Energy 35 (2010) 5309e5319

Contents lists available at ScienceDirect

Energy
journal homepage: www.elsevier.com/locate/energy

Optimal design of distillation column using three dimensional exergy

analysis curves
T.D. Khoa a, M. Shuhaimi a, *, H. Hashim b, M.H. Panjeshahi c
a
Process Development Group, Chemical Engineering Department, Universiti Teknologi PETRONAS, 31750 Tronoh, Perak, Malaysia
b
Faculty of Chemical and Natural Resources Engineering, Universiti Teknologi Malaysia, Malaysia
c
Department of Chemical Engineering, University of Tehran, Iran

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

Article history: This paper presents two main contributions. Firstly, a new exergy graphical method is proposed for
Received 11 February 2010 optimal design of distillation column with minimum exergy lost. The method is applicable to both grass-
Received in revised form root and retrofit cases, respectively. The effect of design and operating parameters of a distillation column
20 July 2010
on the exergy lost is graphically visualized by three dimensional exergy analysis curves. The curve shows
Accepted 22 July 2010
Available online 15 September 2010
the correlations between exergy lost, design and operating parameters of a distillation column. This
technique can be used as an effective method to reduce the simulation effort to search for the optimum
design and operating parameters of a distillation column at minimum exergy lost. Besides, visualization
Keywords:
Exergy lost
also enhances the engineers’ understanding of the column performance. The other contribution is a four-
Diabatic column level idealization concept, which is based on three dimensional graphical exergy analysis curves. The
Process simulation concept defines the effect of transport rate and configuration on exergy lost of distillation column. The
Process optimization effectiveness of the method has been demonstrated on a xylene column, which suggested that an
implementation of feed pre-heater yields a significant reduction in exergy lost by up to 15.5%.
Ó 2010 Elsevier Ltd. All rights reserved.

1. Introduction ecology together. Rasheva and Atanasova [5] evaluated the exergy
efficiency of sulphuric acid production while Chang et al. [6,7]
The method of exergy analysis is widely applied by many presented a two level idealization concept applied to exergy and
researchers in design and optimization of chemical processes since exergoeconomic analysis. A common goal of all the previous works
1970s. Exergy reflects the quality of energy flow in the whole [9e12,19e23,26e33] is improvement in exergy efficiency through
process. Exergy which is not harnessed is lost [13]. Exergy lost can the reduction exergy lost. However, some gaps still remain.
be divided into two parts: inevitable and avoidable. The inevitable First of all, there is no technique to realise the full potential for
exergy lost is also known as unavoidable exergy destruction, irre- exergy improvement since the effect of design and operating
versible exergy lost and intrinsic exergy lost. The exergy lost that is parameters of a distillation column such as number of tray, effi-
possible to be reduced is also known as avoidable exergy destruc- ciency of tray, relative volatility and reflux ratio on the exergy lost
tion, effluent exergy lost or extrinsic exergy lost. Besides, exergy cannot be visualized. Secondly, a two level idealization concept as
lost is also classified according to interaction between system and proposed by Chang and Li [6] has drawbacks when applied to
environment (internal and external exergy lost), or between one a distillation column since the focus was only on the number of
component and the whole plant (endogenous and exogenous tray, temperature of cooling water at condenser and steam pressure
exergy lost). at reboiler. In a case study on the distillation of benzeneetoluene,
Tsatsaronis et al. [1e3] estimated the potential for improving the optimum number of tray was obtained in 23 stages but the
the avoidable part of exergy lost in thermal and energy conversion strategy to achieve the solution was unclear. The method doesn’t
systems by splitting the exergy lost into two parts, which are the indicate the trend to optimum value, as well as the effect of design
endogenous/exogenous and unavoidable/avoidable parts. Ricardo and operating parameters on exergy lost of a column.
[4,8,24,25] conducted exergy analysis of petrochemical industries As a result, an improved graphical method is required to provide
using the Three-E model, which combines energy, economy and insights through simple visualization of a distillation column, which
is capable to simultaneously obtain an optimal design and reduce
* Corresponding author. Tel.: þ60 5 3687622; fax: þ60 5 3656176. exergy lost. The usefulness of the method is demonstrated in a case
E-mail address: [email protected] (M. Shuhaimi). study to optimize the design and operations of a xylene column.

0360-5442/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.energy.2010.07.032
5310 T.D. Khoa et al. / Energy 35 (2010) 5309e5319

Nomenclature Subscripts
act actual
a relative volatility av avoidable
b, q coefficient B bottom product
ExL exergy lost col column
H enthalpy D distillate product
h Murphree efficiency ext extrinsic
N number of tray F feed stream
R reflux ratio in inlet stream
S entropy int intrinsic
To ambient temperature L liquid
x liquid mole fraction min minimum
y vapour mole fraction out outlet stream
z feed mole fraction (total vapour and liquid) theory theory case
4 incline plane surface un unavoidable
V vapour
N infinite number

