutanizer application has 2

columns of diameter 2.3 m with 32 stages each (Table 6). The
DWC system has only one column with diameter of 2.4 m. The
existing column 2 (T-101) from the plant can be retrotted
to a R-DWC assuming that a vapor velocity of 0.9 times the
maximum is acceptable. The 32 stages in C-101 is more than
sufcient for both the bottom and top sections, and the middle
section with 40 trays has to be added. The existing reboiler can
be re-used but not the existing condenser, and hence costing
for a new condenser is included. From Table 6, it can be seen
that there is savings of about 39% in the operating cost by
implementing a new DWC or R-DWC. The payback period for
retrotting the existing C2C system to a R-DWC is 32 months
which is good and acceptable. The column capacity being very
big, duties of the reboiler and condenser are very large, and so
the energy saving potential is high.

3.4.

Ethanol, water and ethyl glycol (EWE) application

For the EWE application, the C2C system has 2 columns of

diameter 0.9 and 0.5 m, with 30 and 7 stages respectively
(Table 7). The DWC system has one column with diameter of
only 1.1 m. For retrotting to a R-DWC, diameter of the existing column (C-100) is not sufcient for use as a DWC even

if vapor velocity of 0.9 times the maximum vapor velocity is

acceptable, which leaves the only option of a new column for
retrotting. The existing reboiler can be re-used but not the
existing condenser. From Table 7, it can be seen that there is
savings of about 31% in the operating cost by implementing
a new DWC compared to the conventional 2-column system.
This gives a clear direction to go for a DWC against using the
C2C system for new plants. The existing C2C system can be
retrotted to a DWC but the payback period is 49 months,
which is not that attractive. The main reasons being the higher
investment cost and fewer saving (in terms of the amount).
The savings in terms of percentage for operating cost is still
attractive (31%). One alternate solution is using cheaper material of construction instead of stainless steel. Carbon steel
instead of stainless steel as the material for the column (with
trays still made of stainless steel) would reduce the capital cost
and hence the payback period. In such a case, retrotting the
existing C2C system to a R-DWC is attractive with a payback
period of 33 months.

3.5.

Ethanol, propanol and butanol (EPB) application

For EPB application, the C2C system has 2 columns of diameter 2.6 and 1.5 m, with 52 and 32 stages respectively (Table 8).
The DWC have only one column with diameter of 2.3 m. The
existing column 1 (C-100) with diameter of 2.6 m can be used
for retrotting since the diameter requirement for the DWC
is only 2.3 m. The number of stages in the existing column
is more than sufcient to satisfy the total stages in the top
and bottom sections of the DWC. The existing condenser and
reboiler can be re-used. Hence, the costing shows the cost for
additional middle section and trays. From Table 8, it can be
seen that there is savings of about 43% in the operating cost
by implementing a DWC in addition to savings of about 30% in
capital cost compared to a C2C system. This gives a clear direction to go for a DWC against the C2C system for new plants.
The payback period for retrotting the conventional column
to a R-DWC is only 18 months.

56

chemical engineering research and design 8 7 ( 2 0 0 9 ) 4760

Table 7 Comparison of conventional 2-columns system, new DWC and retrotted DWC for EWE application
Sizing/cost details

Conventional 2-column system

(with units in brackets)

C-100

C-101

No. of stages
No. of stages (top, middle and bottom)
Diameter of column (m)
Height of column (m)
Condenser duty (kW)
Reboiler duty (kW)
Ratio of Vact /Vmax

30

0.9
24
1,259
1,289
0.8

0.5
7
202
141
0.8

Bare module cost

Column (US$)
Sieve trays (US$)
Condenser (US$)
Reboiler (US$)

162,944
71,702
45,708
58,776

38,098
22,612
22,108
22,699

Capital cost
Total module cost (US$)

421,321

Operating cost
Cooling water (US$/year)
Steam (US$/year)
Total (US$/year)

81,041
285,386
366,427

Total

T-101

DWC (retro)
R-101

35
17, 9, 9
1.1
29
1,033
973
0.8

35
17, 9, 9
1.1
29
1,033
973
0.9

201,042
94,314
67,816
81,475

252,769
96,653
73,828
50,731

252,769
96,653
73,828
0

131,735

553,056

596,064

531,198

13,028
31,151
44,179

94,069
316,537
410,606

66,493
215,313
281,806

66,493
215,313
281,806

521,217

401,019

388,046

1,461
1,430

Total annual cost (US$/year)

3.6.

