Energy 35 (2010) 1224–1231

Contents lists available at ScienceDirect

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

Exergetic analysis of the refrigeration system in ethylene and propylene

production process
F.M. Fábrega, J.S. Rossi, J.V.H. d’Angelo*
Department of Chemical Systems Engineering, University of Campinas, UNICAMP, Chemical Engineering School, P.O. Box 6066, 13083 970, Campinas-SP, Brazil

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

Article history: The exergetic analysis is a tool that has been used successfully in many studies aiming a more rational
Received 29 April 2009 energy use reducing the cost of the processes. With this analysis it is possible to perform an evaluation of
Received in revised form the overall process, locating and quantifying the degradation of exergy. In this context, the present work
30 October 2009
aimed the exergetic analysis of the refrigeration cycles in ethylene and propylene production process,
Accepted 2 November 2009
calculating the loss of exergy, in order to propose changes in the operational variables of the cycles used,
Available online 12 November 2009
trying to reduce the rate of destroyed exergy in the process. The commercial simulator HysysÓ (version
3.2) was used to obtain thermodynamic properties of the process streams and to perform mass and
Keywords:
Refrigeration energy balances. The application of new operational conditions in these cycles resulted in a reduction of
Ethylene about 13% of the losses of exergy for the refrigeration system of the process.
Propylene Ó 2009 Elsevier Ltd. All rights reserved.
Exergy

1. Introduction variables of the process, aiming an economy of energy consump-

tion. With this analysis it is possible to evaluate the individual
Many chemical processes use large scale refrigeration systems performance of each equipment or the general performance of the
that generate cold utilities which are essential in different stages of entire process. The exergetic analysis has been applied with great
the process. Some important examples of these processes are: the success by many authors in different processes [2–14].
production of polyethylene, polypropylene, PVC, PET, ethylene and In this context, the goal of this work is to perform the exergetic
propylene. These refrigeration systems are great energy consumers analysis of the refrigeration cycles used in ethylene and propylene
and the costs of the compression and condensation steps have production process to produce cold utilities which are then used in
a great impact on the cost of the final products. Therefore, it is very the condensers of the distillation columns. This analysis identifies
important that these refrigeration cycles operate in an optimized and quantifies the exergy losses (irreversibilities) with which it is
way in order to reduce production costs. possible to propose changes in the operational conditions of the
Ethylene and propylene, obtained from the pyrolysis of naphtha, refrigeration cycle, minimizing the rates of destroyed exergy in the
are two of the most important products in the petrochemical process.
industries. Nowadays almost 60% of the ethylene world production
use naphtha as raw material and in Brazil this percentage is 90% [1], 2. Methodology
used to produce 3.7 million tons of ethylene per year and 36% of this
are produced by Quattor Petrochemical Co. (former Petroquı́mica The methodology used in this work was divided in three steps:
União – PQU), located in the Petrochemical Pole of Capuava, in simulation and validation of the industrial process of ethylene and
Santo André (Brazil). propylene production, using industrial data from Quattor Petro-
The exergetic analysis is a thermodynamic tool that can be used chemical Co.; simulation and validation of the refrigeration cycles
to evaluate the performance of refrigeration cycles determining the used to produce cold utilities for the process and finally, exergetic
magnitude and location of process irreversibilities (losses of energy analysis of the process. Each one of these steps is detailed as
quality), making possible to study the changes of operational follows.

3. Process simulation and validation

* Corresponding author. Tel.: þ55 19 35213950; fax: þ55 19 35213894.
E-mail addresses: francinemf@feq.unicamp.br (F.M. Fábrega), jakelinerossi@ A simulation of the cold site of an ethylene and propylene
gmail.com (J.S. Rossi), dangelo@feq.unicamp.br (J.V.H. d’Angelo). industrial plant was performed in order to reproduce the process

0360-5442/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.energy.2009.11.001
 F.M. Fábrega et al. / Energy 35 (2010) 1224–1231 1225

