Chinese Journal of Chemical Engineering 23 (2015) 1518–1522

Contents lists available at ScienceDirect

Chinese Journal of Chemical Engineering

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

Process Systems Engineering and Process Safety

Simultaneous optimization of heat-integrated crude oil

distillation systems☆
Yiqing Luo ⁎, Liwen Wang, He Wang, Xigang Yuan
Chemical Engineering Research Center, School of Chemical Engineering and Technology, Tianjin University, Tianjin 300072, China

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

Article history: Crude oil distillation is important in refining industry. Operating variables of distillation process have a critical ef-
Received 10 September 2014 fect on product output value and energy consumption. However, the objectives of minimum energy consumption
Received in revised form 21 January 2015 and maximum product output value do not coordinate with each other and do not lead to the maximum eco-
Accepted 6 February 2015
nomic benefit of a refinery. In this paper, a systematic optimization approach is proposed for the maximum an-
Available online 27 March 2015
nual economic benefit of an existing crude oil distillation system, considering product output value and energy
Keywords:
consumption simultaneously. A shortcut model in Aspen Plus is used to describe the crude oil distillation and
Crude oil distillation the pinch analysis is adopted to identify the target of energy recovery. The optimization is a nonlinear program-
Annual economic benefit ming problem and solved by stochastic algorithm of particle warm optimization.
Energy optimization © 2015 The Chemical Industry and Engineering Society of China, and Chemical Industry Press. All rights reserved.
Particle warm optimization
System engineering

1. Introduction
Crude oil distillation plays an important role in refineries, affecting
economic benefits and energy consumption. The distillation unit splits
crude oil into a series of oil products with different values. The yield
distribution of oil products affects product output value and energy con-
sumption. For example, adding the heat taken from the top by changing
the heat taken from the mid-pumparounds of the atmospheric column
will increase the output of light constituents such as naphtha, which
will increase the output value of the column and result in more energy
consumption. Thus the optimization of crude oil unit should coordinate
the increase in product output value and reduction of energy consump-
tion simultaneously, maximizing the economic benefit.
Some advances in crude oil distillation process include improving
heat exchanger network structure [1–3], investigating the effect of
mixing different crude oils on product profit to increase economic profit
[4,5], optimizing operating parameters of distillation unit to maximize
product profit [6–9], and improving production program by solving
nonlinear programming (NLP) problems [10,11]. Liebmann et al. put
forward design procedures of a heat-integrated conventional crude oil
distillation tower using pinch analysis [12]. Suphanit adjusted reflux,
flow rate of steam stripping and other degrees of freedom to reduce op-
erating cost considering the interaction of distillation column and heat
exchange network [13]. Some improvements were made to Suphanit's
☆ Supported by the National Natural Science Foundation of China (21176178) and the
State Key Laboratory of Chemical Engineering (SKL-ChE-13B02).
⁎ Corresponding author.
E-mail address: luoyq@tju.edu.cn (Y. Luo).
approach and applied to retrofit design [14–17]. Most of previous stud-
ies focused on maximizing the product profit and minimizing the ener-
gy consumption separately, which are related to economic benefit of a
refinery. However, considering the two targets separately may not
give the maximum economic benefit. Besides, boiling point tempera-
tures of refining products were fixed at a predetermined value in
these optimizations. In fact, the boiling point temperature of a refining
product varies in a permissible range, within which the quality of the
product is satisfactory. Therefore, product yield distribution of a crude
oil distillation column is different when optimizing the operating pa-
rameters affecting both product output value and energy consumption.
In this study, we optimize operating parameters including distributions
of product yield and flow rates of stripping steams to maximize the an-
nual economic benefit considering the product output value and energy
consumption simultaneously by a systematic approach based on the
SCFrac shortcut model of Aspen Plus and the pinch analysis. The optimi-
zation is a NLP problem and solved by stochastic algorithm of particle
warm optimization (PSO) algorithm [18]. A case study is used to illus-
trate the application.
2. Process Description
Fig. 1 shows a general flow sheet of the initial distillation in a refin-
ery. Crude oil enters the preheat train, which consists of heat ex-
changers associated with various downstream equipment in the
system. The preheat train raises the temperature of crude significantly
and reduces the overall energy consumption. The heated crude enters
the primary furnace, where a large portion of it evaporates. Then the
vapor–liquid mixture goes to the atmospheric distillation tower

http://dx.doi.org/10.1016/j.cjche.2015.03.008
1004-9541/© 2015 The Chemical Industry and Engineering Society of China, and Chemical Industry Press. All rights reserved.
 Y. Luo et al. / Chinese Journal of Chemical Engineering 23 (2015) 1518–1522 1519