DWC (new)

Alkanes (pentane, hexane, heptane) application

Table 9 provides the comparison of C2C system, new DWC

and R-DWC for the alkanes separation. The C2C system has 2
columns of diameter 0.8 m each, with 38 and 21 stages respectively. A new DWC has only one column with diameter of 0.9 m.
For R-DWC, the existing column is tried for retrotting but the
ratio of actual vapor velocity to the maximum vapor velocity is more than 0.9, which leaves the only option of a new
column. The existing condenser and reboiler can be re-used.
From Table 9, it can be seen that there is savings of about 37%
in the operating cost by implementing a DWC in addition to
savings of about 20% in capital cost, compared to the C2C sys-

tem. This gives a clear direction to go for a DWC against the

C2C system for new plants. For the existing plants, the C2C
system can be retrotted to a R-DWC and the payback period
is 32 months, which is attractive.

3.7.

Analysis of the results

For retrotting the C2C system to a R-DWC, most of the applications show the techno-economic viability. Table 10 provides
the summary of the results including the costing details for the
three cases of all applications. The condenser duties, which
are comparable to reboiler duties, are not shown in this table.
It can be observed from the payback periods in Table 10 that

Table 8 Comparison of conventional 2-columns system, new DWC and retrotted DWC for EPB application
Sizing/cost details
(with units in brackets)

Conventional 2-column system

C-100

C-101

No. of stages
52
No. of stages (top, middle and bottom)
Diameter of column (m)
2.6
Height of column (m)
52
Condenser duty (kW)
6,881
Reboiler duty (kW)
8,012
0.8
Ratio of Vact /Vmax

1.5
29
2,889
2,984
0.8

Bare module cost

Column (US$)
Sieve trays (US$)
Condenser (US$)
Reboiler (US$)

2,039,313
499,787
102,878
183,749

418,948
123,199
73,679
95,349

Capital cost
Total module cost (US$)

3,479,941

Operating cost
Cooling water (US$/year)
Steam (US$/year)
Total (US$/year)

442,924
1,773,868
2,216,792

Total annual cost (US$/year)

Total

32

DWC (new)
T-101

DWC (retro)
R-101

60
10, 40, 18
2.3
59
5,187
6,316
0.8

60
10, 40, 18
2.6
59
5,187
6,316
0.9

2,458,261
622,986
176,557
279,098

1,736,263
453,640
110,320
155,557

1,270,247
336,395
0
0

880,907

4,360,848

3,025,223

1,970,067

185,963
660,662
846,625

628,887
2,434,530
3,063,417

333,883
1,398,371
1,732,254

333,883
1,398,371
1,732,254

3,935,586

2,337,299

2,126,267

9,770
10,996

57

chemical engineering research and design 8 7 ( 2 0 0 9 ) 4760

Table 9 Comparison of conventional 2-columns system, new DWC and retrotted DWC for alkanes application
Sizing/cost details
(with units in brackets)

Conventional 2-column system

C-100

C-101

No. of stages
38
No. of stages (top, middle and bottom)
Diameter of column (m)
0.8
Height of column (m)
28
Condenser duty (kW)
753
Reboiler duty (kW)
817
0.8
Ratio of Vact /Vmax

0.8
17
666
672
0.8

Bare module cost

Column (US$)
Sieve trays (US$)
Condenser (US$)
Reboiler (US$)

155,489
85,621
30,856
46,488

109,679
47,317
23,410
42,312

Capital cost
Total module cost (US$)

394,545

Operating cost
Cooling water (US$/year)
Steam (US$/year)
Total (US$/year)

48,438
180,841
229,279

DWC (new)

Total

T-101

21

DWC (retro)
R-101

48
14, 23, 11
0.9
37
858
934
0.8

48
14, 23, 11
0.9
37
858
934
0.9

265,168
132,938
54,266
88,800

232,565
114,723
32,749
49,708

232,565
114,723
0
0

276,544

671,089

531,729

425,846

42,870
148,804
191,674

91,308
329,645
420,953

55,222
206,833
262,055

55,222
206,833
262,055

555,170

368,311

371,890

1,419
1,489

Total annual cost (US$/year)

retrotting the C2C system to a DWC is attractive for BTX, BTE,

depropanizer, EPB and alkanes separations. For EWE separation, the payback period is slightly higher (4 years) and not
attractive due to higher investment though the percentage
savings in the operating cost is comparable with other applications. Factors which affect the investment and hence the
payback period have been analyzed, and alternate solutions
have been proposed in order to reduce the payback period.
These are discussed below.