Nomenclature R universal constant of ideal gases ¼ 8.314472 (L kPa)/

(K mol)
m_ mass flow (kg/h) n molar volume (L/mol)
Q_ rate of heat transfer (kJ/h) u acentric factor
W _ rate of work (kJ/h)
3 exergetic efficiency Subscripts
Ex rate of exergy (kJ/h) o reference state
EK kinetic energy (kJ/h) kin kinetic
EP potential energy (kJ/h) comp compressor
h specific enthalpy (kJ/kg) cv control volume
P pressure (kPa) D destroyed
s specific entropy (kJ/kg.K) in intlet stream
T temperature (K) phy physical
V volume (m3) j matter stream
g acceleration of gravity (m/s2) mix mixer
v velocity (m/s) pot potential
h height (m) che chemical
TC critical temperature (K) out outlet stream
PC critical pressure (kPa) val valve
TR reduced temperature (¼T/TC) (K) E evaporator

used at Quattor’s industrial site composed by six distillation of coexisting phases in vapor–liquid equilibrium, trying to ach-
columns (Fig. 1 shows the flowchart of the ethylene and propylene ieve better results than the ones obtained with the existing
production plant used in the simulation). The company has equations.
provided industrial data which were used to validate the simula- In the development of this equation, Peng and Robinson were
tion. These data as well as the temperatures obtained from the seeking for the following basic goals: the parameters should be
simulator can be seen in Table 1 for comparison of simulation expressible in terms of the critical properties (PC and TC) and
performance. A more detailed description of the production acentric factor (u); the model should provide a good accuracy
process of ethylene and propylene may be found in the book of near the critical point, particularly for calculations of the
Chauvel and Lefebvre [15]. compressibility factor and liquid density; mixing rules should not
The industrial process described in Fig. 1 uses cracked employ more than a single binary interaction parameter, which
naphtha as raw material. The process starts with a demethanizer should be independent of temperature, pressure and composi-
column that separates hydrogen and methane in the top stream tion and finally, the equation should be applicable to all calcu-
and this is the only column that has four feed streams, all the lations of all fluid properties in natural gas processes. This
others in the process have only one feed stream. The bottom equation of state is generally superior in predicting the liquid
stream is sent to the deethanizer column that removes C2 densities of many materials, especially nonpolar gases, when
compounds in the top of the column, which are sent to the compared to the Soave-Redlich-Kwong equation. In the case of
ethylene/ethane splitter and the bottom products are sent to the pure hydrocarbons, this equation presents an excellent accuracy
depropanizer column which separates C3 compounds in the top for predicting thermodynamic properties, according to many
of the column, which follows to the propylene/propane splitter. other works [18–20].
Bottom products of the depropanizer column are sent to the Peng–Robinson equation of state is given by Equation (1):
debutanizer column where C4 and C5 compounds are separated.
Finally, in the splitter columns, the most important products of RT a
P ¼  (1)
the plant, ethylene and propylene, are obtained with a high nb n2 þ 2bn  b2
purity degree. The cold utilities used in the condensers of the
where a and b are the parameters of the equation of state, given by
demethanizer, deethanizer and ethylene/ethane splitter columns
Equations (2) and (3).
are produced in the refrigeration system of the process (streams
7, 12 and 41 in Fig. 1). In the other columns (depropanizer,
debutanizer and propylene/propane splitter) the cold utility used 0:45724R2 TC2 h i
1=2 2
a ¼ 1 þ f ðuÞð1  TR Þ (2)
in the condensers is cold pressurized water (streams 17, 20 and PC
39 in Fig. 1).
Hydrogenation reactions are used in the process in order to 0:0778RTC
b ¼ (3)
increase ethylene and propylene production. Equilibrium reactors PC
were used in this simulation to perform these reactions. The
composition of cracked naphtha was taken from [16] and it is
f ðuÞ ¼ 0:37464 þ 1:54226u  0:26992u2 (4)
shown in Table 2.
The commercial simulator HYSYSÓ version 3.2 from Aspen It is important to emphasize that the simulation of the ethylene
Technology was used to generate simulated data. The fluid and propylene process was necessary to obtain the thermal duties
package chosen in the simulator for the determination of ther- in the condensers of the columns, enabling the refrigeration cycle
modynamic properties was the Peng–Robinson equation of state simulation. The operational conditions obtained by the simulator
[17]. In their paper they have tried to present a cubic equation of were compared with industrial data and their relative error,
state that was able to predict with great accuracy P-V-T properties calculated by Equation (5), were smaller than 6%.
1226 F.M. Fábrega et al. / Energy 35 (2010) 1224–1231

Methane + Hydrogen Refrigeration Cycle

7
41

15
1 20
Demethanizer 12 Ethylene
2 25
3 33
15
4 Ethylene Ethane
65 Splitter

Reactor
71
39
27
Propylene
Deethanizer
Ethane
Propylene
60 Propane Splitter
60
17

Reactor
160
25

Depropanizer
Propane

60
20

C4

20

Debutanizer

45

C5

Fig. 1. Flowchart of the ethylene and propylene production plant at Quattor used in the simulation.

