VOL. 45NOS. 4A P R I L

2 0 1 2

J O U R N A L O F

C H E M I C A L E N G I N E E R I N G

O F J A PA N

The Society ofChemical Engineers,Japan

JCEJAQ 45(4)245-310(2012)ISSN 0021-9592

[ VOL. 45, NO. 4, APRIL 2012 ]

▶Process Systems Engineering and Safety

Development of Environment for Logical Process Safety Management Based on the Business Process Model (Journal Review)Yukiyasu Shimada ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 245

▶Transport Phenomena and Fluid EngineeringOperation for Fine Particle Dispersion in Shear-Thinning Fluid in a Stirred VesselTakafumi Horie, Nobuhiro Sakano, Eiji Aizawa, Norihisa Kumagai and Naoto Ohmura ––––––––––––––––––––––––––––––––––– 258

Self-Organization of Polymer Films Using Single and Mixed Solvents on Chemically Patterned SurfacesShohei Yasumatsu, Hirotaka Ishizuka, Koichi Nakaso, Patience Oluwatosin Babatunde and Jun Fukai –––––––––––––––––– 265

▶Particle Engineering

Acoustic E�ect on Induction of Cerium Carbonate in Reaction CrystallizationXin-Kuai He, Rayoung Kim, Dongmin Shin and Woo-Sik Kim –––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 272

Definition of the New Mean Particle Size Based on the Settling Velocity in Liquid (SC)Rondang Tambun, Masamitsu Shimadzu, Yuichi Ohira and Eiji Obata ––––––––––––––––––––––––––––––––––––––––––––––––––––– 279

▶Separation Engineering

Improvement of the Deethanizing and Depropanizing Fractionation Steps in NGL Recovery Process Using Dividing Wall ColumnNguyen Van Duc Long and Moonyong Lee –––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 285

▶Chemical Reaction Engineering

Selective Hydrogenation of Unsaturated Aldehyde in a Gas–Liquid–Liquid–Solid Batch Reactor (SC)Hiroshi Yamada, Takashi Miura, Tomohiko Tagawa and Shigeo Goto –––––––––––––––––––––––––––––––––––––––––––––––––––––– 295

▶Process Systems Engineering and Safety

Nonlinear Process Control Using a Direct Adaptive PI ControllerChyi-Tsong Chen and Shih-Tien Peng ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 298

▶Materials Engineering and Interfacial Phenomena

Production of Fine Organic Crystalline Particles by Using Milli Segmented Flow CrystallizerShoji Kudo and Hiroshi Takiyama ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– 305

This article appeared in the Journal of Chemical Engineering of Japan.The attached copy is provided to the author for non-commercial research, education use and sharing with colleagues.Other uses listed below are prohibited:

- Reproduction,- Commercial use,- Posting to personal, institutional or third party websites.

Vol. 45 No. 4 2012 285Copyright © 2012 The Society of Chemical Engineers, Japan

Journal of Chemical Engineering of Japan, Vol. 45, No. 4, pp. 285–294, 2012

Improvement of the Deethanizing and Depropanizing Fractionation Steps in NGL Recovery Process Using Dividing Wall Column

Nguyen Van Duc Long and Moonyong LeeSchool of Chemical Engineering, Yeungnam University, Gyeongsan 712-749, South Korea

Keywords: Distillation, Dividing Wall Column, DWC, Natural Gas Liquid, Fully Thermally Coupled Distillation Column, Heat Integration

In this work, our aim is to utilize a dividing wall column to improve the performance of the deethanizing and depro-panizing fractionation steps in natural gas liquid processing. Starting from an initial conventional column sequence, the initial designs of the conventional dividing wall column and a top dividing wall column are obtained by maintaining the number of trays. In succession, they are optimized to reduce the energy consumption using factorial design. The results show that the conventional dividing wall column and the top dividing wall column can o�er many bene�ts to the system, e.g., curbing the operating cost including refrigeration cost, and minimizing the reboiler and condenser duty. Furthermore, by using a dividing wall column, both the purity of ethane and its recovery rate are increased. The in�u-ence of utility prices on the operating cost saving of the conventional and the top dividing wall columns is also inves-tigated. In addition, to further enhance the dividing wall column performance, heating is integrated on the top and an interreboiling system is installed at the bottom section of the dividing wall column.

Introduction

�e natural gas (NG) industry, which �rst dates back to the early 1900s in the United States, is still an evolving �eld. Although the primary use of NG is as a fuel, it can also be a source of hydrocarbons for petrochemical feed stocks and a major source of elemental sulfur, which is an impor-tant industrial chemical (Kidnay and Parrish, 2006). Due to the clean burning characteristics and the ability to meet stringent environmental requirements, the demand for nat-ural gas has increased considerably over the past few years. �erefore, the recovery of natural gas liquids (NGL) has be-come more and more economically attractive as a number of components in natural gas are o�en isolated from the bulk gas and then sold separately. Consequently, there are numer-ous process con�gurations and methods known to increase NGL recovery from a feed gas. In addition, these processes can provide potential enhancements to the overall facility availability and project economics when employing the in-tegrated concept (Elliot et al., 2005; Mak, 2006). However, one of the challenges in the integration of an NGL recovery process is the di�erence in column operating pressures.