Fig. 1. Aspen model of atmospheric and vacuum distillation units.

named CRUDE. The potion with boiling point lower than 350 °C is into N-1 simple towers sequentially numbered from the top to the bot-
distillated from CRUDE as products, while the atmospheric residue RC tom. SCFrac model requires the following conditions: (1) pressure of
is send to the vacuum distillation named VACUUM to recover additional side products, (2) estimated initial amount of side product withdrawn,
distillates and gas oils. and (3) amount of stripping steam of all products in addition to over-
In this study, all operation conditions and values of operating vari- head product [19]. By specifying the column pressure, the estimated
able are based on an existing refinery in China. After preheated to
about 370 °C at 145 kPa in the atmospheric furnace, the crude oil en-
tered the flash zone of atmospheric column at Stage 29, which
contained 31 theoretical stages. The pressure of the column was main-
tained at 131 kPa at the top and at 147 kPa at the bottom. The temper-
ature of the top condenser was kept at 47 °C. A steam stream was fed at
the bottom of atmospheric column. Three pumparounds were located
between Stages 3 and 1, between Stages 12 and 10, and between Stages
21 and 19. The top distillate product was naphtha, and the side products
from 3 steam strippers (S1, S2, S3) were kerosene, diesel and atmo-
spheric gas oil (AGO), which were drawn from Stages 10, 13 and 24 in
sequence.
The residual oil from the bottom of atmospheric column was heated
to 394 °C in the vacuum furnace and entered the flash zone of vacuum
column, which had 14 theoretical stages with pressure maintained at
2.7 kPa at the top and at 3.9 kPa at the bottom. There were 3
pumparounds, PAV1 located between Stages 2 and 1, PAV2 located be- Fig. 2. SCFrac model in Aspen Plus.
tween Stages 7 and 6, and PAV3 located between Stages 9 and 8. The
side products, vacuum diesel (VD), light vacuum gas oil (LVGO), and product flow, the number of theoretical stage, the steam flow rate,
heavy vacuum gas oil (HVGO), were drawn from Stages 2, 7 and 9, re- and the value range of product property (D86 95% point), the SCFrac
spectively. The bottom product was the residue. In the distillation sys- can give relatively accurate results including product status and heat
tem, naphtha, kerosene, diesel, AGO, and vacuum diesel were direct duties of each condenser, furnace, and pumparound.
products, while LVGO, HVGO and the residue went to downstream
units. 3.2. Energy consumption calculation

3. Crude Oil Distillation Model
3.1. Shortcut model
In a design and retrofit problem of crude oil distillation systems,
shortcut models [15] are more suitable for optimization and can be
solved more quickly than strict models, especially when multiple vari-
ables are considered simultaneously. In this paper, SCFrac shortcut
model in Aspen Plus [19] is used to simulate the crude oil distillation.
The SCFrac model is based on the following assumptions: (1) a complex
column can be divided into a series of simple column sections, as shown
in Fig. 2; (2) liquid flows between sections are negligible; and (3) all
steam flows to the top of the column and is withdrawn with overhead
product. SCFrac model decomposes a complex tower with N products
The pinch analysis method has been proposed to calculate energy re-
covery without knowing the exact structure of a heat exchanger net-
work [20], which is adopted here to obtain theoretical hot (required
furnace heat) and cold (required cooling duty) utility consumption.
We take the minimum temperature difference of 15 °C, used in the Chi-
nese refinery. The crude oil feed is the only cold stream, while products
and pumparounds are hot streams.
4. Optimization
4.1. Optimization variables
In the SCFrac model, the variables we optimize include the flow rate
of four stripping steams (atmospheric column bottom stripping steam
FSTC and three side stripping steams FST1, FST2, and FST3) and the D86
1520 Y. Luo et al. / Chinese Journal of Chemical Engineering 23 (2015) 1518–1522

