Anda di halaman 1dari 24

Review

Current Status and Perspectives of Liqueed Natural Gas (LNG) Plant


Design
Wonsub Lim, Kwangho Choi, and Il Moon*,

Department of Chemical and Biomolecular Engineering, Yonsei University, 50 Yonsei-ro, Seodaemun-gu, Seoul 120-749,
Republic of Korea

GS E&C, GS Yeokjeon Tower, 537 Namdaemun-ro 5-ga, Joong-gu, Seoul 120-722, Republic of Korea
ABSTRACT: Liqueed natural gas (LNG) is attracting great interest as a clean energy alternative to other fossil fuels, mainly
due to its ease of transport and low carbon dioxide emissions, a primary factor in air pollution and global warming. It is expected
that this trend in the use of LNG will lead to steady increases in demand over the next few decades. To meet the growing
demand for LNG, natural gas liquefaction plants have been constructed across the globe. Furthermore, single train capacity has
been increased to strengthen price competitiveness. To achieve greater capacity, more complex refrigeration cycle designs that
combine two or more dierent conventional single refrigeration cycles are being developed to obtain synergistic eects in the
liquefaction process. At the same time, a variety of recent studies have focused on designing suitable processes for oshore and
small-scale plants to improve the protability of stranded gas elds. LNG plants are known to be energy/cost-intensive, as they
require a large amount of power for the processes of compression and refrigeration, and need special equipment such as
cryogenic heat exchangers, compressors, and drivers. Therefore, one of the primary challenges in the LNG industry is to improve
the eciency of the current natural gas liquefaction processes in combination with cost savings. In this paper, we review recent
developments in LNG processes, with an emphasis on commercially available refrigeration cycles. We also discuss recent research
and suggest future directions for natural gas liquefaction processes. Up to this point, most studies have focused on operating cost.
To achieve better results, future studies that investigate optimal design and operation of LNG technologies should consider both
capital cost and operating cost.

respectively.4 It is clear from this table that natural gas is an


attractive option for reducing CO2 emissions.
The geographical mismatch between gas reservoirs and
customer demand is reected in the continuous increase in the
global natural gas trade. One of the dominating factors in the
natural gas trade is the selection of an economically feasible
transport method. Widely used methods to transfer natural gas
trade are pipelines and liqueed natural gas (LNG). For transport
over long distances over 3500 km, LNG is the preferred method
for economical, technical, safety-related, and political reasons.
A comparison of future projections of the global inter-regional
natural gas trade through pipelines and LNG is presented in
Figure 1.3,5 Pipeline and LNG trade is projected to increase
continually in the future. However, the LNG trade is projected
to have a higher growth rate than the pipeline trade. Therefore,
LNG plays a major role in the global natural gas trade.
To satisfy the projected demand, the natural gas supply will
need to increase by almost 50% in the period designated in
Figure 1. Several countries have commercialized natural gas
resources and constructed LNG production facilities to meet
this increasing demand. The global trade in natural gas is
currently under rapid transition because several countries have
been steadily increasing their LNG production capacity. In the

INTRODUCTION

Global energy consumption is projected to increase at an


average rate of 0.91.6% per year.13 Fossil fuels are expected
to remain the dominant energy sources into the foreseeable
future. Natural gas is the fastest growing major energy source in
the world; its consumption is expected to increase at an average
rate of 1.41.6% per year from 2008 to 2035,2,3 due to its lower
environmental impact than other fossil fuels. Combustion of
natural gas results in lower emission of CO2 and other pollutants
than combustion of coal and oil. Table 1 shows the amounts of
CO2 emitted from the combustion of natural gas, coal, and oil,
Table 1. Pounds of Air Pollutants Produced Per Billion BTU
of Energya4
pollutant

natural gasb

oilc

coald

carbon dioxide
carbon monoxide
nitrogen oxides
sulfur dioxide
particulates
formaldehyde
mercury