  Thermodynamic properties of all streams of the process,

industrial-simulated
relative error ð%Þ ¼  $100 (5) necessary for the calculation of the destroyed exergy in the
industrial 
refrigeration cycle, were obtained via simulation using HYSYSÓ. It is
important to say that the heat exchanger ‘‘CW’’ uses pressurized
water at 5 bar, as cold utility, with an inlet temperature of 29.50  C
4. Simulation of refrigeration cycles (302.65 K) and outlet temperature 49.00  C (322.15 K).
Tables 3 and 4 show the operational conditions provided by
The second step of the methodology was the simulation of the Quattor Petrochemical. These data were used to input data into the
refrigeration system based on the one from Quattor where ethylene simulator in order to perform a simulation as close as possible from
and propylene are used as refrigerants. All operational conditions the real process. The process streams indicated in these tables may
were provided by the company and are presented in Tables 3 and 4. be seen in Figs. 2 and 3 respectively.
It is relevant to point out that this refrigeration cycle is an Table 5 shows the relative error of refrigerant mass flow and
integrated cycle, very complex, containing valves, compressors, compressor power between industrial data (provided by Quattor)
separators, heat exchangers and several mixers, as shown in Figs. 2 and the values obtained from simulations. In this comparison the
and 3. The cycle presented in Fig. 2 uses ethylene as refrigerant and variables of the ethylene cycle were used as input data, analyzing as
supplies the cold utilities for the condenser of the demethanizer output data the ones from the propylene cycle. For both variables
column. The one in Fig. 3 uses propylene to produce a stream of the relative error obtained by Equation (1) was smaller than 10%,
cold utility for the condensers of the deethanizer and ethylene- making possible to conclude that the simulation has shown good
ethane splitter columns. agreement.

Table 1
(1) (2)
Industrial data from Quattor and simulated data from Hysys for the process.
(1)
Distillation column Number of trays Feed tray(1) Temperature in top tray (K) Pressure in top
tray (kPa)(1)
Quattor(1) Simulation(2) Relative error (%)
Demethanizer 65 15; 20; 25 and 33 176.15 175.15 0.57 3500
Deethanizer 60 27 262.15 262.85 0.27 2650
Ethylene and Ethane Splitter 71 15 243.15 242.25 0.37 1910
Depropanizer 60 25 288.15 284.95 1.11 780
Propylene and Propane Splitter 120 60 313.15 295.35 5.68 1090
Debutanizer 45 20 315.15 304.45 3.40 340
 F.M. Fábrega et al. / Energy 35 (2010) 1224–1231 1227

Table 2 Table 4
Cracked naphtha composition at 288.15 K and 4000 kPa pressure [16]. Operational conditions given by Quattor for the heat exchangers of the propylene
cycle (Fig. 3).
Components Composition (molar fraction)
Hydrogen 0.3410 Stream Equipment Exit Temperature (K)
Methane 0.0421 2 CW 313.15
Ethylene 0.3350 8 Heat Exchanger A 291.15
Ethane 0.2628 15 Heat Exchanger B 277.15
Acetylene 0.0040 23 Heat Exchanger C 249.15
Propylene 0.0070 31 Heat Exchanger D 233.15
Propane 0.0010 3 Separator, þ18  C 291.15
Propadiene 0.0010 7 Separator, þ4  C 277.15
1-Butene 0.0005 14 Separator, 24  C 249.15
13-Butadiene 0.0040 22 Separator, 40  C 233.15
n-Butane 0.0005
n-Pentane 0.0010