Table 1 4.3. Optimization methodology

Bounds of variable values

Variable Range of D86 95% point/°C Variable Range of flow rate/kg·h−1 The optimization of the crude oil distillation system is a NLP problem
DN 170–185 FST1 0–1000
and solved by a PSO algorithm, which is a population based stochastic
DK 250–260 FST2 0–1000 optimization technique. In PSO, the potential solutions, named particles,
DD 310–320 FST3 0–1000 ‘fly’ through the problem space following the current optimum parti-
DA 355–365 FSTC 4000–9000 cles. As a point in a D-dimensional space, each particle updates accord-
DVD 360–365
ing to the experience of itself and other particles. The algorithm is
simple in concept and gives better results in a faster and cheaper way.
The basic equations of PSO are given below.
95% points of naphtha (DN), kerosene (DK), diesel (DD), AGO (DA) and
i i i
vacuum diesel (DVD). Of each distillate product, the D86 95% point is xkþ1 ¼ xk þ vkþ1 ð4Þ
constrained to an adoptable range, in which the quality and product    
price could be regarded the same, as shown in Table 1. i i i i g i
vkþ1 ¼ ωk vk þ c1 r 1 pk −xk þ c2 r2 pk −xk ð5Þ

4.2. Objective function

In each update step k, for each particle i, position xi is updated by
Eq. (4) and step velocity vik + 1 is calculated by Eq. (5), based on the
How to improve the product output value and how to reduce the en-
best position of particle, pik, and that of group, p gk. Luo et al. [18] have pro-
ergy consumption are currently two common concerns in refineries.
posed a discrete binary particle swarm optimization algorithm, which
They are affected by interrelated operating parameters. Optimizing
can be used to optimization problems containing integer variables and
one of them does not mean maximum economic benefit of a refinery.
achieve satisfactory results. In this study we adopt this algorithm. c1
Considering product output value and energy consumption simulta-
and c2 are the two positive parameters, and in this work, c1 = c2 = 2;
neously will realize the maximum economic benefit. Here the maxi-
ωk = 0.729 is the inert weight; r1 and r2 are the uniform random num-
mum annual economic benefit is the objective function, expressed by
bers between 0 and 1.
Eq. (1). For comparison, we also use the maximum product output
In this study, PSO algorithm is connected with the SCFrac model by
value and the minimum energy consumption as objective functions, for-
the Visual Basic and Aspen Plus Active X automation sever [19]. When
mulated by Eqs. (2) and (3), respectively, with the annual economic
any one of the following conditions is satisfied, the termination condi-
benefit calculated.
tion is satisfied.
(a) The
 maximum number of iteration is reached;
 k
(b)  maxZ −kmaxZ b10−3 ; Δk ¼ 50 (Z is the objective function).
kþΔk

maxZ
maxðAnnual PROFIT Þ ¼ ðPRODUCT−ENERGY Þ  8400 ð1Þ

X Table 3
PRODUCT ¼ ð F i  P i Þ− F C  P C ði ¼ N; K; D; A; V; RÞ ð2Þ ASTM D86 data of two crude oils

ER crude DG crude

ENERGY ¼ ðQ F P F þ Q CW   P CW Þ  3600 Liquid volume/% Temperature/°C Liquid volume/% Temperature/°C

X
þ F ST j  P ST ð j ¼ 1; 2; 3; C Þ ð3Þ 5 150.6 5 160. 8
10 189.9 10 213.3
30 313.4 30 333.7
50 394.4 50 421.8
where PRODUCT is the product output value (USD·h−1), the product 70 464.8 70 508.7
value minus the cost of feedstock, ENERGY is the system energy cost 90 573.6 90 655.7
(USD·h−1), Annual PROFIT is the annual economic benefit of the system 95 720.2 95 763.5
100 866.8 100 871.4
(USD·a−1), Fi is the optimal flow rate of product i (m3·h−1), FC is the
flow rate of crude oil feed (m3·h−1), Pi and PC are the prices of product
i and crude oil (USD·m−3), respectively, Q F and Q CW are required fur-
nace heat and cooling duty determined by pinch analysis (MW), respec-
tively, and FSTj is the flow rate of stripping steam j (kg·h−1). All these 5. Case Study and Results
values are functions of stripping steam flow rate and D86 95% point of
each product. The optimization variables are searched within the ranges The existing crude oil distillation system deals with two crude oils,
in Table 1. Product/utility prices are shown in Table 2, remaining the Erlianhaote crude oil (ER crude) and Dagang crude oil (DG crude).
same when the D86 95% points of products are within the ranges in Both throughputs are 5208000 m3·a−1. The ASTM D86 data in Table 3
Table 1. Working time of the refinery per year is 8400 h. show that DG crude oil is heavier than ER crude oil.