3.7.1.

Capacity of the plant

Table 10 presents retrotting results corresponding to the

capacities of the existing plants. Here, the longer payback
period in some cases is because of limited use of existing
equipment. Table 11 presents retrotting results at capacities
that maximize re-use of existing equipment. Recall that the
existing columns could not be used for retrotting to DWCs (at
100% feed capacity) for EWE and alkanes applications. A quick
calculation shows that, at reduced capacities, the existing col-

Table 10 Summary of the techno-economics of various cases for all applications

Applications and cases

Reboiler duty
(kW)

Investment
cost US$

Operating cost
US$/year

Payback period
(months)

Savings
US$/year

BTX
C2C
DWC
R-DWC

2,398
1,355
1,355

931,642
795,154
504,562

682,323
384,065
384,065

298,258
298,258

43.7%
43.7%

20

BTE
C2C
DWC
R-DWC

2,058
1,335
1,335

1,505,965
951,648
431,757

585,456
378,800
378,800

206,656
206,656

35.3%
35.3%

25

15,440
9,409
9,409

9,142,917
10,489,082
5,595,675

5,454,306
3,329,996
3,329,996

2,124,309
2,124,309

38.9%
38.9%

32

EWE
C2C
DWC
R-DWC

1,430
973
973

553,056
596,064
531,198

410,606
281,806
281,806

128,800
128,800

31%
31%

49

EPB
C2C
DWC
R-DWC

10,996
6,316
6,316

4,360,848
3,025,223
1,970,067

3,063,417
1,732,254
1,732,254

1,331,162
1,331,162

43.5%
43.5%

18

Alkanes
C2C
DWC
R-DWC

1,489
934
934

671,089
531,279
425,846

420,953
262,055
262,055

158,897
158,897

37.7%
37.7%

32

Depropanizer/debutanizer
C2C
DWC
R-DWC

58

chemical engineering research and design 8 7 ( 2 0 0 9 ) 4760

Table 11 Cost per unit feed rate at capacity which maximizes re-use of existing equipment
Description
Feed ow rate at 100%
Capacity factor
Reboiler duty
Investment
Operating cost per year
Savings per year
Payback period

Units

BTX

BTE

kg/h
%
kW/(kg/h of feed)
US$/(kg/h of feed)
US$/(kg/h of feed)/year
US$/(kg/h of feed)/year
Months

9228
100%
0.15
55
42
32
20

9228
100%
0.14
47
41
22
25

umn can be utilized and hence results in a better payback

period. The reduced capacity factor is shown in Table 11; the
investment cost shown in this table for each application corresponds to the reduced capacity (as per the capacity factor). In
addition, Table 11 gives required investment and savings per
unit feed rate. Note that distillation columns in some companies may be operating at lower capacities depending on the
product demand and market conditions. Hence, for such situations, retrotting the existing C2C system to a DWC may be
desirable provided the investment has an acceptable payback
period.
It is also worthwhile to analyze if both the existing columns
can be retrotted to two DWCs. Each may have different capacity but the capacity of both columns put together is likely to be
more than the existing capacity eventually. This option has to
be studied in detail taking into account product demand and
ability of upstream and downstream equipments to take care
of the increased capacity. The benets are not just savings in
the operating cost but also signicantly increased product output. In the depropanizer/debutanizer application discussed
above, the existing plant has 2 columns of diameter of 2.3 m
with 32 stages. The diameter requirement for retrotting to
DWC is 2.4 m, which is very close to the existing diameter. One
of the existing 2 columns can be retrotted to a DWC operating at reduced capacity based on vapor velocity of 0.8 times
the maximum velocity. Alternately, as stated above, if vapor
velocity of 0.9 times the maximum velocity is acceptable, one
existing column with a diameter of 2.3 m can be retrotted
to DWC at 100% of the original design capacity. In both these
cases, the payback period works out to less than 3 years. As an
additional option, if both the columns are retrotted to DWCs
operating at 90 or 100% (depending on vapor velocity) of the
original design capacity, then the plant can process 80 or 100%
more feed. Similarly, the 2 columns in the C2C system for alkanes separation can be retrotted to 2 DWCs, each operating at
80% of the original design capacity; then the plant can process
60% more feed.

3.7.2.