2. The new exergy analysis method where yn is the actual vapour molar composition at stage n and y*n is
the equilibrium composition. As shown in Fig. 1, let the true-
2.1. The specific factors of distillation column equilibrium curve be specified by a0 and the pseudo-equilibrium
curve be specified by ai. The subscript ‘i’ refers to index of the
The McCabeeThiele diagram provides a useful basic tool for pseudo-equilibrium curve at different relative volatilities such that
analysis of distillation column operation and design. For a binary i ¼ (1, 2.). A range of ai is between (1,a0). For distillation process,
system comprising component A and B, the true-equilibrium curve, the composition of feed stream, distillate and bottom products are
which comprises equilibrium vapour and liquid compositions over specified by process; hence points F, D and B are fixed. A simple
a range of boiling temperatures at specific pressure, is built from distillation column is divided into a top and a bottom section
experimental data. A pseudo-equilibrium curve is then redrawn depicted by upper and lower operating lines, respectively. The
between the operating lines and the true-equilibrium curve at upper operating line is drawn from point D based on Eq. (4):
different Murphree efficiency. The relative volatility is assumed to
be constant at given pressure, the true- or pseudo-equilibrium ynþ1 ¼ ½R=ðR þ 1Þxn þ xD =ðR þ 1Þ (4)
curves are also a function of relative volatility (a): The q-line, or feed line, is drawn from point F using Eq. (5):

  
y ¼ aA;B x= 1 þ aA;B  1 x (1) y ¼ ½q=ðq  1Þx  zF =ðq  1Þ (5)
In Eq. (1), the vapour composition (y) can be estimated from the where q ¼ (HV  HF)/(HV  HL). The lower operating line is drawn
liquid composition (x) and relative volatility (a). Relative volatility from point B to the intersection of upper operating line and q-line
is defined as ratio between vapoureliquid equilibrium ratios of by Eq. (6):
components A and B.
ynþ1 ¼ ½ðb þ 1Þ=bxn  xB =b (6)
aA;B ¼ ðyA =xA Þ=ðyB =xB Þ (2)
where b ¼ [R(zF  xB) þ q(xD  xB)  (xD  zF)]/(xD  zF). The detail
The Murphree efficiency is defined as: derivation of Eqs. (1e6) are explained in other literatures [14e17].
  The two operating lines can move upward or downward depending
h ¼ ðyn  ynþ1 Þ= y*n  ynþ1 (3) on reflux ratio and the thermal condition of feed stream, which is

1
True–equilibrium curve Vapor at the top (V1 )
D
Pseudo–equilibrium curve

Upper operating line

Stg. 1 Liquid at the top (reflux, L 0 )
q–line (feed line)

y
F
Feed (F)

Lower operating line

B Stg. n Vapor at the bottom (Vn+1)
0
0 xB zF xD 1
x Liquid at the bottom (Ln )

Fig. 1. The McCabeeThiele for operations of simple binary distillation mixtures. Fig. 2. Input and output streams in a column.
 T.D. Khoa et al. / Energy 35 (2010) 5309e5319 5311

Fig. 3. Ideal adiabatic column (infinite configuration & transport rate). Fig. 5. Finite configuration & infinite transport rate.

specified by q. The stair line that is limited by two operating lines section. This distillation column is defined as completely reversible
and the equilibrium curve shows the number of tray in the column. and diabatic [8,9,24]. Diabatic column conceptually uses a heat
Consequently, the parameters of distillation process depicted in exchanger at every stage. It means that heat transfer process at
McCabeeThiele diagram are composition of feed stream (zF), every stage is significant, and consequently, the exergy lost of the
distillate (xD) and bottom (xB), the relative volatility (a0, a1, a2.ai), distillation column is zero.
reflux ratio (R), q-value, Murphree efficiency (h) and number of tray In industry, the adiabatic column is more popular because of
(N). These specific factors effect on the column is the exergy lost. cheaper investment and easier operation. This distillation column
Fig. 2 shows the input and output streams of a simple column. contains only one condenser at top and only one reboiler at bottom.
The exergy lost of column (ExLcol), which is the sum of exergy lost at The heat transfer process at every stage is negligible. This causes
every stage, can be estimated as follow: the operating lines to move apart from the true-equilibrium curve,
X X thus causing the lost of exergy. This amount of exergy lost has to be
ExLcol ¼ ðExin Þ  ðExout Þ accepted when using adiabatic column.
However, the exergy lost will be at minimum when the oper-
¼ ðHF þ HL0 þ HVnþ1  HLn  HV1 Þ  To ðSF þ SL0 þ SVnþ1
ating lines touch the true-equilibrium curve as shown on Fig. 3. This
 SLn  SV1 Þ condition refers to an infinite transport rate with perfect efficiency
(7) (h ¼ 1) and infinite configuration (NN). Furthermore, reflux ratio
reaches minimum value (Rmin 0). The upper operating line that
where, H, S and To are enthalpy flow, entropy flow and ambient crosses the intersection between feed line and true-equilibrium
temperature, respectively. curve determines the minimum reflux ratio. Therefore, exergy lost
is minimum and is called exergy lost for ideal adiabatic,
2.2. Operating condition and exergy lost of distillation column ExLadiabatic ¼ ExL(a0,Rmin 0).
When transport rate is limited, the Murpree efficiency decreases
For reversible operation, where the operating lines and the true- (h < 1) as shown in Fig. 4. The true-equilibrium curve moves to
equilibrium curve coincide, there are infinite number of trays. The pseudo-equilibrium curve as specified by ai. The result is an
condenser at top is combined into a single rectifying section, and increase of exergy lost although the reflux ratio reaches Rmin i. The
similarly, the reboiler at bottom is merged into a single stripping exergy lost is defined as ExL(ai,Rmin i), where i refers to index of