Mafi et al. [8] also performed an exergetic analysis of a similar Ex ¼ Exphy þ Exche þ Expot þ Exkin (6)
refrigeration system for the same process (ethylene and propylene
When evaluated relative to the environment, the kinetic and
production) but they did not mention if their case study is a real potential energies of a system are in principle fully convertible to
process or not. They have used in their propylene refrigeration
work as the system is brought to rest relative to the environment,
system three temperature levels (5  C, 20  C and 35  C) and in and so they correspond to the kinetic and potential exergies,
the ethylene refrigeration system they used two temperature levels respectively, as shown in Equations (7) and (8).
(65  C and 101  C). In this work a real industrial process (from
Quattor Petrochemical Co.) was studied and simulations were _ 2
mv
validated comparing the results obtained with the ones from the Exkin ¼ EK ¼ (7)
2
industry and some different temperature levels were used: four (18
 C, 4  C, 24  C and 40  C) for the propylene system and three
_
Expot ¼ EP ¼ mgh (8)
(55  C, 75  C and 101  C) for the ethylene system; which
makes the refrigeration systems of this work a little bit more Some assumptions were adopted to perform exergetic analysis,
complex. which are: all systems are operating at steady state; variation of
potential and kinetic energies in all equipments are neglected;
5. Exergetic analysis there are not chemical reactions or changes in the refrigerant
composition and all equipments operate adiabatically. With these
The third step of the methodology used in this work is the assumptions potential, kinetic and chemical exergies are null,
exergetic analysis. In thermodynamics, the exergy of a system is remaining only the physical exergy.
defined as the maximum work possible to obtain from a system In order to prove that kinetic and potential exergy components
during a process that brings this system into equilibrium with its may be neglected in this work, two control volumes in the ethylene
surroundings (which is at a reference state with a temperature To cycle – the compressor (considering the three stages) and one heat
and a pressure Po). After the system and its surroundings reach exchanger (CW) – see Fig. 2, were analyzed using the First Law of
equilibrium, the exergy of the system will be zero [11]. In an exergy Thermodynamics, to evaluate the variation of kinetic and potential
balance, the so called ‘‘destroyed’’ exergy represents the real loss in energy (since they correspond to kinetic and potential exergy
the quality of energy that cannot be identified by means of an components). In this analysis, a basis of calculation that considered
energy balance, since a conservation of energy will always be these components of energy as being only 1% of the entire work
considered [21]. Exergetic analysis has been used by many authors consumed in the compressor and 1% of the enthalpy variation of
to perform the evaluation of the efficiency of industrial process one of the streams in the heat exchanger, was used. With this
[7–14]. To perform the exergetic analysis an electronic spreadsheet procedure it is possible to show that the variation of height and
was used in this work. velocity in these control volumes needed to be very high, more than
The exergy may be split into four components (physical, they are in a real process, in order to be significant. For example it is
chemical, potential and kinetic) as shown in Equation (6). impossible to build a compressor of 468 m high, so if it is much
smaller than this, certainly the potential energy may be neglected.
The same observation is valid for kinetic energy. The results of this
Table 3 analysis are presented in Table 6. This is a common procedure
Operational conditions provided by Quattor for the heat exchangers of the ethylene
adopted in many other works that used exergetic analysis [7–14].
cycle (Fig. 2).
So Equation (6) may be reduced to Equation (9) as follows.
Process Stream Equipment Exit Temperature (K)
1 CW 471.35 Exphy ¼ ðh  ho Þ  To ðs  so Þ (9)
2 CW 313.15
3 Heat Exchanger A 298.15 where h is the specific enthalpy (kJ/kg), s is the specific entropy (kJ/
4 Heat Exchanger B 281.15 kg.K), both evaluated at T and P of each process stream; h0 and s0
5 Heat Exchanger C 253.15
are, respectively, the specific enthalpy and specific entropy evalu-
6 Heat Exchanger D 236.15
7 Separator, 55  C 218.15
ated at the reference state, which in this paper are To ¼ 298.15 K and
12 Separator, 75  C 198.15 Po ¼ 100 kPa.
18 Separator, 101  C 172.15 Equation 10 presents the complete exergy balance of any system
28 Compressor Stage 1 346.29 [22]. When an accumulation of exergy occurs in the system, the
29 Compressor Stage 2 613.75
derivation term is employed to represent the rate of increasing or
30 Compressor Stage 3 517.95
decreasing exergy in the system.
1228 F.M. Fábrega et al. / Energy 35 (2010) 1224–1231

20

-101° C

18 -75° C

21
16
23 22 -55° C
12
15
Chiller -101°C - C2

14 10
24 17
24
Chiller -75°C - C2 9
Cd DeC1
7
8
11
19 Chiller -55°C - C2

13
25

26

28
6
27

Stage 1

Stage 2
29
Stage 3
30
1 2 3 4 5

CW A B C D

Fig. 2. Refrigeration system with ethylene as refrigerant.