117,000
40
92
0.6
7.0
0.750
0.000

164,000
33
448
1122
84
0.220
0.007

208,000
208
457
2591
2744
0.221
0.016

No post-combustion removal of pollutants. bNatural gas burned in


uncontrolled residential gas burners. cNo. 6 fuel oil burned in an oilred utility boiler. (Conversion factor: No. 6 fuel oil at 6.287 million
BTU per barrel and 1.03% sulfur content.). dBituminous coal burned
in a spreader stoker. (Conversion factor: Bituminous coal at 12 027
BTU per pound and 1.64% sulfur content.).
2012 American Chemical Society

Special Issue: Process Engineering of Energy Systems


Received:
Revised:
Accepted:
Published:
3065

April 1, 2012
December 14, 2012
December 14, 2012
December 15, 2012
dx.doi.org/10.1021/ie302877g | Ind. Eng. Chem. Res. 2013, 52, 30653088

Industrial & Engineering Chemistry Research

Review

Figure 1. Future projections of the global inter-regional natural gas


trade.3,5.

Figure 3. Cost breakdown of the LNG value chain.7.

past 5 years, there has been a 40% increase in the global LNG
production capacity. The liquefaction capacity is projected
to more than double from 8 trillion cubic feet in 2008 to 19
trillion cubic feet in 2035.2
A LNG project is considered one of the most expensive
energy projects and comprises a set of unique values in the
supply chain (known as the LNG value chain), as schematized
in Figure 2.6 The percentage of average capital costs for each
component in the LNG value chain is presented in Figure 3;7
there are noticeable dierences between the costs of each
element in conventional natural gas elds. The cost of a
liquefaction plant is the greatest in the value chain, accounting
for more than 40% of the total cost. The costs of the other
components account for nearly equal portions of the remaining
total. Among the components, the costs of exploration and
production vary greatly according to natural gas eld. Accurate
cost projection of a LNG project depends on a suite of factors
including site conditions, safety, and traded volumes. From an
economic perspective, the liquefaction plant is the most
important component in the LNG value chain.

Algeria, the single mixed refrigerant (SMR) process was


adopted to simplify the complex equipment conguration of
the cascade liquefaction process using three pure refrigerant
cycles. Due to the low thermodynamic eciency of the SMR
process, it was substituted with the propane precooled mixed
refrigerant (C3MR) process comprising a propane precooling
cycle and a mixed refrigerant cycle. This process was rst used
in the Lumut (Brunei) LNG plant in 1972.9 Since then, the
C3MR process designed by Air Products and Chemicals, Inc.
(APCI) has remained the dominant liquefaction process in the
LNG plant market; over 60% of currently installed base load
LNG plants use this process. Only three processesnamely
Cascade, SMR, and C3MRwere applied in base load LNG
plants from the 1970s to the 1990s.
Steady progress has been made in liquefaction technology,
with improvements in thermodynamic eciency, a reduction in
capital cost, and an expanded capacity per train. These continuous
improvements have promoted sti competition among liquefaction technology providers and created new challenges for
developing liquefaction technology.
The main purpose of the majority of early LNG projects was
capacity expansion with the addition of production facilities
to existing operating gas elds. However, to meet the rapid
growth in demand for LNG, a number of new gas elds for
use as LNG plant construction sites have been explored, and
vendors are competing to develop new high-performance,
low-cost technologies and processes. Due to innovations in

1. LIQUEFIED NATURAL GAS (LNG) PLANTS


1.1. Developments in LNG Plants. The worlds rst
commercial liquefaction plant was built in 1964 at Arzew,
Algeria and was based on the cascade liquefaction process,8
which was then applied in the Kenai, Alaska LNG plant in
1969. In the early 1970s at Marsa El Brega, Libya and Skikda,

Figure 2. LNG value chain.6.