Depropanizer/
debutanizer

EWE

EPB

Alkanes

90,170
100%
0.10
62
37
24
30

4405
75%
0.22
45
64
29
19

18,030
100%
0.35
109
96
74
18

8604
80%
0.11
28
30
18
18

Material of construction

Material of construction of equipments invariably affects the

capital cost. Material of construction for a column is one
grade less than that for the internals such as trays, packing and distributors, depending upon the corrosive nature
of uids in the process. This is seen widely in industrial
cases due to economic reasons. In some cases where the
process requires higher grade materials, internals are made
with the solid material of that grade. However, columns are
made of carbon steel and then inner-lined with the higher
grade as required for the process conditions. Certain mills
now produce cladded materials and supply directly to the
market. Such cladded materials from the mill exhibit high
quality very similar to the solid material of the same grade.
Table 12 gives a summary of the payback period if material
of construction is varied; here, SS refers to stainless steel
and CS refers to carbon steel. Table 12 presents data only
for depropanizer/debutanizer, EWE and alkanes where the
payback period for retrotting to a R-DWC was higher than
other 3 applications. For example, depropanizer/debutanizer
application with SS, the payback period is 32 months which
can be reduced to 15 months by using CS, which is in the
very attractive range. Hence, the material of construction
should be re-evaluated considering all possibilities, which
may result in more economic benets for the retrotting
exercise. Similarly, EWE and alkanes applications also turn
more attractive for retrotting with the change in material
of construction. The results in Table 12 are based on the
change in material of construction of column only. If required,
this can be extended to column internals, condenser and
reboiler.

3.7.3.

Related issues

The chemical/petrochemical plants usually have an annual

maintenance period of 1014 days, during which major repair
and replacement are carried out. The retrotting discussed
above has been assumed to take place within this mainte-

Table 12 Effect of material of construction for retrotting a C2C system to a DWC: depropanizer/debutanizer, EWE and
alkanes separation
Description

Material for column

(with SS internals)

Investment (US$)

Operating cost
(US$/year)

Savings
(US$/year)

Payback period
(months)

Depropanizer/
debutanizer

SS
CS lined with SS
CS

5,595,675
4,124,038
2,652,402

3,329,997
3,329,997
3,329,997

2,124,310
2,124,310
2,124,310

32
23
15

EWE

SS
CS lined with SS
CS

531,198
440,021
348,845

281,806
281,806
281,806

128,800
128,800
128,800

49
41
33

Alkanes

SS
CS lined with SS
CS

425,846
343,614
261,381

262,055
262,055
262,055

158,898
158,898
158,898

32
26
20

chemical engineering research and design 8 7 ( 2 0 0 9 ) 4760

nance period, and so production loss during the retrotting

time has not been considered. If the retrotting is carried
outside the regular maintenance period, then the associated
production loss needs to be considered for techno-economic
evaluation. On the other hand, several countries have introduced incentives and tax rebates to industries investing in
energy conservation projects. Singapore is one such country;
for details, see http://www.nccc.gov.sg/incentive/home.shtm
(accessed in June 2008). These incentives and increasing energy cost, which make retrotting to a DWC even
more favorable, are not taken into account in the above
techno-economic evaluation. Other potential benets from
retrotting an existing C2C system to a DWC are space savings and availability of a spare column for other applications
or capacity increase.

4.

Conclusions

Retrotting the C2C systems to DWCs for several industrial

applications has been studied in this work, to reduce the
energy consumption and hence the operating cost of distillation. The results of this study conrm the lower energy
requirements and consequently cost savings due to use of
DWCs for new plants as well as show the signicant potential for retrotting existing C2C systems. The savings in the
operating cost are more than 30% in all the applications for
new as well as retrotted DWC compared to a C2C system.
The returns from retrotting a C2C system to a DWC are
very good in 5 of the 6 applications considered, with a payback period of less than 3 years. In the EWE application,
the payback period is about 4 years. Two ways of reducing this by changing material of construction or retrotting
2 columns to 2 DWCs, are discussed. The payback period
reduces to 33 months for the EWE application if material of
construction for the column (and not internals) is changed
to carbon steel. Alternately, capacity of the separation system can be increased by retrotting both columns to DWCs
in this and other applications. In general, energy savings by
retrotting to a DWC translate to reduction of fossil fuels consumption and thus contribute to sustainability of chemical
processes.

Appendix A. Retrot procedure and

implications at site
While retrot is an attractive option, its implementation is
more difcult than the new plants. The factors, which affect
the retrotting exercise, are:

techno-economic feasibility of retrotting operation;

existing plant condition;
location of the existing columns and exchangers;
experienced contractors and other resources to perform
retrotting at site;
project management with a clear retrotting execution
plan.
Of the above, the rst one has already been discussed in the
study described in the main text. It is very important to come
out with a good and reliable report on techno-economics of
the retrotting and project execution plan to present them to
the Management/Board and obtain approval. The other factors
are briey discussed below.