! !
! m
X m
X
n
X   _ D ¼ W
_ cv þ _ i Exi _ i Exi
dExcv To _ Ex m  m (11)
¼ 1 Q  W _ cv  Po dVcv
dt Tj j dt i¼1 in i¼1 out
j¼1
! ! The term of the exergy transfer associated to work transfer
m
X Xm
ðW_ cv Þ is used only in the exergetic balance of compressors. So, for
þ _ i Exi
m  m_ i Exi _ D
 Ex (10)
i¼1 i¼1
the other components of the cycle, except compressors, the
in out
quantity of destroyed (dissipated) exergy may be calculated by
where: dExcv =dt: represents the time rate of change of the exergy of Equation (12).
Pn _
the control system; j ¼ 1 ð1  To =Tj Q j Þ: this sum represents the ! !
exergy transfer associate with heat transfer considering that Q_ j m
X m
X
_ D ¼
Ex _ i Exi
m  _ i Exi
m (12)
represents the time rate of heat transfer at the location on the
boundary where the instantaneous temperature is Tj; i¼1 in i¼1 out
W _ cv  Po dVcv =dtÞ: where W
_ cv ðW _ cv represents the time rate of
Through the exergetic analysis it was possible to identify and
energy transfer by work, other than flow work. The accompanying quantify the losses of exergy in the entire system and then it is
availability transfer is given by ðW _ cv  Po dVcv =dtÞ, where dVcv/dt is
P possible to propose some changes in the operational variables of
the time rate of change of volume. ð m _
i ¼ 1 mi Exi Þin : this term the refrigeration cycle in order to reduce these losses. The current
accounts for the time rate of exergy transfer accompanying mass operational conditions and the new proposed ones were evaluated
P
flow and flow work at the inlet of the system. ð m _
i ¼ 1 mi Exi Þout : this based on the exergetic efficiency theory, which is presented in
term accounts for the time rate of exergy transfer accompanying more details by [5,6,11,23].
mass flow and flow work at the outlet of the system. Ex _ D : finally,
this term accounts for the time rate of exergy destruction due to
irreversibilties within the control volume. 6. Results and discussion
In this work the process is considered to be operating at
steady state conditions, so the terms related to variation with In the exergetic analysis of the system, after the identification
time are null [5,6,22]. Also all components of the cycle are and quantification of the exergy losses, some operational variables
considered to be adiabatic, so there are no losses of energy due to of the refrigeration cycle were changed, trying to reduce these
heat transfer between the system and its surroundings. With losses.
these assumptions Equation 10 may be rewrite leading to Equa- Fig. 4 shows the rate of destroyed exergy for both ethylene and
tion 11 as follows. propylene cycles, considering the simulated process and actual
 F.M. Fábrega et al. / Energy 35 (2010) 1224–1231 1229

-40ºC

31
22
28
Heat Exchanger D
27 18
32 -24ºC
29
Chiller -40ºC - C3
30 23
33 14
19
Heat Exchanger C +4ºC
Frac Etil 26
24 11
20 7
17
+18ºC
Chiller -24ºC - C3 15 10

25 12 5
21 9
Heat Exchanger B
Cd DeC2 13 4
16 3

6
Chiller -4ºC - C3

8 Heat Exchanger A
34

35

36

37
Stage 1

Stage 2

38 Stage 3
39
40
41
Stage 4
1

CW

Fig. 3. Refrigeration system with propylene as refrigerant.