3066

dx.doi.org/10.1021/ie302877g | Ind. Eng. Chem. Res. 2013, 52, 30653088

Industrial & Engineering Chemistry Research

Review

Figure 4. Conventional processes of a base load LNG plant.

its volume by a factor of more than 600.5 The liquefaction


process requires a signicant amount of energy due to the large
and complicated refrigeration systems and the high capital cost of
the equipment required such as compressors, heat exchangers,
and other cryogenic equipment. Liquefaction is thus known as an
energy intensive or capital intensive process.
After liquefaction of natural gas, it can be stored in specially
designed containers. Storage tanks and associated equipment
should meet rigorous design standards due to the characteristics
of LNG.

liquefaction technology and related equipment such as heat


exchangers, gas turbines, compressors, and other utilities,
LNG production costs per tonne have decreased substantially.
For instance, liquefaction costs were around $560/tonne in
1995, but decreased to $222/tonne in 2004.10 However, recent
liquefaction costs have increased to some extent due to an
increase in engineering, procurement, and construction (EPC)
costs. The liquefaction costs commissioned in 20052008 were
around $430/tonne. The construction costs in 20092013 are
expected to reach $830/tonne.11
The size of a single liquefaction train has been increasing to
strengthen price competitiveness, with the goal of using fewer
trains to achieve the same capacity. Furthermore, the protability
of stranded gas elds has increased greatly. Development of
liquefaction process eciency has focused on the replacement of
expansion valves with liquid expanders or two-phase expanders,
power reduction in the liquefaction process, or an increase in
LNG production throughput at the same power. In the early
days of the industry, the capacity of the majority of trains was
approximately 12 million tonnes per annum (MTPA). Today,
the largest train has a capacity of 7.8 MTPA.10
1.2. LNG Plant Processes. In general, a LNG plant mainly
consists of pretreatment, liquefaction, and storage facilities.
Figure 4 shows the processes performed in a conventional base
load LNG plant. Natural gas is a mixture of hydrocarbons
comprising methane, ethane, propane, and butane, among others.
It also contains water, carbon dioxide, oxygen, nitrogen, and trace
components of other gases. Before natural gas liquefaction,
impurities such as water, carbon dioxide, oxygen, nitrogen, and
hydrogen sulde must be removed to prevent equipment damage
caused by internal corrosion or solids formed during cooling.
The natural gas pretreatment process consists of three main
steps. The removal of acid gases, such as CO2 and H2S, before
liquefaction is an important step in LNG production, because
if the CO2 concentration exceeds a certain limit the gas will
freeze in the pipelines. The next step is a dehydration process to
remove water to avoid freezing. The last step is to remove
mercury from the gases to protect the aluminum heat exchangers
from corrosion.
Following pretreatment, the natural gas undergoes a liquefaction process in which it is cooled to 161 C at atmospheric
pressure so that a phase transition takes place, allowing the
natural gas to become liquid.12 Liquefying natural gas reduces

2. NATURAL GAS LIQUEFACTION PROCESS


2.1. Fundamental Principles. Liquefaction of natural gas
requires the removal of sensible and latent heat over a wide
range of temperatures using one or more refrigerants,13 and
thus a complicated refrigeration systemeither compression
refrigeration or absorption refrigerationis required. In LNG
plants, a compression refrigeration cycle is generally used.
The ideal compression refrigeration cycle is represented in
Figure 5a.14 The cycle works as follows. In the rst step (12),
a refrigerant absorbs heat at constant pressure and temperature
in an evaporator; in the second step (23), pressure is increased
at constant entropy in the compressor. In the third step (34),
desuperheating and condensation, in which enthalpy decreases
at a constant pressure, are performed in a condenser. In the
last step (34), the refrigerant returns to its original pressure at
constant entropy in an expander.
Improvements in refrigeration cycle eciency can be achieved
by modifying the cycle conguration. Vaporization of the liquid
in the evaporator provides refrigeration. Thus, compression
work can be reduced by adding an intermediate level. This
intermediate level is added between two temperature levels, with
the vapor formed after the rst pressure reduction separated
from the liquid and fed directly into the high-pressure
compressor, as shown in Figure 6a and b.14 The intermediate
level allows the insertion of an intercooler to reduce the amount
of superheat in the compressors. However, it is not feasible to
use an intercooler in a lower temperature cycle. In this case,
an economizer (phase separator) can be substituted with a
presaturator (vapor cooler) to reduce the superheat to zero, as
shown in Figure 6c and d.14 The presaturator can reduce the
inlet temperature to the next compression level by direct contact
with the liquid refrigerant. However, presaturation requires a
3067