59

A.1. Existing plant condition

The equipment to be retrotted as well as the equipments/
systems connected with it should be checked/monitored for
their condition to perform the new operation. If any of them is
not suitable from the point of process/thermal design and/or
mechanical stability/integrity, then such equipment have to
be considered for modications and added to the cost.

A.2. Location of the existing equipment for retrotting

The distillation column, for example, may have some structures for pipe supports. If this equipment is subject to removal
or modication, then the piping attached to it may also be
affected. The piping has to be separated from the attachments
to the column, and temporary supports have to be provided to
hold the piping until completion of the modication to put
this back to its original position. If the column has to be lifted
or some sections need to be replaced, the location and access
are important to plan the crane and other type of facility. It is
important to know where to park the crane to lift the column.
If the space is not available near the column, then it will be
parked at a distance, the crane should have sufcient capacity and overhang to handle the column from such a distance.
Another reason which location affects is the process condition
for downstream equipments such as net pump suction head
for pumps.

A.3. Experienced contractors and resources

Experienced contractors and resources ensure smooth
retrotting; they will take adequate safety measures as well
as plan in detail on the execution plan. The resources available with them in terms of people, machines and equipment
help a safe and successful retrotting. The experienced contractors shall be technically sound to understand and carry
out the necessary testing as per the local statutory regulations
with respect to that industry. For example, hot welding, cutting or modication require non-destructive testing such as
X-ray/gamma ray, ultrasonic testing and dye penetration testing to check if the quality of the welding is acceptable as per
the codes and regulations. Subsequently, the column is subjected to pressure testing such as hydrostatic testing, leak test
such as pneumatic testing to check for possible leaks as well
as the mechanical integrity of the equipment. In some industries, special testing such as helium testing has to be done
to check if the equipment is sound and good for startup and
operation. There are codes and standards (ASME Section VIII,
Division 1) available for a repair or rework of the equipments,
which experienced contractors, will be aware of to complete
the retrotting in full compliance. The vessels and exchangers have to be re-registered for operation as per the statutory
regulations.

A.4. Project management and clear retrotting plan

Project management plays a key role in preparing a clear
retrotting plan. The project plan is usually prepared by the
project management team to achieve the objectives of the
project. The project management monitors and controls the
budget, resource planning and time schedule. Time is the
essence in the retrot operation as this has to be done in the
minimal time. A shutdown should be planned to ensure that
the complete operation can be carried out successfully meet-

60

chemical engineering research and design 8 7 ( 2 0 0 9 ) 4760

ing the budget, available resources and within the planned

time. Otherwise, this will result in additional cost in terms of
production loss.
Planning of resources, appointing the right contractors,
evaluating the technical and commercial risks, etc. are key
activities of planning. All the affected systems need to be put
back for proper operation. In addition, the statutory approvals
from the local authorities are required if certain equipments
such as columns, heat exchangers, piping, etc. are to be modied at site. Coordination and communication plans have to
be drafted and executed properly. Budget depends upon the
complete planning of the operation, contingency, etc. Organizing all the lead activities within the estimated cost leads
the project to the successful completion.

References
Abdul Mutalib, M.I. and Smith, R., 1998, Operation and control of
dividing wall distillation columns. Part 1: Degrees of freedom
and dynamic simulation. Trans IChemE, Part A, 76: 308318.
Abdul Mutalib, M.I. and Smith, R., 1998, Operation and control of
dividing wall distillation columns. Part 2: Simulation and pilot
plant studies using temperature control. Trans IChemE, Part
A, 76: 319334.
Adrian, T., Schoenmakers, H. and Boll, M., 2004, Model predictive
control of integrated unit operations: Control of a divided wall
column. Chem Eng Process, 43: 347355.
Agrawal, R. and Fidkowski, Z.T., 1998, Are thermally coupled
distillation columns always thermodynamically more efcient
for ternary distillation? Ind Eng Chem Res, 37: 34443454.
Amminudin, K.A., Smith, R., Thong, D.Y.C. and Towler, G.P., 2001,
Design and optimization of fully thermally coupled
distillation columns. Part 1: Preliminary design and
optimization methodology. Trans IChemE, Part A, 79: 701715.
Bek-Pedersen, E. and Gani, R., 2004, Design and synthesis of
distillation systems using a driving-force-based approach.
Chem Eng Process, 43: 251262.
Blancarte-Palacios, J.L., Bautista-Valdes, M.N., Hernandez, S.,
Rico-Ramrez, V. and Jimenez, A., 2003, Energy-efcient
designs of thermally coupled distillation sequences for
four-component mixtures. Ind Eng Chem Res, 42: 51575164.
Bruggemann, S. and Marquardt, W., 2004, Rapid screening of
design alternatives for nonideal multi-product distillation
processes. Comput Chem Eng, 29: 165179.
Dunnebier, G. and Pantelides, C.C., 1999, Optimal design of
thermally coupled distillation columns. Ind Eng Chem Res, 38:
162176.
Emtir, M., Rv, E. and Fony S Z., 2001, Rigorous simulation of
energy integrated and thermally coupled distillation schemes
for ternary mixtures. Appl Therm Eng, 21: 12991317.
Fidkowski, Z. and Krolikowski, L., 1987, Minimum energy
requirements for thermally coupled distillation systems.
AIChE J, 33: 643653.