operational conditions. The high destruction of exergy observed in they are responsible for the greatest losses. To reach this it is
the mixers of the ethylene cycle is due the fact that an uncontrolled necessary to reduce the great difference of temperature and pres-
mixing of the streams is done, without considering their potential sure between the streams that feed the mixer, since the existing
to produce work, since they have different temperatures and potential between them is lost without producing work, which
pressures, resulting in a great lost of exergy. It is important to means that a lot of exergy is lost. Then the temperatures in the
comment that in the ethylene cycle it was possible to make some outlet streams of the compressors (streams 28, 29 and 30 – Fig. 2)
changes in the operational variables in order to perform a sensi- were changed, causing, consequently a change in their pressures. It
tivity analysis to evaluate their influence on the exergy losses. In the is not possible to change pressure in stream 30 because this is
propylene cycle, the sensitivity of these variables is too high, so a constraint of the process. Some tests changing these variables
they cannot be changed randomly, showing that this should be were performed and Table 7 presents the best set of operational
done in a more systematic way, such as by an optimization method. conditions (temperatures and pressures), from the ones tested that
That is why this work has focused in the analysis of the ethylene led to a minimum exergy loss.
cycle.
When trying to propose new operational conditions the major
objective was the reduction of destroyed exergy in the mixers, since
Table 6
Analysis of the influence of kinetic and potential energy in control volumes of
ethylene refrigeration cycle.

Table 5 Compressor Heat Exchanger

Relative error of industrial and simulated data. (CW)

Ethylene Cycle Propylene Cycle Work (kJ/h) 3.6  107 –

DH12 ethylene stream (kJ/h) – 25.6  106
Refrigerant mass Industrial (Quattor) 78 500 545 000 DEC (kJ/h) 3.6  105 25.6  104
flow (kg/h) Simulated (Hysys) 78 500a 600 000 Dv2 – necessary variation of square velocity 9172 3242
Relative error (%) – 9.17 between inlet and outlet streams (m2/s2)
Compressor Industrial (Quattor) 10 35 DEP (kJ/h) 3.6  105 25.6  104
power (MW) Simulated (Hysys) 10a 36 Dh – necessary variation of height between 468 165
Relative error (%) – 2.78 inlet and outlet streams (m)
a
input data used in the simulator.
1230 F.M. Fábrega et al. / Energy 35 (2010) 1224–1231

20 20 Actual Conditions
Ethylene Cycle Proposed Conditions

Destroyed Exergy (kJ/h x 10 -6 )

Destroyed Exergy (kJ/h x 10 -6 )

Propylene Cycle
15 15

10 10

5 5

0 0

Fig. 4. Rate of destroyed exergy in the ethylene and propylene cycles. Fig. 5. Detroyed exergy in the ethylene cycle: current and proposed operational
conditions.

After proposing new operational conditions, another exergetic ! !

m
P m
P
analysis was made and the best set of variables shown in Table 7 _ i Exi
m  _ i Exi
m
caused a decrease of 13% in the exergetic losses in the entire system. i¼1 in i¼1 out
3comp ¼ _
(14)
The exergetic losses for the compressor, heat exchanger and 4 W i
mixers were also reduced for the ethylene cycle in 12.0%, 18.9% and
19.6%, respectively, as shown in Fig. 5. !
m
P
In this work the concept of exergetic efficiency [24] was _ i Exi
m
defined as the ratio between the sum of the exergy of output i¼1
3val ¼ !out (15)
streams and the sum of the exergy of input streams (material, m
P
heat and/or work) of a particular control volume according to _ i Exi
m
i¼1 in
Equation (13).
!
n
P _ out Exout
m
_
Ex 3mis ¼ ! (16)
i m
P
i¼1 m_ i Exi
3 ¼ !out (13)
n
P i¼1 in
_
Ex i
i¼1 in  
_ out Exout  m
m _ in Exin cold stream
The calculation of the exergetic efficiency allows: the qualifi- 3E ¼ (17)
ðm_ in Exin  m
_ out Exout Þhot stream
cation of energy, making possible to select the most efficient energy
source in a particular case; the identification of specific sites where It was also observed that the destruction of exergy in the heat
exergetic losses (irreversibility) occur and, consequently, the exchanger of the ethylene cycle was reduced to approximately zero,
conditions for process improvement; and finally an optimization of but an acceptable temperature difference between the streams was
the processes. kept. It is important to emphasize that the change in operational
Table 8 shows the exergetic efficiency for the equipments of the conditions involving evaporators and expansion valves requires
ethylene cycle that have presented a significant variation in their a more specific study such as the use of an optimization method, for
efficiency, emphasizing that the compressor efficiency was example, the application of the thermoeconomic optimization
calculated considering the three stages of compression as just one methodology by [25], because it will be necessary to implement
compressor. The exergetic efficiency was calculated using Equa- changes in the operational conditions following a systematic way
tions 14, 15, 16 and 17 respectively for compressors, valves, mixers, and not just a trial and error procedure.
heat exchangers and evaporators. The coefficient of performance, COP, is a fundamental parameter
in the analysis of refrigeration systems, helping to indicate
parameters that have great influence over the energetic perfor-
Table 7 mance of the cycle. It is defined by the relation between the sum of
Actual and final proposed conditions in the temperatures of mixers in the ethylene
cycle (Fig. 2).
Table 8
Stream Equipment Actual conditions Final proposed conditions Exergetic efficiency for the ethylene cycle (actual and proposed conditions).