Table 2
Prices of products and utilities [4]

Item Price Unit Item Price Unit

−3
PN (naphtha) 1003.8 USD·m PF (fuel oil) 0.017 USD·MJ−1
PK (kerosene) 861.1 USD·m−3 PCW (cooling water) 4.74 × 10−3 USD·MJ−1
PD (diesel) 810.2 USD·m−3 PST (steam) 0.0055 USD·kg−1
PA (AGO and VD) 767.5 USD·m−3 PDG 629.0 USD·m−3
PV (LVGO and HVGO) 718.2 USD·m−3 PER 647.0 USD·m−3
PR (residue) 449.0 USD·m−3 PC 597.0 USD·m−3
 Y. Luo et al. / Chinese Journal of Chemical Engineering 23 (2015) 1518–1522 1521

Table 4
Results of optimal operating parameters

Parameter ER crude DG crude

Minimum energy Maximum product Maximum annual Minimum energy Maximum product Maximum annual
consumption output value economic benefit consumption output value economic benefit

FN/m3·h−1 76.3 82.3 80.4 56.2 59.2 58.5

FK/m3·h−1 41.0 42.7 40.8 48.3 45.1 42.5
FD/m3·h−1 43.8 39.0 41.8 49.2 47.2 48.4
FAGO/m3·h−1 67.2 64.1 66.4 41.9 37.3 39.7
DutyC1/kW 5196 4905 5013 4195 4352 4212
DutyC2/kW 6748 6157 6492 5573 5390 5499
DutyC3/kW 13527 10425 11595 10066 9633 9826
DutyCON/kW 12945 14057 13876 9985 11024 10231
FVD/m3·h−1 9.3 11.7 10.6 13.5 15.1 14.8
FLVGO/m3·h−1 144.2 143.8 142.1 123.6 122.3 120.5
FHVGO/m3·h−1 68.45 69.07 71.11 80.3 83.9 84.2
FR/m3·h−1 168.6 162.1 164.6 188.3 182.8 184.0
DutyV1/kW 4320 3996 4110 3520 3980 3919
DutyV2/kW 8638 8215 8342 7989 7753 7676
DutyV3/kW 19086 17469 18503 18013 16869 17094
Energy consumption × 10−6/USD·a−1 9.98 14.20 10.65 15.78 24.33 20.33
Product output value × 10−6/USD·a−1 318.43 329.74 328.17 258.26 303.06 302.17
Annual economic benefit × 10−6/USD·a−1 308.45 315.54 317.52 242.48 278.73 281.84

The bold-faced values in this Table 4 represent the minimum energy consumption, the maximum product output value and the maximum annual economic benefit the system can obtain
under the three objective function respectively, but they can not be achieved simultaneously.

Table 4 shows the results of optimal operating parameters. For ER
crude, with the minimum energy consumption, the maximum product
output value, and the maximum annual economic benefit as optimiza-
tion objectives, annual economic benefits are 308.45, 315.54, and
317.52 millionUSD·a−1, respectively. Maximizing the annual economic
benefit by considering product output value and energy consumption
simultaneously brings the refinery more economic benefit. When the
minimum energy consumption is the target, the energy cost decreases
to 9.98 millionUSD·a−1, but the product output value and the annual
economic benefit are also lower. As shown in column 2 of Table 5,
the drawing rates of naphtha DN and kerosene DK are less, while the
drawing rates of diesel DD are higher compared with those in column 4.
This indicates that lighter constituent moves down in the column and
more heat is recovered at the expense of less product output value.
When the maximum product output value is the target, Table 5 shows
that the product output value increases to 329.74 millionUSD·a−1 as a re-
sult of lighter constituent moving up. Consequently, more energy cost,
14.20 millionUSD·a−1, is needed. For DG crude, we have similar results.
Both cases indicate that with the maximum annual economic benefit as
the objective considering product output value and energy consumption
simultaneously bring the refinery the largest economic benefit. More-
over, light crude (ER crude) containing more light constituent leads to
more annual economic benefit as a result of more product output value
and less energy consumption. With heavier crude (DG crude), the prod-
uct output value and the annual economic benefit decreases, and the
energy consumption increases a lot. Thus the refinery should pay more
attention to heat recovery units and the heat exchange network should
have certain design margin for different crude feeds.
The stochastic optimization based on the strategy developed in this
study is used to optimize the distillation operating parameters and
heat recovery of an existing refinery in a single framework. For ER and
DG crude oils, the maximum annual economic benefit is obtained by
62 and 46 iterations, respectively. It is computationally efficient and ef-
fective for determining operating parameters with the maximum annu-
al economic benefit as objective.
6. Conclusions
Crude oil distillation columns interact strongly with associated heat
recovery system, which affects the annual economic benefit of the sys-
tem. In this paper, we took the maximum annual economic benefit as
the objective considering the product output value and energy con-
sumption simultaneously. A shortcut model (SCFrac) in Aspen Plus
was applied to describe the crude oil distillation process, giving relative-
ly accurate boiling point temperature of refining products, flow rate,
and heat duties of each condenser, furnace and pumparounds. The
pinch analysis was adopted to identify energy recovery under different
operation conditions. The optimization framework was solved by the
stochastic PSO algorithm in a fast way. The study shows that the distri-
bution of refining products has a major influence on annual economic