Fig. 4. Infinite configuration & finite transport rate. Fig. 6. Finite configuration & infinite transport rate.
5312 T.D. Khoa et al. / Energy 35 (2010) 5309e5319

Reflux ratio (R)

0 1 2 3 4 5 η
0
1
2
0.95
4 0.9
Direction of decreasing
6 efficiency 0.85
8 0.8
10 0.75

Nu mb er o f tray (N)
12 0.7
14 0.65
16 0.6
18 0.55
20 0.5
22 0.45
24 0.4
26 0.35
28 0.3
30 0.25
Fig. 7. Finite configuration & infinite transport rate.
Fig. 9. Correlation between efficiency, reflux ratio, number of tray (benzeneetoluene
pseudo-equilibrium curve. In contrast, when only the configuration system).
is limited while the efficiency is perfect, then the exergy lost will
increase as shown Figs. 5e7. In this case, the number of tray is number of tray reaches minimum value as reflux ratio goes to
theoretical, Ntheory. The exergy lost is defined as ExL(a0,Rmin i) in infinity. These correlations are validated with the data published by
Fig. 5, ExL(a0,Ract) in Fig. 6 and ExL(a0,Rtheory) in Fig. 7, respectively. McCabe et al. [17] with 2% maximum error. In the validation, the
In an actual case as shown in Fig. 8, the distillation column cannot reflux ratio is in the range of 1.45e2.16, the range of Murphree
operate at Ract  Rmin i because of the limitation of both transport efficiency is between 35% and 100% and the number of tray varies
rate (h < 1, at ai) and configuration due to the actual number of tray, from 10 to 30.
Nact. The exergy lost is defined as actual exergy lost, ExLact ¼ ExL The correlations between ExL, R, N and h are shown in Fig. 10. For
(ai,Ract). constant number of tray, the correlation between exergy lost and
reflux ratio is linear. Since, minimum exergy lost is at perfect effi-
ciency (h ¼ 1), decreasing efficiency increases reflux ratio and
2.3. The correlation between exergy lost, reflux ratio, number of
exergy lost. An important observation is that all the lines corre-
tray and efficiency
sponding to different number of tray are coincident.
In general, the degree of freedom for the correlations between
The correlation between exergy lost (ExL) from one side and
reflux ratio, number of tray and efficiency is two. The exergy lost of
reflux ratio, number of tray and efficiency on the other side, are
column at given feed condition can be determined by a simplified
studied in benzeneetoluene system. The feed stream at 92  C, 1 atm
function of reflux ratio, in which reflux ratio is as function of the
and 1000 kmol/h is composed of equimolar mixture of benzene and
number of tray and efficiency.
toluene. The distillate and bottom product compositions are 95 mol
% benzene and 90 mol% toluene, respectively. The system is simu-
2.4. The three dimensional exergy analysis curves
lated by Aspen HYSYS version 2006 [18]. The simulation results in
Fig. 9 shows that for constant number of trays, increasing reflux
An operating state of distillation column that is specified by
ratio reduces the efficiency. In case of constant reflux ratio, the
design and operating parameters (reflux ratio, number of tray,
higher number of tray, the lower is the efficiency. For constant
efficiency, increasing the reflux ratio decreases the number of tray.
Reflux ratio reaches minimum value as infinite stages. Similarly, 3500

3000
N
8
2500
Ex erg y lo st o f co lu mn (Ex L,M J/h )

10
12
2000 14
N=8; η = 1 16
18
1500
20
22
1000
N=10; η = 1 24
26
28
500 N=20; η = 1
30
N=30; η = 1
0
0 1 2 3 4 5 6
Reflux ratio (R)

Fig. 10. The correlation between exergy lost, reflux ratio, number of tray, efficiency
Fig. 8. Actual distillation (finite configuration & transport rate). (benzeneetoluene system).
 T.D. Khoa et al. / Energy 35 (2010) 5309e5319 5313

ND < NC < NB ¼ NE ¼ Nact < NA ¼ NF ¼ NN (8)

The a-curves indicate equilibrium curves. This curve will move
up as relative volatility decreases. Therefore, the lowest curve
corresponds to the true-equilibrium curve (a0). The a-curves are
all parallel and the vertical distance of two curves represents the
exergy lost between the two pseudo-equilibrium curves. All
the points in one a-curve are the same with respect to the
relative volatility (equilibrium curve) as points A, B, C and D,
points E and F.

1 < aE ¼ aF ¼ ai < aA ¼ aB ¼ aC ¼ aD ¼ a0 (9)

The h-curves indicate Murphree efficiency. This curve will move up
as efficiency decreases. Therefore, the lowest curve corresponds to
perfect efficiency (h ¼ 1). All the points in one h-curve are the same
with respect to the efficiency value as points A, B, C and D.

hF < hE < hA ¼ hB ¼ hC ¼ hD ¼ 1 (10)

All the points in the same height from XeY plane are the same with
respect to the reflux ratio and exergy lost of column as points D and
E, points C and F.
Fig. 11. The three dimensional exergy analysis curves.