dx.doi.org/10.1021/ie302877g | Ind. Eng. Chem. Res. 2013, 52, 30653088

Industrial & Engineering Chemistry Research

Review

Figure 5. Compression refrigeration cycles.14

higher vapor refrigerant ow rate. Two or more cycles with


dierent refrigerants may be operated at the same heat exchanger
as shown in Figure 7.14 The cascade refrigeration cycle is
employed to provide very low-temperature refrigeration in cases
where a single refrigerant cannot be used for operation due a
wide range of operating temperatures.
The liquefaction process is completed when the natural gas
is cooled to a temperature in the two-phase region. A simple
Linde liquefaction process, which depends solely on throttling
expansion, is shown is Figure 8.15 Feed gas is mixed with
the uncondensed portion of the gas from the previous cycle,
and the mixture is compressed by a multistage compressor.
The compressed gas is precooled to ambient temperature and
can be cooled even further by refrigeration. The high-pressure

gas is cooled by a return gas stream in a heat exchanger and


then expanded through a throttle valve, and the outlet stream
exists in an equilibrium state of liquid and vapor phases. The
outlet stream is ashed in the separator, producing liqueed
product at its bottom. Gas from the separator is used to cool
the high-pressure gas stream in the heat exchanger above.
The reverse-Brayton cycle, as shown in Figure 9,15 uses gas
phase refrigerant. This cycle is widely used for cryogenic liquefaction, and it forms the basis of the expander process in natural gas
liquefaction. The fundamental principles of a variety of gas liquefaction processes and how these processes work eciently have
been reviewed by Barron,16 Walker,17 and Timmerhaus and
Flynn.18 Natural gas liquefaction processes have been developed
by combining characteristics of dierent refrigeration cycles.
3068

dx.doi.org/10.1021/ie302877g | Ind. Eng. Chem. Res. 2013, 52, 30653088

Industrial & Engineering Chemistry Research

Review

Figure 6. Multistage compression refrigeration cycles.14

in the worlds rst LNG plant in CAMEL. This process comprises three separate refrigeration cycles as shown in Figure 10.13
The three refrigerant cycles are typically operated at three
evaporation temperature levels with multistage compression.
High-pressure propane is condensed by ambient air or cooling
water in the multistage compression step. In the propane cycle,
the propane is used to cool the natural gas and the other
two refrigerants to 30 C. The ethylene cycle then cools the
natural gas and methane to about 100 C. Finally, the methane
is used to produce LNG at 160 C.19
A kettle-type heat exchanger is employed in the propane
cycle, and a coil-wound heat exchanger (CWHE) is used in the
ethylene and methane cycles. Steam turbines contribute to drive
compressors for each refrigerant cycle and condensers uses
cooling water as a coolant. The three-train capacity is a mere 1.1
MTPA.
Phillips Cascade.19 The LNG plant in Kenai applied an early
version of the Phillips cascade process similar to that used in the
CAMEL plant. This process also uses propane, ethylene, and
methane cycles, but the single train capacity is 50% greater than
that of the three trains at CAMEL. This process is considered
to be the rst that employed gas turbine/compressor sets and a
plate-n heat exchanger (PFHE) in each refrigeration cycle.19
Phillips Optimized Cascade (POC).2023 A new version of
the cascade process, known as Phillips Optimized Cascade
(POC), was developed and applied to a Trinidad LNG plant in
1999 by ConocoPhillips as shown in Figure 11.21 This process
also uses three pure refrigerants (propane, ethylene, and methane),
and each cycle is operated separately at multiple pressure levels.

Figure 7. Cascade refrigeration cycle.14

2.2. Commercial Natural Gas Liquefaction Processes.


Many liquefaction processes have been developed and applied in
LNG plants over the last few decades. The processes can be
classied into three general categories based on the type of
refrigeration cycle and equipment used: a cascade process using
pure refrigerants, a mixed refrigerant process using refrigerant
mixtures, and an expander process using expanders instead of
JouleThomson (JT) valves. Each category is described in
detail in the next three subsections. Additional classication is
possible based on whether or not a precooling process is applied.
2.2.1. Cascade. Technip/Air Liquide Cascade.19 The rst
cascade process, designed by Technip/Air Liquide, was applied
3069

dx.doi.org/10.1021/ie302877g | Ind. Eng. Chem. Res. 2013, 52, 30653088

Industrial & Engineering Chemistry Research

Review

Figure 8. Simple Linde liquefaction process.15.