Distillation is the primary separation process used in chemical processing industries. While this unit operation has many advantages, one major drawback is its large energy requirement (Schultz et al., 2002). �is can signi�cantly in-�uence the overall plant pro�tability. �erefore, the increas-ing cost of energy has forced industry to reduce its energy

consumption. In addition, the tendency to rationalize and to reduce the usage of fossil energy resources, and the tighter environmental regulations caused by, for example, the e�ort to prevent climate change, have generated the need to em-ploy new and more e�cient separation methods (Knapp and Doherty, 1990; Malinen and Tanskanen, 2009).

Several approaches have been proposed to develop alter-natives for energy savings. One of them is the integration of the column with the rest of the process (Linnho� et al., 1983; Jiménez et al., 2003). Figure 1(a) illustrates the di-rect conventional distillation sequence for ternary distilla-tion. �ere is a reboiler and a condenser in every distillation column. �e number of product �ows in the columns is always the same, i.e. two. However, in a more novel distilla-tion scheme, i.e. in a prefractionator scheme, as illustrated in Figure 1(b), there is one extra product �ow, known as a side stream, in the second column of the column sequence (Malinen and Tanskanen, 2009). �rough the implementa-tion of two interconnecting streams (one in the vapor phase and the other in the liquid phase) between the two columns, thermally coupled distillation systems can be built; such sys-tems have been particularly studied for the separation of ter-nary mixtures, and have been shown to provide signi�cant energy savings with respect to both the conventional direct and indirect distillation sequences. �e three most widely analyzed thermally coupled systems are the sequence with a side recti�er, the sequence with a side stripper and the fully thermally coupled system (or Petlyuk column), which are illustrated in Figures 2(a), (b) and Figure 3(a), respec-tively (Glinos et al., 1986; Fidkowski and Krolikowski, 1990; Triantafyllou and Smith, 1992; Hernández and Jiménez, 1996; Jiménez et al., 2003).

Research Paper

Received on August 30, 2011; accepted on December 3, 2011Correspondence concerning this article should be addressed to M. Lee (E-mail address: mynlee@yu.ac.kr).

286 Journal of Chemical Engineering of Japan

Of the above possible thermal coupling arrangements, the side stripper and the side recti�er have been widely used in re�nery distillation and cryogenic air separation, re-spectively (Seidel, 1935; Watkins, 1979; Amminudin et al., 2001). �e Petlyuk column provides a fully interconnected structure, with two thermal couplings that result in vapor interconnections �owing back and forth between the col-umns (Hernández et al., 2005). �ese sequences have been shown to provide energy savings of up to 30% over the con-ventional direct and indirect distillation sequences (Tedder and Rudd, 1978; Alatiqi and Luyben, 1985; Triantafyllou and Smith, 1992; Finn, 1993; Hernández and Jiménez, 1996, 1999; Blancarte-Palacios et al., 2003).

However, the Petlyuk column undergoes strong interac-tions between the two columns because of their thermal in-tegration, which causes di�culties in design and operation. To solve this problem as well as to reduce the capital cost, a vertical wall is installed in the central section of the column, dividing the column into the prefractionator and the main column, as schematically drawn in Figure 3(b). �is ar-rangement is called the dividing wall column (DWC), which is conceptually the same as the Petlyuk column due to their thermodynamically equivalent arrangements (Amminudin et al., 2001). By moving the dividing wall to the top or the bottom of the column, the top dividing wall column (TDWC) or the bottom dividing wall column (BDWC) can be constituted, respectively.

Although DWCs can potentially reduce energy use and

investment costs (Midori et al., 2000; Olujić et al., 2009; Dejanović et al., 2010; Long et al., 2010; Long and Lee, 2011), their design and optimization is challenging, involv-ing the solving of multivariable problems with variables that interact with each other and need to be optimized simulta-neously. A common di�culty associated with detailed simu-lation is related to the estimation of the number of stages in each section (Dejanović et al., 2010). Statistical approaches have shown potential for this, with DWCs' structures hav-ing been optimized by response surface methodology (Long and Lee, 2012). �is method can be easily and e�ciently implemented using Hysys and Minitab. Another statistical approach, factorial design, is widely used in experimen-tal investigation (Montgomery, 1991; Cestari et al., 1996; Persson and O. Åström, 1997; Carvalho et al., 1999; Cestari et al., 2008). However, factorial design can only give relative values. Regardless, this method is a useful method to design experiments or simulation in both laboratory and industrial settings.

A set of improved deethanizing and depropanizing frac-tionation steps in NGL processing, such as interreboiling, intercondensing, thermal coupling, and “thermo-mechan-ical” coupling, has been studied (Manley, 1997). Inspired by this study, the our aim of study is, therefore, to further enhance the performance of the deethanizing and depro-panizing fractionation steps in NGL processing using DWC and heat integration. Factorial design was employed to �nd the optimum structure of DWC. Furthermore, a comparison

(a)

(b)

Fig. 1　Schematic diagram of the (a) direct sequence and (b) prefrac-tionator arrangement

(a)

(b)

Fig. 2　Schematic diagram of the thermally coupled distillation se-quences: (a) system with side recti�er and (b) system with side stripper

Vol. 45 No. 4 2012 287

between the conventional DWC and the TDWC was per-formed.