RA ¼ Rmin 0 < RB ¼ Rtheory < RC ¼ RF ¼ Rmin i < RD ¼ RE

Murphree efficiency, relative volatility, and exergy lost) is repre-
sented by one point in a system of three axes as shown in Fig. 11. ¼ Ract ð11Þ
The axes X, Y, Z are reflux ratio, number of tray and exergy lost of
 
column, respectively. In constructing the three dimensional exergy
ExLA ¼ ExLadiabatic ¼ ExLðRmin 0 Þ < ExLB ¼ ExL Rtheory
analysis curves, reflux ratio, number of trays and Murphree effi-
ciency are pre-defined for the simulator, while exergy lost of < ExLC ¼ ExLF ¼ ExLðRmin i Þ < ExLD ¼ ExLE
column is calculated by Eq. (7). The data of enthalpy flow, entropy ¼ ExLðRact Þ ð12Þ
flow and flow rate are estimated by simulator using Pen-
geRobinson thermodynamic property package. The Three dimen-
sional exergy analysis curves, which are constructed by collection
all points from the simulation, provides a visualization of an
operating state in a distillation column. All curves are in an incline
plane surface (4), which is perpendicular to XeZ plane, and is
limited by true-equilibrium curve a0 at perfect efficiency (h ¼ 1).
The Figs. 3e8, which represent the critical states of distillation
column, are depicted by points A, F, C, D, B and E in Fig. 11,
respectively. The N-lines indicate number of tray. All the points in
one N-line are the same with respect to the number of tray as points
B and E, points A and F.

Fig. 13. The three dimensional exergy analysis curves for benzeneetoluene distillation
Fig. 12. Correlation between efficiency, reflux ratio, number of tray, relative volatility. column.
5314 T.D. Khoa et al. / Energy 35 (2010) 5309e5319

Actual operation (point E)

ExLext ExLav
4th level idealization Type II
rd
3 level idealization Type I (point F)
ExLact (Thermodynamic equilibrium ExLun
operation, point B) ExLint

2nd level Ideal adiabatic Column operation

ExLadiabatic (point A)

1st level idealization Reversible operation

(Reversible Diabatic Column)

Fig. 14. The four-level idealization concept.

The projection of (4) plane into XeZ plane is only one line as Fig. 14. In this figure, each idealization level corresponds to one
Fig. 10, which expresses the linear correlation of exergy lost and operation state of distillation column. The 1st level is reversible
reflux ratio. The projection of (4) plane into XeY plane as Fig. 12 is operation, in which the operating lines and the true-equilibrium
the same form of Fig. 9, which expresses the relations of four curve are coincident. This means, it is a reversible diabatic column.
parameters, namely reflux ratio, number of tray, relative volatility The 2nd level idealization is ideal adiabatic column as point A in
and Murphree efficiency. Fig. 11, the 3rd level idealization is thermodynamic equilibrium
The (4) plane moves downward as decreasing of q-value. This operation with infinite transport rate h ¼ 1 and finite configuration
means that there are many parallel planes corresponding to q- Nact (see point B in Fig. 11). Finally, the 4th level idealization is
values. In general, an operating state of distillation column is infinite configuration NN and finite transport rate h < 1 (see point F
verified by three degrees of freedom. One is the q-value while the in Fig. 11). For actual operation, both configuration and transport
other two can be any from the five parameters: reflux ratio, number rate are limited, Nact and h < 1 as depicted on point E in Fig. 11.
of tray, relative volatility, Murphree efficiency and exergy lost. On one hand, exergy lost of a distillation column can be reduced
Therefore, for variable q-value, the exergy lost of column is a func- from actual level to the third level by transport rate improvement.
tion of q-value and reflux ratio. This is because the difference between actual and third levels is the
The three dimensional exergy analysis curves of benze- amount of exergy lost caused by limitation of transport rate or tray
neetoluene system are plotted in Fig. 13. Properties of feed and efficiency. On the other hand, the difference between actual and the
products streams are described in Section 2.3. The range of reflux fourth level is the amount of exergy lost caused by limitation of
ratio is between 1.29 and 5.00. The Murphree efficiency is ranges configuration or the number of trays. As a result, exergy lost of
from 30 to 100% and the number of tray is between 8 and 30. column can be reduced from actual level to fourth level by
improving the configuration such as increasing number of tray.
2.5. The four-level idealization concept

Based on the three dimensional exergy analysis curves, four 3. Comparison with Chang & Li and Tsatsaronis & Park
idealization levels of distillation column are defined, as shown in concepts

In the concept by Chang & Li [6], the intrinsic exergy lost (ExLint)
Table 1
is defined as the exergy destruction of the process system under
Comparison results from Chang’s work and three dimensional exergy analysis
method. thermodynamic equilibrium operation, indicating perfect effi-
ciency as shown in Fig. 7. This condition is graphically visualized as
Specifications Chang’s work Three dimensional
point B in Fig. 11. Therefore, the extrinsic exergy lost (ExLext) is the
(CHEMCADÔ exergy
V.5.06) analysis method difference between actual exergy lost and intrinsic exergy lost:
(Aspen HYSYS version  
2006) ExLint ¼ ExL Rtheory ¼ ExLB (13)
Initial data
Number of stages 14 14  
Efficiency of trays (%) 100 100 ExLext ¼ ExLE  ExLB ¼ ExLðRact Þ  ExL Rtheory (14)
Feed flow rate (kmol/h) 1000 1000
Temperature of feed (K) 298 298
Pressure of feed (atm) 1 1
Eq. (14) is written for constant N.
Mole fraction of benzene 0.50 0.5
in feed
Mole fraction of toluene 0.50 0.5 Table 2
in feed Simulated data of benzeneetoluene distillation column.
Mole fraction of benzene 0.95 0.95
No. stages Minimum reflux ratio Minimum exergy lost Exergy saving (%)
in distillate
(MJ/h)
Mole fraction of toluene 0.90 0.90
in bottom 14 1.423 758
16 1.354 706 6.86
Calculated results Error (%) 18 1.315 676 10.82
Reflux ratio 0.929 0.941 1.27 20 1.292 660 12.93
Distillate flow rate (kmol/h) 525 530 0.94 22 1.279 650 14.25
Bottom flow rate (kmol/h) 475 470 1.05 24 1.271 644 15.04
Condenser duty (MJ/h) 31364.70 31710.62 1.09 26 1.266 640 15.57
Reboiler duty (MJ/h) 41573.43 42060.43 1.15 28 1.263 638 15.83
Intrinsic exergy lost (MJ/h) 1752.67 1730.57 1.26 30 1.261 637 15.96
 T.D. Khoa et al. / Energy 35 (2010) 5309e5319 5315