Table 5
Results of optimized variables

Variable ER crude DG crude

Minimum energy Maximum product Maximum annual Minimum energy Maximum product Maximum annual
consumption output value economic benefit consumption output value economic benefit

DN/°C 176.0 184.3 183.7 178.2 184.2 184.0

DK/°C 251.1 258.0 256.3 256.5 258.8 257.6
DD/°C 316.5 313.2 314.1 313.8 310.7 310.2
DA/°C 358.7 364.1 362.4 364.7 362.6 362.1
DVD/°C 364.1 363.7 362.8 363.0 364.5 363.6
FST1/kg·h−1 469.3 472.6 452.6 495.2 530.2 503.6
FST2/kg·h−1 507.2 493.7 465.6 609.4 635.3 594.5
FST3/kg·h−1 678.3 669.5 656.7 529.5 507.7 501.2
FSTC/kg·h−1 5479.2 5211.9 5384.2 5692.3 5796.2 5760.1
1522 Y. Luo et al. / Chinese Journal of Chemical Engineering 23 (2015) 1518–1522

benefit compared with energy recovery. The type of crude oil also has an pik the past best position of particle i at time k
effect on the profit and energy cost. The model and optimization meth- Q CW required cooling duty calculated from heat recovery system,
od developed in this paper could help chemical engineers to improve kW
the operation of existing refineries, make decisions on purchasing QF required furnace heat calculated from heat recovery system,
crude, find optimal process conditions, and improve the performance kW
of an integrated refinery. r1, r2 uniform random numbers between 0 and 1
vik step velocity of particle i at time k
Nomenclature ωk inertia weight
c1, c2 positive constants
DA ASTM D86 95% point of AGO, °C
DD ASTM D86 95% point of diesel, °C References
DK ASTM D86 95% point of kerosene, °C
[1] M. Gadalla, M. Jobson, R. Smith, Increase capacity and decrease energy in existing re-
DN ASTM D86 95% point of naphtha, °C finery distillation columns, Chem. Eng. Prog. (April 2003) 44–50.
DVD ASTM D86 95% point of vacuum diesel, °C [2] P. Promvitak, K. Siemanond, S. Bunluluesriruang, V. Raghareutai, Retrofit design of
DutyCON heat duty of atmospheric column condenser, kW heat exchanger networks of crude oil distillation unit, Chem. Eng. Trans. 18 (2009)
99–104.
DutyC1 heat duty of pumparound PAC1, kW [3] F. Friedler, Process integration, modelling and optimisation for energy saving and
DutyC2 heat duty of pumparound PAC2, kW pollution reduction, Appl. Therm. Eng. 30 (2010) 2270–2280.
DutyC3 heat duty of pumparound PAC3, kW [4] R.K. More, V.K. Bulasara, R. Uppaluri, V.R. Banjara, Optimization of crude distillation
system using aspen plus: effect of binary feed selection on grass-root design, Chem.
DutyV1 heat duty of pumparound PAV1, kW Eng. Res. Des. 88 (2010) 121–134.
DutyV2 heat duty of pumparound PAV2, kW [5] V. Mittal, J. Zhang, X.T. Yang, Q. Xu, Analysis for crude and vacuum distillation sys-
DutyV3 heat duty of pumparound PAV3, kW tem, Chem. Eng. Technol. 34 (11) (2011) 1854–1863.
[6] L.C. Liau, T.C. Yang, M. Tsai, Expert system of a crude oil distillation unit for process
FAGO volume flow rate of AGO, m3·h−1 optimization using neural networks, Expert Syst. Appl. 26 (2) (2004) 247–255.