1.　Conventional Distillation Sequence

Figure 4 illustrates the conventional distillation sequence and its current operating conditions. �e critical tempera-ture for ethane is approximately 32.3°C; therefore, the NGL deethanizer column requires a refrigerated condenser when producing relatively pure ethane. In order to minimize the refrigeration costs, the deethanizer was designed and oper-ated at relatively high pressures of about 450 psia (31.0 bar). �e resulting condenser temperature of nearly 13.3°C (for relative pure ethane) is also high enough to preclude the formation of hydrates which can plug the equipment. With 18 theoretical stages, the deethanizer column produces a partially condensed overhead vapor stream. Whereas, the depropanizer, possessing 34 theoretical stages, was de-signed and operated at about 250 psia (17.2 bar), as com-mercial propane can be condensed with cooling water at this pressure (Manley, 1997). �e feed composition, tem-perature, and pressure are described in Table 1. �e simula-tion work was performed using the simulator Aspen Hysys V7.1 (Aspen Technology, Inc.). �e Peng–Robinson equa-

(a)

(b)

Fig. 3　Schematic diagram of the (a) Petlyuk column and (b) dividing wall column

Fig. 4　Simpli�ed �ow sheet illustrating the separation train of two conventional columns

Table 1　Feed conditions of the mixture

Feed conditions

Component Mass �ow [kg/h] Liquid volume fractions [%]

Methane 991.66 0.50Ethane 87179.69 37.00Propane 87268.17 26.00i-Butane 26803.59 7.20n-Butane 57180.35 14.80i-Pentane 20649.83 5.00n-Pentane 14600.65 3.50n-Hexane 17559.16 4.00n-Heptane 9099.56 2.00Temperature [°C] 55.83Pressure [bar] 31.37

Table 2　Column hydraulics, energy performance, and product speci-�cations of the existing columns sequence

Deethanizer Depropanizer

Number of trays 18 34Tray type Sieve SieveColumn diameter [m] 6.2 4.9Number of �ow paths 1 1Tray spacing [mm] 609.6 609.6Max �ooding [%] 85.99 83.51Condenser duty [MW] 14.67 26.03Reboiler duty [MW] 28.95 21.54Purity of C2 [Liq. Vol.

Fractions %]95.13

Recovery of C2 [%] 94.58Purity of C3 [Liq. Vol.

Fractions %]90.14

Recovery of C3 [%] 92.47

288 Journal of Chemical Engineering of Japan

tion of state that supports the widest range of operating con-ditions and the greatest variety of systems was used for the prediction of vapor–liquid equilibrium of these simulations (Aspen Technology, 2009). Table 2 presents the conditions and product speci�cations for each column in the conven-tional distillation sequence. From the base case simulation model, it shows that the energy consumption of the deetha-nizer and depropanizer are 28.95 and 21.54 MW, respec-tively.

All columns were designed with the maximum �ooding of near 85%, so as to prevent �ooding in the columns. To determine the maximum �ooding of a particular column, the rating mode is simulated with column internal speci�-cations including the type of trays, column diameter, tray spacing, and number of passes. Table 2 lists the parameters necessary to de�ne the existing hydraulic features of the col-umns.

It should be noted that, in order to estimate the operating cost of the proposed sequences, the utility cost data from published literatures (Krajnc and Glavic, 1996; Biegler et al., 1997), shown in Table 3, was used. If a process stream needs to be operated below 27°C, cooling is certainly required, and thus a refrigeration cycle needs to be considered. Refrigerating a stream is far more expensive (on an energy basis) than lowering the temperature with cooling water, or even raising the temperature with steam. Consequently, re-frigeration is generally not a desirable option and other al-ternative processes should be considered �rst (Biegler et al., 1997). Importantly, in the conventional column sequence, the condensers in the deethanizer and depropanizer utilize a refrigeration system and water, respectively.

2.　Integrating Using CDWC

2.1　Conventional dividing wall column (CDWC)In the NGL recovery process, the operating pressure is

an important factor a�ecting the processing cost. Increasing the column pressure increases the temperature of the top vapor stream, i.e. decreases the refrigeration cost. However, it increases the burden on the reboiler, the condenser and also the vessel wall, resulting in a thicker column with a relatively large diameter. �erefore, the optimum pressure of the column is achieved by balancing the investment cost and operating cost. Furthermore, it has been discovered that the ine�cient remixing of propane and butane in the bottom of the deethanizer can be eliminated by reducing the deetha-nizer pressure to match the depropanizer and by thermally coupling the two columns (Manley, 1997). It is also known that a relatively low pressure (about 17.5 bar) deethanizer

requires more refrigeration utilities, but has a much smaller diameter and requires less reboiler duty than a high pressure (about 31.0 bar) one. In addition, the wall thickness, weight and cost of the column will be further reduced because of the lower operating pressure. �erefore, the DWC, operated at approximately 17.5 bar, is applied herein to minimize the investment cost and the energy consumption.