225 221 212

236 238 231 233 BFW Steam

Cooler –1(53.5MW) Cooler – 2 (17.649MW) Cooler – 3 (15.6MW)

222 213
226
209

RECEIVER

236
239 247

201
XYLENE COLUMN

160

285

Furnace (99.132MW)

260
252 254

Fig. 15. Process flow diagram for system xylene column.

Table 1 shows the comparison between the method by Chang & Based on property of the three dimensional exergy analysis
Li and the method of three dimensional exergy analysis curves for curves in Fig. 11, the relations between intrinsic exergy lost and
benzeneetoluene column. The error of intrinsic exergy lost is 1.26% unavoidable exergy destruction follows Eq. (17)e(21):
and hence negligible.
In optimizing benzeneetoluene column, Chang considers ExLE  ExLB ¼ ExLF  ExLA (17)
number of tray, cooling water temperature in condenser and steam or
pressure in reboiler as variables. Base on the results, the column
exergy lost was shown to be solely dependent on the number of ExLact  ExLint ¼ ExLun  ExLadiabatic (18)
tray where it decreases as the number of tray increases from 17 to
23. However, Chang’s selection of the optimum number of tray to or
be at 23 is quite arbitrary. Table 2 list the limitation of operating
ExLext ¼ ExLun  ExLadiabatic (19)
conditions of benzeneetoluene column as variation in number of
trays. Increasing number of tray decreases minimum reflux ratio or
and exergy lost of column, consequently the quantity of exergy  
saving will be raised. The results show correlations of exergy lost, ExLðRact Þ  ExL Rtheory ¼ ExLðRmin i Þ  ExLðRmin 0 Þ (20)
number of tray and reflux ratio, as well as trend of minimizing
exergy lost. or
In the concept proposed by Tsatsaronis & Park [1], the
unavoidable exergy destruction (ExLun) corresponds to the oper- Ract  Rtheory ¼ Rmin i  Rmin 0 (21)
ating condition shown in Fig. 4. This condition is graphically visu-
Both of the previous methods do not show the effect of design
alized as point F in Fig. 11. Therefore, the avoidable exergy
parameters on exergy lost. In this work, the method of three
destruction (ExLav) is the difference between actual exergy lost and
dimensional exergy analysis curves shows that reducing reflux
the unavoidable exergy destruction:
ratio or increasing number of trays will decrease column exergy
ExLun ¼ ExLðRmin i Þ ¼ ExLF (15) lost, as well as correlations of design and operating parameters.
Moreover, trend of minimum exergy lost is clearly visualized in the
three dimensional exergy analysis curves. Consequently, the full
ExLav ¼ ExLE  ExLF ¼ ExLðRact Þ  ExLðRmin i Þ (16)
potential for exergy lost improvement on a distillation column may
Eq. (16) is written for constant a. be fully exploited.
5316 T.D. Khoa et al. / Energy 35 (2010) 5309e5319

Table 3
Compositions of feed streams (base case) in xylene column.

No Compositions, 160 201 No Compositions, 160 201

kmol/h kmol/h
1 Cyclohexane 0.00 0.03 23 Indane 5.38 0.00
2 Benzene 0.00 0.42 24 n-Decane 1.78 0.00
3 Toluene 3.37 11.11 25 Iso butyl 0.19 0.00
benzene
4 2-Methyl 0.03 0.00 26 Sec butyl 0.19 0.00
heptane benzene
5 4-Methyl 0.02 0.00 27 Meta diethyl 0.91 0.00
heptane benzene
6 3,4-Dimethyl 0.05 0.00 28 1-Methyl-3-n-propyl 3.06 0.00
hexane benzene
7 3-Methyl 0.01 0.00 29 n-Butyl benzene 0.19 0.00
heptane
8 n-Octane 0.11 0.00 30 1-Methyl-4-n-propyl benzene 1.25 0.00
9 Iso propyl 0.08 0.55 31 Para diethyl benzene 0.37 0.00
cyclo Pentane
10 Ethyl Benzene 56.78 28.52 32 1-Methyl-2-n-propyl benzene 0.64 0.00
11 Para xylene 151.23 474.35 33 5-Ethyl meta xylene 4.85 0.00
12 Meta Xylene 343.16 1071.79 34 2-Ethyl para xylene 3.50 0.00
13 Ortho Xylene 166.88 453.82 35 4-Ethyl meta xylene 3.15 0.00
14 3-Ethyl heptane 0.34 1.03 36 4-Ethyl ortho xylene 6.12 0.00
15 Cumene 2.92 0.00 37 1,2-Dimethyl-3-ethyl benzene 1.54 0.00
16 n-Propyl benzene 19.67 0.00 38 1,2,4,5-Tetramethyl benzene 3.97 0.00
17 Meta ethyl toluene 38.33 0.62 39 1,2,3,5-Tetramethyl benzene 2.78 0.00
18 Para ethyl toluene 17.94 0.00 40 1,2,3,4-Tetramethyl benzene 2.98 0.00
19 Ortho ethyl toluene 13.33 0.00 41 Naphthalene 6.53 0.78
20 1,3,5-Trimethyl benzene 39.26 10.10 42 C11 aroma 6.86 0.00
21 1,2,4-trimethyl benzene 109.96 0.00 43 C12 aroma 1.23 0.00
22 1,2,3-Trimethyl benzene 20.88 0.00 Total mole flow, kmol/hr 1041.82 2053.12