Fidkowski, Z.T., Doherty, M.F. and Malone, M.F., 1993, Feasibility

of separations for distillation of nonideal ternary mixtures.
AIChE J, 39: 13031321.
Hairston, D., Parkinson, G., Kamiya, T., Aquino, R.D. and Ondrey,
G., 1999, The divide in distillation: while divided-wall systems
seek acceptance, catalytic units bask in new applications.
Chem Eng, 106(4): 32.
Harlvorsen, I.J. and Skogestad, S., 2003, Minimum energy
consumption in multicomponent distillation. 1: Vmin diagram
for a two-product column. Ind Eng Chem Res, 42: 596604.
Hewitt, G., Quarini, J. and Morrell, M., 1999, More efcient
distillation. Chem Eng, 1619.
Jimenez, A., Ramirez, N., Castro, A. and Hernandez, S., 2003,
Design and energy performance of alternative schemes to the
Petyluk distillation system. Trans ChemE, Part A, 81: 518524.
Kaibel, G., 1987, Distillation columns with vertical partitions.
Chem Eng Technol, 10: 9298.
Kim, Y.H., 2006, A new fully thermally coupled distillation
column with postfractionator. Chem Eng Process, 45: 254263.
Kolbe, B. and Wenzel, S., 2004, Novel distillation concepts using
one-shell columns. Chem Eng Process, 43: 339346.
Lek, C.M., Rangaiah, G.P. and Hidajat, K., 2004, Distillation:
revisiting some rules of thumb. Chem Eng, (September): 5055.
Muralikrishna, K., Madhavan, V.K.P. and Shah, S.S., 2002,
Development of dividing wall distillation column design space
for a specied separation. Trans IChemE, Part A, 80: 155166.
Petyluk, F.B., Platonov, V.M. and Slavinskii, D.M., 1965,
Thermodynamically optimal method of separating
multicomponent mixtures. Int Chem Eng, 5(3): 555561.
Rangaiah, G.P., Chan, J.T.Y. and Hidajat, K., 2006, Optimization of
complex distillation columns. Chem Eng, (September): 6064.
Rev, E., Emtir, M., Szitkai, Z., Mizsey, P. and Fonyo, Z., 2001, Energy
savings of integrated and coupled distillation systems.
Comput Chem Eng, 25: 119140.
Schultz, M.A., Stewart, D.G., Harris, J.M., Rosenblum, S.P., Shakur,
M.S. and OBrien, D.E., 2002, Reduce costs with dividing wall
columns. Chem Eng Prog, (May): 6471.
Seider, W.D., Seader, J.D. and Lewin, D.R., (2003). Product & Process
Design Principles: Synthesis, Analysis and Evaluation (2nd Ed.).
(John Wiley, New York).
Shah, P.B., 2002, Squeeze more out of complex columns. Chem
Eng Prog, (July): 4655.
Sinnott, R.K., (2005). Chemical Engineering Design (4th edition).
Coulson & Richardsons Chemical Engineering Series (Elsevier
Butterworth Heinemann, Oxford).
Triantafyllou, C. and Smith, R., 1992, The design and
optimization of fully thermally coupled distillation columns.
Trans IChemE, Part A, 70: 118132.
Turton, R., Bailie, R.C., Whiting, W.B. and Shaeiwitz, J.A., (2003).
Analysis, Synthesis and Design of Chemical Processes. (Prentice
Hall, Upper Saddle River, NJ).
Wright, R.O., 1949, Fractionation Apparatus, U.S. Patent No.
2,471,134.