T (K) P (kPa) T (K) P (kPa) Equipment Actual Conditions (%) Proposed Conditions (%)
28 Compressor Stage 1 518.00 34 375 514.20 36 942 Compressor 82.10 81.60
29 Compressor Stage 2 613.75 66 014 553.10 39 984 Mixer (MIX – 100) 56.80 58.00
30 Compressor Stage 3 346.24 1588 383.10 1588 Heat Exchange (E 101) 81.30 84.70
 F.M. Fábrega et al. / Energy 35 (2010) 1224–1231 1231

14.0 some limitations to be considered when implementing this meth-

Compressor Power (MW)

odology, since some modifications of conditional operations may

12.0
be not possible for an industrial process in operation. The exergetic
10.0 analysis is more suitable for design process.
8.0 The calculations of the COP for the refrigeration cycles were
performed before and after the changes in the operational condi-
6.0
tions and an improvement of more than 31% in the COP could be
4.0 observed. It was also analyzed the influence of compressor power
2.0 in the COP of ethylene refrigeration cycle.
0.0
0.0 0.5 1.0 1.5 2.0
COP Acknowledgments

Fig. 6. Influence of compressor power in COP. The authors would like to thank Chem. Eng. Reinaldo Antônio
Cardoso and Quattor Petrochemical Co. (former PQU) for their
support in the development of this work.
the rate of heat removed from the hot fluid in all the evaporators
and the sum of the rate of work done by all the compressors of the
cycle.
In this work, calculations of the COP for the refrigeration cycles References
were performed before and after the changes in the operational
[1] Abiquim – Associação Brasileira da Indústria Quı́mica. Ethylene world
conditions and an increase of more than 31% in the value of the COP production use naphtha as raw material and in Brazil. See also. Campinas, SP:
could be observed. The influence of compressor power in the ABIQUIM, http://www.abiquim.org.br; 2009.
ethylene refrigeration cycle was also studied. As shown in Fig. 6 it [2] Ouadha A, En-Nacer M, Adjlout L, Imine O. Exergy analysis of a two-stage
refrigeration cycle using two natural substitutes of HCFC22. International
can be observed that the decrease in the compressor power causes
Journal of Exergy 2005;2(1):14–30.
great influence in the COP, increasing it. It is important to [3] Araújo AB, Brito RP, Vasconcelos LS. Exergetic analysis of distillation processes –
remember that the application of an optimization technique is a case study. Energy 2007;32(7):1185–93.
needed to evaluate the ideal compressor power. The black square in [4] Hinderink AP, Kerkhof FPJM, Lie ABK, De Swaan J, Arons J, Van Der Kooi HJ.
Exergy analysis with a flowsheeting simulator – I. Theory; calculating exergies
Fig. 6 indicates the value of compressor power (10.0 MW) consid- of material streams. Chemical Engineering Science 1996;51(20):4693–700.
ering actual conditions of the cycle and the black triangle indicates [5] Szargut J, Morris DR, Steward FR. Exergy analysis of a thermal, chemical and
the value of compressor power (6.9 MW) after the implementation metallurgical processes. New York: Hemisphere Publishing Corporation; 1988.
[6] Ahern JE. The exergy method of energy systems analysis. Toronto: John Wiley
of the new operating conditions. & Sons Inc; 1980.
[7] Morosuk T, Tsatsaronis G. A new approach to the exergy analysis of absorption
7. Conclusions refrigeration machines. Energy 2008;33(6):890–907.
[8] Mafi M, Naeynian SM, Amidpour M. Exergy analysis of multistage cascade low
temperature refrigeration systems used in olefin plants. International Journal
In this paper an exergetic analysis of the refrigeration cycles of of Refrigeration 2009;32(2):279–94.
ethylene and propylene production process was performed to [9] Wang Q, Yanzhong L, Xi C. Exergy analysis of liquefied natural gas cold energy