4. Case study e xylene column compositions of two feed streams and properties of process
streams are listed in Tables 3 and 4.
4.1. Description of system
4.2. Results and discussion
The xylene column, which is described in Fig. 15, is a unit
operation in an integrated refinery e petrochemical plant operating
The plots of reflux ratio versus efficiency and reflux ratio versus
at 2671 MT/year. The column consists of 115 sieve trays, and two
products purification are depicted in Figs. 16 and 17, respectively.
feed streams at tray 23 and 49. The C9-Aromatic composition is
Decreasing of reflux ratio increases efficiency but decreases the
0.0005 mass fraction in distillate stream, the xylene composition is
products purification. Distillate is the main product in this case.
0.005 mass fraction in bottom stream. The condenser system
Thus, purification of distillate product is very important to the
consists of three coolers, namely Cooler-1, Cooler-2 and Cooler-3. A
plant. Correlation between exergy lost and reflux ratio is linear as
furnace supplies energy at the bottom of the column. The
shown in Fig. 18. Decreasing of reflux ratio decreases all the exergy
lost in the column, condenser and reboiler. The summation of all
exergy lost reflects the exergy lost of overall system. Therefore,
Table 4
Real data of process streams (Base case) in xylene column. reflux ratio is an important factor because it affects the exergy lost
in column, condenser and reboiler. Based on the above results, the
Stream No. Temperature,  C Pressure, kg/cm2 g Flow rate, kg/h
201 197.2 10.99 218,000
160 193.5 10.19 115,877 100
209 239.3 7.14 1,070,032
225 239.3 7.51 659,920 95
226 223.0 6.64 659,920
236 218.1 1.22 2,249,013 90
238 218.4 1.20 2,249,013
85
221 239.3 7.51 217,700
Efficiency (%)

222 212.0 6.64 217,700

80
231 206.0 0.79 725,028
233 208.6 0.77 725,028 75
212 239.3 7.40 192,412
213 229.0 6.64 192,412 70
BFW 142.8 73.90 25,500
Steam 209.6 18.40 25,500 65
236 223.0 6.64 1070,031
60
239 223.0 6.64 777,432
247 223.0 6.64 292,599 55
252 284.4 8.35 2243,339 0 1 2 3 4 5 6
254 284.4 8.35 41,278
Reflux ratio (R)
260 284.4 8.35 2202,061
285 294.8 8.66 2202,061
Fig. 16. Correlation between efficiency and reflux ratio at 115 stages.
 T.D. Khoa et al. / Energy 35 (2010) 5309e5319 5317

0.002 xb 120
C 9 composition in distillate (mass fraction)

0.0018 0.0005
0.0006
0.0016 0.0007 100
0.0008
0.0014 0.0009
0.0012 Direction of decreasing 0.001 80
0.002
0.001
xylene composition in

No . stag es
0.003
bottom, xb 0.004
0.0008 60
0.005
0.0006 0.006
0.007
0.0004 40
0.008
0.009
0.0002
0.01
0.02 20
0
0 1 2 3 4 5 6 7 8 0.03
0.04
Reflux ratio
0.05 0
0.0000 0.1000 0.2000 0.3000 0.4000 0.5000
Fig. 17. Correlation between reflux ratio and purification of distillate, bottom product. Exergy lost, MW

Fig. 19. Profile of exergy lost of column in base case.

column should be operated with reflux ratio as low as possible

70 while keeping the purity of distillate product at an acceptable level.
In the base case (as real case), the column parameters in the
three dimensional exergy analysis curves are reflux ratio (2.63),
60
efficiency of tray (85%), number of tray (115), distillate at 0.0005
mass fraction of C9-Aromatic, bottom product at 0.005 mass
50 system
Table 6
Exergy lost (MW) of different cases in xylene column.
40
ExL (MW)