A shortcut design procedure was applied to construct the initial CDWC structure (Amminudin et al., 2001; Long et al., 2010; Lee et al., 2011; Long and Lee, 2011) based on the conventional column con�guration shown in Figure 5. In this conventional con�guration, the �rst column corre-sponds to the prefractionator section in the CDWC. �e rectifying section of the second column and the stripping section of the third column represent the top and bottom sections of the CDWC, respectively. Both the stripping sec-tion of the second column and the rectifying section of the third column are equivalent to the divided wall section of the CDWC. Furthermore, both the bottom stream from the second column and the top stream from the third column refer to the side stream of the CDWC (Long et al., 2010). Consequently, the structure of the CDWC can be divided into four sections: the prefractionator section for the feed mixture; the top and bottom sections above and below the divided wall; and the divided wall section.

�e new CDWC was maintained with 34 stages for a fair comparison with the other sequences. �e main design vari-ables including the internal vapor and liquid �ow to the prefractionator, the number of trays in the top (N1), middle (N2), feed stripping (N3), and side product (N4) section were optimized. In this work, the number of trays in the

Table 3　Utilities cost data

Utility Price [USD/(kW·a)]

Refrigerant 377Cooling water 90

Steam 300Electricity 104

Fig. 5　A three-column distillation system for the initial design of DWC structure

Vol. 45 No. 4 2012 289

dividing wall sections both for the prefractionator and the main fractionators are �xed as the same values as N2. A fac-torial design was employed to analyze their interactive e�ect and to reach an optimum in terms of the reboiler consump-tion. �e simulation run data were �tted to a second order polynomial model, and subsequently regression coe�cients were obtained. �e generalized second order polynomial model used in the factorial design is as given by Eq. (1).

= <

= + + +01

k

i i ij i ji i j

Y β β X β X X ε (1)

Here, Y is the predicted response (energy consumption), Xi is the uncoded or coded value of the variables, β0 is a con-stant, βi is the main e�ect coe�cient, βij is the interaction e�ect coe�cient, and ε is the error. �e second order poly-nomial coe�cients and optimum value of the reboiler con-sumption were obtained using the Minitab so�ware.

Table 4 lists the factors and levels used in this case study. A total of 16 simulation runs were used to �nd the opti-mum values of the 4 parameters associated with the CDWC structure. For each run, the internal vapor and liquid �ow to the prefractionator were varied to minimize the energy con-sumption while meeting the required product purity. �e factorial design can cover the main and interaction e�ects of the parameters over the whole range of selected parameters. According to the sparsity-of e�ects principle in factorial de-sign, it is most likely that the main (single factor) e�ects and two-factor interactions are the most signi�cant e�ects, and the higher order interactions are negligible (Nejad et al., 2011). In other words, higher order interactions such as three-factor interactions are very rare and considered as the residual, which are dispersed randomly. Figure 6 pres-ents the main e�ect plots of design variables N1, N2, N3 and N4 on the reboiler duty. �e analysis of the e�ect of the principal factors illustrated that, in the considered range of parameters, the side product section (N4) is the most signi�cant variable in achieving the minimum energy con-sumption. Figure 7 shows the plots of interactions between the N1, N2, N3 and N4. �e best result of the reboiler con-sumption was found at the coded levels with the number of trays in the top, middle, feed stripping, and side product section of 1, −1, −1, and 1, respectively. From these re-sults, the natural value of the variables can be derived. �en, the recycle vapor and liquid �ow were optimized to achieve the minimum energy consumption and a satisfactory level of product purity. Figure 8 reveals a simpli�ed �ow sheet illustrating the resulting CDWC system. According to the

rigorous simulation results, the duty of the condenser and reboiler are 18.92 and 28.60 MW, respectively. Using the CDWC can save 31.74% of the operating cost. Furthermore, compared with the two columns with diameters of 6.2 and

Fig. 6　Main e�ects plots of design variables N1, N2, N3, and N4 to the reboiler duty

Fig. 7　Interaction plots between main design variables N1, N2, N3, and N4

Table 4　Coded levels of factors

FactorLevels

−1 1

Top section 2 4Middle section 18 22Feed stripping section 11 15Side product section 3 6

Fig. 8　Simpli�ed �ow sheet illustrating the CDWC

290 Journal of Chemical Engineering of Japan

4.9 m, respectively, the use of CDWC with a diameter of 5.5 m, as shown in Table 5, could give a saving of 33.63% in the total annual cost (TAC) when operating at low pres-sure. Furthermore, the purity of C2 is increased from 95.13 to 97.73% and its recovery also is improved from 94.58 to 98.21% as compared to the conventional column sequence.