recovering cycles. International Journal of Energy Research 2005;29(1):65–78.
identify and quantify exergetic losses in this system, and then new [10] Yumrutaş R, Kunduz M, Kanoglu M. Exergy analysis of vapor compression
operational conditions were proposed for the refrigeration cycles in refrigeration systems. Exergy, an International Journal 2002;2(4):266–72.
order to reduce these losses. [11] Kotas TJ. The exergy method of thermal plant analysis. Malabar: Butterworths;
1995.
The simulations done for the ethylene and propylene produc-
[12] Kwak H-Y, Kim D-J, Jeon J-S. Exergetic and thermoeconomic analyses of power
tion process involving the distillation columns and the refrigeration plants. Energy 2003;28(4):343–60.
cycles were compared with current industrial data and a good [13] Rivero R, Rendón C, Gallegos S. Exergy and exergoeconomic analysis of a crude
agreement between industrial and simulated data was obtained, oil combined distillation unit. Energy 2004;29(12–15):1909–27.
[14] Yang JL, Ma YT, Guan HQ. Exergy analysis of transcritical carbon dioxide
showing that the methodology adopted is reliable for the purpose refrigeration cycle with an expander. Energy 2005;30(7):1162–75.
of an exergetic analysis of the system. [15] Chauvel A, Lefebvre G. Petrochemical processes – synthesis – gas derivatives
Through the exergetic analysis it was observed that the highest and major hydrocarbons. New York: Gulf Publishing Company; 1989.
[16] Kaiser V, Heck G, Mestrallet J. Optimize demethanizer pressure for maximum
rates of destroyed exergy were concentrated in the mixers (40%), ethylene recovery. Hydrocarbon Processing 1979;58(6):115–21.
compressors (22%) and in the heat exchangers (20%). New opera- [17] Peng D-Y, Robinson DB. A new two-constant equation of state. Industrial and
tional conditions for the ethylene cycle were proposed and with Engineering Chemistry Fundamentals 1976;15(1):59–64.
[18] Brown JS. Predicting performance of refrigerants using the Peng–Robinson
them it is possible to reduce the total destroyed exergy in equation of state. International Journal of Refrigeration 2007;30:1319–28.
approximately 13% for the entire refrigeration system. These new [19] Stegou-Sagia A, Kakatsios X, Damanakis M. Ozone friendly binary blends R 32/
conditions were proposed considering real industrial possibilities R 134a and the ternary R 407b. Energy 2004;29:2537–52.
[20] Artemenko S, Mazur V. Azeotropy in the natural and synthetic refrigerant
and limitations.
mixtures. International Journal of Refrigeration 2007;30(5):831–9.
It is relevant to say that changes of operational conditions [21] Tsatsaronis G. Thermoeconomic analysis and optimization of energy systems.
require also a specific study such as the use of an optimization Progress in Energy and Combustion Science 1993;19(3):227–57.
[22] Moran MJ, Shapiro HN. Fundamentals of engineering thermodynamics. 6th ed.
method, for example, the application of thermoeconomic optimi-
Toronto: John Wiley & Sons, INC; 2007.
zation methodology, because it will be necessary to implement [23] Mehrpooya M, Jarrahian A, Pishvaie MR. Simulation and exergy-method
changes in the operational conditions following a systematic way analysis of an industrial refrigeration cycle used in NGL recovery units.
and not just a trial and error procedure. International Journal of Energy Research 2006;30(15):1336–51.
[24] Bejan A, Tsatsaronis G, Moran M. Thermal design & optimization. New York:
The methodology applied has shown to be appropriate to Wiley-Interscience; 1996.
improve the performance of the industrial process, but there are [25] Lozano MA, Valero A. Theory of exergetic cost. Energy 1993;18(9):939–60.