Base case R ¼ 2.63

Efficiency of trays: 85%
30 Temperature of feed streams
reboiler Stream 201: 197.2  C
Stream 160: 193.5  C
20 Properties of products
condenser
Mass fraction of C9-aromatic in distillate: 0.0005
Mass fraction of xylene in bottom: 0.005
Condenser (total) 5.082
10
Feed heater 0
column Column 4.035
Furnace 17.365
0
Total 26.482
0 2 4 6 8
Reflux ratio Case 1 R ¼ 2.18
Efficiency of trays: 85%
Fig. 18. Correlation between exergy lost and reflux ratio at N ¼ 115; h ¼ 85%. Temperature of feed streams
Stream 201: 197.2  C
Stream 160: 193.5  C
Properties of products
Table 5
Mass fraction of C9-Aromatic in distillate: 0.0005
Comparison real data and simulated data of base case in xylene column.
Mass fraction of Xylene in bottom: 0.02
Specifications Plant data Simulated Error (%) Condenser (total) 4.408
data (HYSYS V.2006) Feed heater 0
Column 3.478
Number of stages 115 115
Furnace 15.992
Efficiency of trays (%) 85 85
Total 23.878 (saving 9.8%)
Reflux ratio 2.63 2.63
Mass fraction of C9-aromatic 0.0005 0.0005 Case 2 R ¼ 2.26
in distillate Efficiency of trays: 85%
Mass fraction of xylene in 0.005 0.005 Temperature of feed streams (at boiling point)
bottom Stream 201: 262.4  C
Distillate flow rate (kg/h) 292,599 293,152 0.19 Stream 160: 268  C
Bottom flow rate (kg/h) 41,278 40,631 1.57 Properties of products
Total condenser duty (MW) 86.749 85.893 0.98 Mass fraction of C9-aromatic in distillate: 0.0005
Reboiler duty (MW) 99.132 98.235 0.90 Mass fraction of xylene in bottom: 0.005
Exergy lost total (MW) n/a 26.482 Condenser (total) 4.516
Condenser n/a 5.082 Feed heater 1.171
Feed heater n/a 0 Column 2.658
Column n/a 4.035 Furnace 14.034
Furnace n/a 17.365 Total 22.379 (saving 15.5%)
5318 T.D. Khoa et al. / Energy 35 (2010) 5309e5319

60 be kept constant. In Fig. 17, reflux ratio is decreased to 2.18 at the

q same distillate purity and exergy lost as shown in Table 6. More-
50 1.71 over, as shown in Fig. 20, the thermal condition of the feed stream
1.00 affects the exergy lost of overall system, including exergy lost of
0.90 feed pre-heater. Reducing q-value decreases exergy lost, because
40 0.80
the required energy is supplied at feed and bottom. This causes the
0.70
ExL, MW

0.60
column to reach diabatic state. Feed stream should be saturated
30
0.50 liquid (q ¼ 1.00) because of hydraulic stability of the column. The
0.40 three dimensional exergy analysis curves for the xylene column is
20 0.30 shown in Fig. 21. In the optimum case, installation of feed pre-
Direction of 0.20
heater together with the reduction in reflux ratio to 2.26, result in
0.10
10
decreasing q 0.00 the exergy lost to be reduced to 15.5% as shown in Table 6.

0
5. Conclusion
0 1 2 3 4 5 6
Reflux ratio In general, exergy lost of a distillation column depends on the
feed conditions, purity of the distillate and bottom products,
Fig. 20. Correlation between exergy lost of system, reflux ratio and q-value.
number of tray, efficiency and reflux ratio. When the input data of
a distillation process is fixed, meaning that composition of feed,
distillate and bottom products are pre-defined, and the three
fraction of xylene, and exergy lost. These parameters are listed in dimensional exergy analysis curves can be constructed with three
Table 5. The profile of column exergy lost is represented in Fig. 19. degrees of freedom. Therefore, the exergy lost of column is a func-
Most of the losses are located at the two feed trays and at the tion of q-value and reflux ratio, in which reflux ratio is a function of
bottom section trays. The optimum case is selected from the two out of three parameters, i.e. number of tray, Murphree effi-
correlation between reflux ratio and products purity. Distillate is ciency and relative volatility. In practice, exergy lost of distillation
main product of the process, so the purity of distillate (or xD) must column depends on reflux ratio and thermal condition of feed or q-

Fig. 21. The three dimensional exergy analysis curves for xylene column.
 T.D. Khoa et al. / Energy 35 (2010) 5309e5319 5319