2.2　Heat integrated CDWCIn the present study, the top vapor stream is divided fur-

ther into two gas streams. �e �rst stream, a re�ux stream, is compressed to a pressure greater than that of the CDWC. �is re�ux is then cooled in a heat exchanger due to the second stream, which then is water-cooled in a condenser. Figure 9 is a schematic illustration of this con�guration, particularly showing how a portion of the overhead vapor stream from the deethanizer is recycled back to the top of the deethanizer. With this con�guration, the temperatures of the top vapor stream of the CDWC a�er being compressed and the re�ux stream are approximately 62 and 32°C, re-spectively, thereby allowing for water to condense both streams.

To minimize the energy consumption, the bottom liquid product can be subcooled while interreboiling the bottom section of the DWC (Manley, 1997). As a result, this sys-tem achieves a substantial saving of 47.58% in the operat-ing cost as compared to the conventional column sequence. Note that the conventional column sequence with the same integration strategy can save up 18.28% in terms of operat-ing cost.

3.　Integrating Using TDWC

3.1　TDWC�e thermally coupled low pressure deethanizer and

depropanizer were, therefore, studied to reduce the energy consumption (Manley, 1997). However, this process requires two columns, which causes an increase in the investment cost. �us, to decrease the refrigeration cost and investment cost, the TDWC, which is equivalent to a combination of both columns, was applied. Consequently, the system shown

in Figure 10 is built by moving the wall up to the top of the column, thereby separating this column into three sec-tions: one is the prefractionator section for the feeding the mixture with its own condenser (C1) to produce C2, another is the bottom section below the divided wall, and the last is the divided wall section. Compared to the CDWC, the TDWC does not have the middle section and has one more condenser. �is could result in a higher investment cost as compared to the CDWC. Furthermore, the liquid split ratio in this column cannot be adjusted except for the vapor split ratio, i.e. this kind of column could not demonstrate the full advantage of the conventional DWC in terms of energy saving.

�e main purpose in this part is to decrease the refrig-eration duty a�ecting the operation of the natural gas pro-cessing to a great extent. �us, this column can be utilized to integrate two columns, whose condensers are cooled by di�erent kinds of cooling resources; particularly, refrigera-tion needs to be used in the deethanizer, while the depro-

Fig. 9　Simpli�ed �ow sheet illustrating the CDWC with modi�ca-tions

Fig. 10　Simpli�ed �ow sheet illustrating the TDWC

Table 5　Design results for the CDWC

Parameters Value

Liquid split* 0.36Vapor split* 0.65Re�ux rate [kg/h] 194706Reboiler duty [MW] 28.60Diameter [m] 5.50Top pressure [bar] 17.50Max �ooding [%] 86.93Purity of C2 [Liq. Vol. Fractions %] 97.73Recovery of C2 [%] 98.21Purity of C3 [Liq. Vol. Fractions %] 90.14Recovery of C3 [%] 92.47

* Note that the split is de�ned as the ratio between the �ow into the prefrac-tionator to the total �ow into the dividing wall section.

Vol. 45 No. 4 2012 291

panizer just needs water to condense the top vapor stream. As a result, the refrigeration cost decreases; in particular, as compared with the conventional column sequence, the TDWC requires less than 0.79 MW. Furthermore, the con-denser (C2) and the reboiler duty can be curbed by 46.36 and 25.73%, respectively. In addition, the purity and recov-ery of C2 also increase from 95.13 to 96.71% and 94.58 to 96.77%, respectively, as compared with the conventional col-umn sequence. �e cross sectional area of the top section of the TDWC is the sum of the prefractionator section area and the top section area of the main fractionator. �e diam-eter of the middle section can be calculated from the cross sectional area of the middle section. Table 6 lists the diam-eter of the top and bottom sections of the TDWC. Based on the results in Table 6, the length of 6.1 m was chosen as the diameter of TDWC. Overall, using TDWC can save TAC by 24.02% as compared to the conventional distillation se-quence.

In the literature, there is almost no data on the study of TDWC or any comparison between the CDWC and TDWC. �us, a comparison of their performances was performed in our study. As mentioned earlier, the TDWC has one more condenser and the liquid split ratio cannot be adjusted as compared to CDWC, causing a higher investment cost and more steam consumption, respectively. Table 7 summarizes the comparison of the performances between CDWC and TDWC. It can be seen that the TDWC requires less refrig-erator duty, having a cost that is a�ected by the electricity utility price, and more reboiler duty, having a cost that is mainly a�ected by the fuel price. Furthermore, the TDWC requires a larger diameter than the CDWC. In addition, the in�uence of utility prices on the operating cost saving of the CDWC and TDWC was studied with varying the ratio of re-frigerant/steam cost. Table 8 reveals that when the cost ratio of refrigerant/steam almost triples, the TDWC o�ers more bene�ts in terms of the operating cost saving as compared to

Table 6　Diameter of the TDWC

Prefractionator Mainfractionator (top) TDWC (top) TDWC (bottom) TDWC

Diameter [m] 4.5 3.1 5.5 6.1 6.1Max �ooding [%] 83.29 82.13 84.64Position of section 17th tray–34th tray 1st tray–16th tray

Table 7　Comparison of di�erent structural alternatives for improving the performance of deethanizer and depropanizer