FC volume flow rate of crude oil, m3·h−1 [7] K. Basak, K.S. Abhilash, S. Ganguly, D.N. Saraf, On-line optimization of a crude distil-
FD volume flow rate of diesel, m3·h−1 lation unit with constraints on product properties, Ind. Eng. Chem. Res. 41 (2002)
FHVGO volume flow rate of HVGO, m3·h−1 1557–1568.
[8] S.V. Inamdar, K.S. Gupta, D.N. Saraf, Multi-objective optimization of an industrial
FK volume flow rate of kerosene, m3·h−1 crude distillation unit suing the elitist non-dominated sorting genetic algorithm,
FLVGO volume flow rate of LVGO, m3·h−1 Chem. Eng. Res. Des. 82 (A5) (2004) 611–623.
FN volume flow rate of naphtha, m3·h−1 [9] N. Zhang, R. Smith, I. Bulatov, J. Klemeš, Sustaining high energy efficiency in existing
processes with advanced process integration technology, Appl. Energy 101 (2013)
FSTC steam mass flow rate of main atmospheric column, kg·h−1 26–32.
FST1 steam mass flow rate of stripper S1, m3·h−1 [10] C.W. Cao, X.S. Gu, X. Zhong, Chance constrained programming models for refinery
FST2 steam mass flow rate of stripper S2, kg·h−1 short-term crude oil scheduling problem, Appl. Math. Model. 33 (2009) 1696–1707.
[11] A.M. Alattas, I.E. Grossmann, I. Palourivera, Integration of nonlinear crude distillation
FST3 steam mass flow rate of stripper S3, kg·h−1 unit models in refinery planning optimization, Ind. Eng. Chem. Res. 50 (11) (2011)
FVD volume flow rate of vacuum diesel, m3·h−1 6860–6870.
PA price of AGO and vacuum diesel, USD·m−3 [12] K. Liebmann, V.R. Dhole, M. Jobson, Integrated design of a conventional crude oil dis-
tillation tower using pinch analysis, Trans. IChemE Part A 76 (A) (1998) 335–347.
PC price of feed crude oil, USD·m−3 [13] B. Suphanit, Design of Complex Distillation System(Ph.D. Thesis) UMIST, Manches-
PCW price of cooling water, USD·MJ−1 ter, UK, 1999.
PD price of diesel, USD·m−3 [14] M. Gadalla, M. Jobson, R. Smith, Optimization of existing heat-integrated refinery
distillation systems, Trans. IChemE PArt A 81 (2003) 147–152.
PDG price of Dagang crude, USD·m−3 [15] M. Gadalla, M. Jobson, R. Smith, Shortcut models for retrofit design of distillation col-
PER price of Erlianhaote crude, USD·m−3 umns, Trans. IChemE Part A 81 (2003) 971–986.
PF price of fuel oil, USD·MJ−1 [16] V. Rastogi, Heat-integrated Crude Oil Distillation System Design(Ph.D. Thesis)
PK price of kerosene, USD·m−3 UMIST, Manchester, UK, 2006.
[17] L. Chen, Heat-integrated Crude Oil Distillation System Design(Ph.D. Thesis) UMIST,
PN price of naphtha, USD·m−3 Manchester, UK, 2008.
PR price of residue, USD·m−3 [18] Y.Q. Luo, X.G. Yuan, Y.J. Liu, An improved PSO algorithm for solving non-convex NLP/
PST price of stripping steam, USD·kg−1 MINLP problems with equality constraints, Comput. Chem. Eng. 31 (2007) 153–162.
[19] Aspen Technology, Aspen Plus User Guide, Volumes 1, 2 and 3, 2009.
PV price of LVGO and HVGO, USD·m−3 [20] B. Linnhoff, E. Hindmarsh, The pinch design method of heat exchanger networks,
p gk the past best position of group in the swarm at time k Chem. Eng. Sci. 38 (1983) 745–763.