value. The optimum reflux ratio should be selected in order to [9] Rosjorde A, Kjelstrup S. The second law optimal state of a diabatic binary tray
distillation column. Chemical Engineering Science 2005;60:1199e210.
maintain purification of the products.
[10] Antonio BA, Romildo PB, Luís SV. Exergy analysis of distillation processeseA
A new method of three dimensional exergy analysis curves is case study. Energy 2007;32:1185e93.
proposed in this paper and carries the potential to become a signifi- [11] Ishida M, Taprap R. Application of energy e utilization diagram for graphic
cant method in the design and optimization of a distillation column. exergy analysis of multicomponent distillation columns. Journal of Chemical
Engineering of Japan 1992;25(4):396e402.
It is a convenient graphical method that is able to effectively illus- [12] Budiman A, Ishida M. Three-dimensional graphical exergy analysis of
trate the effects of the design and operating parameters on exergy a distillation column. Journal of Chemical Engineering of Japan 1996;29
lost. This paper also demonstrates that the new method can be (4):662e8.
[13] Desmond EW. Advanced thermodynamics for engineers. London: Arnold;
conveniently used to determine the optimum values for the column 1997.
operating variables as well as quantifying intrinsic and extrinsic [14] Perry RH, Green DW, Maloney JO. Perry’s chemical engineerings’ handbook.
exergy destruction together with avoidable and inevitable exergy 7th ed. New York: McGraweHill; 1997.
[15] Richardson JF, Harker JH, Backhurst JR. Coulson and Richardson’s chemical
lost. The application of the method to optimize xylene column engineering e volume 2: particle technology and separation processes. 5th ed.
results in significant 15.5% decrease in exergy lost. This is achieved by Great Britain: Butterworth Heinemann; 2002.
reduction of reflux ratio from 2.63 to 2.18, and implementation of [16] Page SB, William LL, Joseph PS. Design of distillation column control systems.
New York: Edward Arnold; 1985.
a feed heater. [17] McCabe WL, Smith JC, Harriott P. Unit operations of chemical engineering. 7th
This work has also introduced the four-level idealization ed. New York: McGraw e Hill; 2005.
concept of distillation column operations. While the previous [18] Aspen HYSYS version (20.0.0.6728). Cambridge, Massachusetts U.S.A: Aspen
Technology, Inc.; 2006.
three-level idealization concept corresponds with unlimited
[19] Kim S-M, Oh S-D, Kwon Y-H, Kwak H-Y. Exergoeconomic analysis of thermal
transport rate and limited configuration, the four-level idealization systems. Energy 1998;23:393e406.
corresponds with limited transport rate and unlimited configura- [20] Bejan A, Dan N, Cacuci DG, Schutz W. Exergy analysis of energy conversion
tion. The concept shows that the effects of rate phenomena and during the thermal interaction between hot particles and water. Energy
1998;23:913e28.
configuration on exergy lost of distillation process are quantified, [21] Rosen MA. Second-law analysis of aquifer thermal energy storage systems.
providing engineers with better understanding to make decisions Energy 1999;24:167e82.
in the design and operations of distillation columns. [22] Leites IL, Sama DA, Lior N. The theory and practice of energy saving in the
chemical industry: some methods for reducing thermodynamic irreversibility
in chemical technology processes. Energy 2003;28:55e97.
Acknowledgements [23] Kwak H-Y, Kim D-J, Jeon J-S. Exergetic and thermoeconomic analyses of
power plants. Energy 2003;28:343e60.
[24] Rivero R, Garcia M, Urquiza J. Simulation, exergy analysis and application of
Financial supports from Universiti Teknologi PETRONAS and diabatic distillation to a tertiary amyl methyl ether production unit of a crude
Ministry of Higher Education Malaysia for the project are oil refinery. Energy 2004;29:467e89.
acknowledged. [25] Ricardo R, Consuelo R, Salvador G. Exergy and exergoeconomic analysis of
a crude oil combined distillation unit. Energy 2004;29:1909e27.
[26] Budiman A, Ishida M. A new method for disclosing internal phenomena in
References a distillation column by use of material-utilization diagram. Energy
2004;29:2213e23.
[1] Tsatsaronis G, Park MH. On avoidable and unavoidable exergy destructions [27] Morosuk T, Morosuk C, Feidt M. New proposal in the thermodynamic analysis
and investment costs in thermal systems. Energy Conversion and Manage- of complex heat regeneration systems. Energy 2004;29:2517e35.
ment 2002;43:1259e70. [28] Martin C, Villamanan MA, Chamorro CR, Otero J, Cabanillas A, Segovia JJ. Low-
[2] Morosuk T, Tsatsaronis G. A new approach to the exergy analysis of absorption grade coal and biomass co-combustion on fluidized bed: exergy analysis.
refrigeration machines. Energy 2008;33:890e907. Energy 2006;31:330e44.
[3] Frank C, Tsatsaronis G, Gao Z. Avoidable thermodynamic inefficiencies and costs [29] Maruoka N, Akiyama T. Exergy recovery from steelmaking off-gas by latent
in an externally fired combined cycle power plant. Energy 2006;31:1472e89. heat storage for methanol production. Energy 2006;31:1632e42.
[4] Ricardo R. Application of the exergy concept in the petroleum refining and [30] Vijayaraghavan S, Goswami DY. A combined power and cooling cycle modi-
petrochemical industry. Energy Conversion and Management 2002;43: fied to improve resource utilization efficiency using a distillation stage. Energy
1199e220. 2006;31:1177e96.
[5] Rasheva DA, Atanasova LG. Exergy efficiency evaluation of the production of [31] Zhu P, Feng X. Allocation of cumulative exergy in multiple product separation
sulfuric acid from liquid sulfur. Exergy, an International Journal 2002;2:51e4. processes. Energy 2007;32:137e42.
[6] Chang H, Li J. A new exergy method for process analysis and optimization. [32] Suphanit B, Bischert A, Narataruksa P. Exergy loss analysis of heat transfer
Chemical Engineering Science 2005;60:2771e84. across the wall of the dividing-wall distillation column. Energy
[7] Chang H, Chuang SC. Process analysis using the concept of intrinsic and 2007;32:2121e34.
extrinsic exergy losses. Energy 2003;28:1203e28. [33] Huang K, Shan L, Zhu Q, Qian J. Design and control of an ideal heat-integrated
[8] Ricardo R. Exergy simulation and optimization of adiabatic and diabatic binary distillation column (ideal HIDiC) system separating a close-boiling ternary
distillation. Energy 2001;26:561e93. mixture. Energy 2007;32:2148e56.