Structural alternativeConventional column sequence

CDWC TDWC

Conventional column sequence with heat integration CDWC with

heat integrationTDWC with

heat integrationDeethanizer Depropanizer Deethanizer Depropanizer

Tray number 18 34 34 34 18 34 34 34Column diameter

[m]6.2 4.9 5.5 6.1 6.2 4.9 5.5 6.1

Refrigeration duty [MW]

14.67 0.00 18.92 13.88 0 0.00 0.00 0.00

Cooling-water Condenser duty [MW]

0.00 26.03 0.00 13.96 16.38 26.25 31.01 35.38

Compressor duty [MW]

0.00 0.00 0.00 0.00 3.34 0.00 14.30 10.41

Reboiler duty [MW] 28.95 21.54 28.60 37.50 28.91 19.85 25.96 33.68Reboiler duty saving

[%]0.00 43.36 25.73 3.43 48.58 33.29

Operating cost sav-ing [%]

0.00 31.74 22.94 18.28 47.58 37.57

TAC saving [%] 0.00 33.63 24.02 16.61 43.95 33.99Purity of C2 [Liq.

Vol. Fractions %]95.13 97.73 96.71 95.13 97.72 97.18

Recovery of C2 [%] 94.58 98.21 96.77 94.58 98.22 97.45

Table 8　In�uence of utility prices on the operating cost saving of the CDWC and TDWC

Price [USD/(kW·a)] Refrigerant 300 377 600 800 1000Steam 300 300 300 300 300

Cost ratio of refrigerant/steam 1.00 1.25 2.00 2.67 3.33Operating cost saving [%] CDWC 34.88 31.74 24.19 18.85 14.49

TDWC 23.85 22.94 20.75 19.21 17.95

292 Journal of Chemical Engineering of Japan

the CDWC. �erefore, building a new plant with the CDWC or TDWC has to be considered carefully based on the en-ergy price and the equipment cost.

3.2　Heat integrated TDWCIn order to take further advantage of energy e�ciency,

the integrated con�guration applied to the CDWC was also employed to the TDWC, as shown in Figure 11. Table 7 shows that the operating cost and the TAC can be curbed up to 37.57 and 33.99% as compared with the conventional column sequence. �e results indicate that the CDWC with heat integration could obtain the largest TAC saving as com-pared to other sequences.

Conclusions

In this paper, the integration of the deethanizer and depropanizer has been presented to reduce costs and en-ergy consumption. Factorial design is employed to opti-mize the DWC with an easy and e�cient implementation of Hysys and Minitab. �e interaction between the variables can be identi�ed or quanti�ed by the factorial design. �e use of CDWC can gain many bene�ts as well as decreasing the TAC by 33.63%. When the refrigerant cost is too high, the TDWC, combining the deethanizer and depropanizer into a single fractionation column, reduces the pressure of the deethanizer to that of the depropanizer and locates the deethanizer and rectifying section of the depropanizer in the top portion of a single dividing wall column, and can be applied as a replacement. By using this kind of column, not only the energy (refrigeration energy, reboiler and con-denser duty) and the investment cost (replacing two col-umns with a column and eliminating the deethanizer re-boiler) can be curbed, but also the purity and recovery of ethane can be increased. In addition, reducing the number of columns, condensers, reboilers, as well as the vessel wall thickness, and weight as a result of the lower operating pres-sure, can also provide bene�ts in the investment cost. Based on these results, a comparison of the performances between the CDWC and TDWC has been analyzed. Furthermore, to further enhance the DWC performance, heating has been integrated on the top and an interreboiling system has been installed at the bottom section of the DWC.

Acknowledgment

�is research was supported by a grant from the Gas Plant R&D Center funded by the Ministry of Land, Transportation and Maritime A�airs (MLTM) of the Korean government.

Appendix:　Column Cost Correlations

a.　Sizing the column: �e column diameter was determined by the column �ooding condition that �xes the upper limit of vapor veloc-ity. Operating velocity is normally between 70 and 90% of the �ooding velocity (Sinnott, 2005; Premkumar and Rangaiah, 2009); In this study, 85% of the �ooding velocity was used.

b.　Capital cost: Guthrie’s modular method was applied (Biegler et al., 1997). �e investment cost for conventional distillation is the total cost of the column and the auxiliary equipment, such as the reboiler and condenser. For the DWC, it included the additional dividing wall cost. In this study, for cost updating, the Chemical Engineering Plant Cost Index of 575.4 was used.

Updated bare module cost ( ) ( 1)BMC UF BC MPF MF = × × + (2)

where UF is the update factor:

present cost index / base cost indexUF= (3)

BC is bare cost: For vessels:

00 0

α βL DBC BC L D

= × × (4)

For the heat exchanger:

00

αSBC BC S

= × (5)

Area of heat exchanger:

ΔQS U T= (6)

where MPF is the material and pressure factor; MF is the module factor (a typical value), which is a�ected by the base cost. D, L, and S represent the diameter, length and area, respectively.

Compressor cost was interpolated (Peters and Timmerhous, 1991).c.　Operating cost (Op):

steam CW refrigerant electricityOp C C C C= + + + (7)

where Cstream is the cost of steam; CCW is the cost of cooling water; Crefrigerant is the cost of refrigerant and Celectricity is the cost of electricity.

d.　Total annual cost (TAC) (Smith, 1995)

(1 )capital cost(1 ) 1

n

ni iTAC Op

i +

= × ++

(8)

where i is the fractional interest rate per year and n is the number of years.

Nomenclature

B = bottom product �ow [kg/h]C2 = ethane [—]C3 = propane [—]C1 = the condenser of the prefractionator [—]C2 = the condenser of the main fractionator [—]D = top product �ow [kg/h]F = feed �ow [kg/h]RL = liquid split ratio [—]RV = vapor split ratio [—]

Fig. 11　Simpli�ed �ow sheet illustrating the TDWC with modi�ca-tions

Vol. 45 No. 4 2012 293

S = side product �ow [kg/h]

‹Subscript and Superscript›L = liquidV = vapor

Literature Cited

Alatiqi, I. M. and W. L. Luyben; “Alternative Distillation Con�gura-tions for Separating Ternary Mixtures with Small Concentration of Intermediate in the Feed,” Ind. Eng. Chem. Process Des. Dev., 24, 500–506 (1985)

Amminudin, K. A., R. Smith, D. Y. C. �ong and G. P. Towler; “Design and Optimization of Fully �ermally Coupled Distillation Col-umns: Part 1: Preliminary Design and Optimization Methodol-ogy,” Chem. Eng. Res. Des., 79, 701–715 (2001)

Aspen Technology ed.; Aspen HYSYS �ermodynamics COM Inter-face, Version V7.1, Aspen Technology Inc., Burlington, U.S.A. (2009)

Biegler, L. T., I. E. Grossmann and A. W. Westerberg; Systematic Meth-ods of Chemical Process Design, pp. 110–141, Prentice Hall Inc., Upper Saddle River, U.S.A. (1997)

Blancarte-Palacios, J. L., M. N. Bautista-Valadés, S. Hernández Cas-tro, V. Rico-Ramírez and A. Jiménez; “Energy-E�cient Designs of �ermally Coupled Distillation Sequences for Four-Component Mixtures,” Ind. Eng. Chem. Res., 42, 5157–5164 (2003)

Carvalho, M. L., J. M. S. Cabral and M. R. Aires-Barros; “Cutinase Stability in AOT Reversed Micelles: System Optimization Using the Factorial Design Methodology,” Enzyme Microb. Technol., 24, 569–576 (1999)

Cestari, A. R., C. Airoldi and R. E. Bruns; “A Fractional Factorial De-sign Applied to Organofunctionalized Silicas for Adsorption Opti-mization,” Colloids Surf., A, 117, 7–13 (1996)

Cestari, A. R., E. F. S. Vieira and J. A. Mota; “�e Removal of an Anion-ic Red Dye from Aqueous Solutions Using Chitosan Beads—�e Role of Experimental Factors on Adsorption Using a Full Factorial Design,” J. Hazard. Mater., 160, 337–343 (2008)

Dejanović, I., L. Matijašević and Ž. Olujić; “Dividing Wall Column-A Breakthrough towards Sustainable Distilling,” Chem. Eng. Process., 49, 559–580 (2010)

Elliot, D., W. R. Qualls, S. Huang and J. J. (Roger) Chen; “Bene�t of Integrating NGL Extraction and LNG Liquefaction Technology,” AIChE Spring National Meeting, 5th Topical Conference on Natu-ral Gas Utilization (TI) Session 16c-Gas, pp. 10–14, Atlanta, U.S.A. (2005)

Fidkowski, Z. and L. Krolikowski; “Energy Requirements of Noncon-ventional Distillation Systems,” AIChE J., 36, 1275–1278 (1990)

Finn, A. J.; “Consider �ermally Coupled Distillation,” Chem. Eng. Prog., 10, 41–45 (1993)

Glinos, K., I. Nikolaides and F. Malone; “New Complex Column Ar-rangements for Ideal Distillation,” Ind. Eng. Chem. Process Des. Dev., 25, 694–699 (1986)

Hernández, J. G. S., S. Hernández and A. Jiménez; “Analysis of Dynam-ic Properties of Alternative Sequences to the Petlyuk Column,” Comput. Chem. Eng., 29, 1389–1399 (2005)

Hernández, S. and A. Jiménez; “Design of Optimal �ermally-Coupled Distillation Systems Using a Dynamic Model,” Chem. Eng. Res. Des., 74, 357–362 (1996)

Hernández, S. and A. Jiménez; “Design of Energy-E�cient Petlyuk Sys-tems,” Comput. Chem. Eng., 23, 1005–1010 (1999)

Jiménez, A., N. Ramírez, A. Castro and S. Hernández; “Design and Energy Performance of Alternative Schemes to the Petlyuk Distil-

lation System,” Chem. Eng. Res. Des., 81, 518–524 (2003)Kidnay, A. J. and W. R. Parrish; Fundamentals of Natural Gas Pro-

cessing, 1st ed., pp. 1–23, Taylor and Francis, Boca Raton, U.S.A. (2006)

Knapp, J. P. and M. F. Doherty; “�ermal Integration of Homogeneous Azeotropic Distillation Sequences,” AIChE J., 36, 969–984 (1990)

Krajnc, M. and P. Glavic; “Prices of Utilities and Process Structure,” Comput. Chem. Eng., 20, 183–188 (1996)

Lee, S. H., M. Shamsuzzoha, M. Han, Y. H. Kim and M. Y. Lee; “Study of the Structural Characteristics of a Divided Wall Column Using the Sloppy Distillation Arrangement,” Korean J. Chem. Eng., 28, 348–356 (2011)

Linnho�, B., H. Dunford and R. Smith; “Heat Integration of Distillation Columns into Overall Processes,” Chem. Eng. Sci., 38, 1175–1188 (1983)

Long, N. V. D. and M. Y. Lee; “Improved Energy E�ciency in Debottle-necking Using a Fully �ermally Coupled Distillation Column,” Asia-Pac. J. Chem. Eng., 6, 338–348 (2011)

Long, N. V. D. and M. Y. Lee; “Dividing Wall Column Structure Design Using Response Surface Methodology,” Comput. Chem. Eng., 37, 119–124 (2012)

Long, N. V. D., S. H. Lee and M. Y. Lee; “Design and Optimization of a Dividing Wall Column for Debottlenecking of the Acetic Acid Pu-ri�cation Process,” Chem. Eng. Process., 49, 825–835 (2010)

Manley, D. B.; “Deethanizer/Depropanizer Sequences with �ermal and �ermo-Mechanical Coupling and Component Distribution,” U.S. Patent, No. 5,673,571 (1997)

Mak, J.; “Con�gurations and Methods for Improved NGL Recovery,” U.S. Patent, No. 7,051,552 (2006)

Malinen, I. and J. Tanskanen; “�ermally Coupled Side-Column Con-�gurations Enabling Distillation Boundary Crossing. 1. An Over-view and a Solving Procedure,” Ind. Eng. Chem. Res., 48, 6387–6404 (2009)

Midori, S., S. N. Zheng and I. Yamada; “Analysis of Divided-Wall Col-umn for Extractive Distillation,” Kagaku Kogaku Ronbunshu, 26, 627–632 (2000)

Montgomery, D. C.; Design and Analysis of Experiments, 3rd ed., pp. 197–386, John Wiley & Sons, New York, U.S.A. (1991)

Nejad, S. J., H. Abolghasemi, M. A. Moosavian and M. G. Maragheh; “Fractional Factorial Design for �e Optimization of Supercriti-cal Carbon Dioxide Extraction of La3+, Ce3+ and Sm3+ Ions from a Solid Matrix Using bis(2,4,4-trimethylpentyl)Dithiophosphin-ic Acid+Tributylphosphate,” Chem. Eng. Res. Des., 89, 827–835 (2011)

Olujić, Ž., M. Jödecke, A. Shilkin, G. Schuch and B. Kaibel; “Equip-ment Improvement Trends in Distillation,” Chem. Eng. Process., 48, 1089–1104 (2009)

Persson, K. and O. Åström; “Fractional Factorial Design Optimization of the Separation of Pilocarpine and Its Degradation Products by Capillary Electrophoresis.,” J. Chromatogr. B, 697, 207–215 (1997)

Peters, M. S. and K. D. Timmerhous; Plant Design and Economics for Chemical Engineers, 4th ed., pp. 523–536, McGraw-Hill, New York, U.S.A. (1991)

Premkumar, R. and G. P. Rangaiah; “Retro�tting Conventional Column Systems to Dividing Wall Columns,” Chem. Eng. Res. Des., 87, 47–60 (2009)

Schultz, M. A., D. G. Stewart, J. M. Harris, S. P. Rosenblum, M. S. Shakur and D. E. O’Brien; “Reduce Costs with Dividing Wall Col-umns,” CEP, 64–71 (2002)

Seidel, M.; “Verfahren zur Gleichzeitigen Zerlegung von Ver�üssig-ten Gasgemischen und anderen Flüssigkeitsgemischen mit mehr als zwei Bestandteilen durch Rekti�kation,” German Patent No.

294 Journal of Chemical Engineering of Japan

610,503 (1935)Sinnott, S. K.; Chemical Engineering Design, 4th ed., Coulson & Rich-

ardson’s Chemical Engineering Series, Vol. 6, pp. 566–568, Else-vier Butterworth Heinemann, Oxford (2005)

Smith, R.; Chemical Process Design, pp. 405–426, McGraw Hill, New York, U.S.A. (1995)

Tedder, D. W. and D. F. Rudd; “Parametric Studies in Industrial Distil-

lation: Part I. Design Comparisons,” AIChE J., 24, 303–315 (1978)Triantafyllou, C. and R. Smith; “�e Design and Optimization of Fully

�ermally Coupled Distillation Columns,” Chem. Eng. Res. Des., 70, 118–132 (1992)

Watkins, R. N.; Petroleum Re�nery Distillation, 2nd ed., pp. 27–60, Gulf Publishing, Houston, Texas, U.S.A. (1979)