relatively low. Therefore, this process is suitable for large


capacity trains. The current capacity per train is 5.2 MTPA.
2.2.2. Mixed Refrigerant. APCI Single Mixed Refrigerant
(SMR).19 The rst single mixed refrigerant (SMR) process was
developed by APCI and applied in the Marsa el-Brega plant in
1970. This process employs a single cycle with a mixture of
nitrogen and hydrocarbons (methane, ethane, propane, etc.) as
a refrigerant instead of several pure refrigerants as in the cascade
process. The condensation and evaporation steps are carried out
in a single cycle over a wide range of temperatures to cool the
natural gas to about 160 C.19 The cycle uses CWHE as the
main cryogenic heat exchanger (MCHE).
This process was developed to decrease the amount of
equipment required compared to the cascade process; however,
a relatively greater amount of power is required for this process
than the cascade process due to the larger refrigerant ows in
former process.
Technip/Air Liquide TEALARC.24 The TEALARC twopressure single mixed refrigerant process was developed by
Technip/Air Liquide and applied in the Skikda plant. This
process uses a single mixed refrigerant cycle with a 33% increase
in capacity per train. This process consists of two refrigeration
cycles, precooling and liquefaction, as shown in Figure 12.24 In
the precooling cycle, the refrigerant used in the liquefaction
cycle is precooled and partially condensed by a mixed refrigerant
composed mainly of ethane and propane. The liquefaction cycle
cools the natural gas by a mixed refrigerant comprising mainly
methane and ethane.
Black & Veatch Pritchard PRICO.2527 The PRICO process,
one of the well-known SMR processes, was developed by Black
& Veatch Pritchard.25 This process was used in the Skikda
(Algeria) plant in the 1970s. This process uses a single mixed
refrigerant composed of a set of nitrogen and hydrocarbons
such as methane, ethane, propane, butane, and pentane.22,26 As
shown in Figure 13,25 cooling and liquefaction steps are
performed at several pressure levels. The mixed refrigerant is
compressed by a compressor and then condensed in the main
heat exchanger. The refrigerant is expanded through a Joule
Thomson (JT) valve and then evaporated as it returns through
the main heat exchanger. This process uses PFHEs in cold
boxes. The axial compressors are driven by steam turbines.22
APCI Propane Precooled Mixed Refrigerant (C3MR).2832
The propane precooled mixed refrigerant (C3MR) process
developed by APCI is dominant in the LNG plant market.
A basic schematic of the C3MR process, which consists of two
main stages, namely propane precooling and mixed refrigerant
(MR) stages, is shown in Figure 14.28 The precooling cycle
cools the natural gas to around 40 C at three or four dierent
pressure levels using a pure propane refrigerant. This cycle may

Figure 9. Reverse-Brayton cycle.15

Figure 10. Classic cascade process.13

The process has evolved from the original cascade process,


in that the methane cycle is now an open cycle or a feed-ash
system, rather than a closed cycle. This improvement enables
a separate fuel gas compressor to be eliminated, as well as
allowing stored vapors and vapors from tanker loading to be
used for reliquefaction rather than being routed directly to fuel
or are, thereby increasing LNG production.20 The parallel
arrangement of gas turbine/compressor sets in each refrigerant
cycle increases availability and allows easier operation. Further,
the process conguration enables the same amount of power to
be used in each cycle.21
Refrigeration and liquefaction are achieved in a series of
PFHEs arranged in vertical cold boxes. Precooling can be carried
out in a core-in-kettle type exchanger. The refrigerants are
compressed by centrifugal compressors driven by gas turbines.22
The latest POC process uses highly ecient aero-derivative gas
turbines for the liquefaction process.23
The cascade process requires relatively high capital and
maintenance costs due to the amount of equipment needed for
the refrigerant cycle, even though the power requirements are
3070

dx.doi.org/10.1021/ie302877g | Ind. Eng. Chem. Res. 2013, 52, 30653088

Industrial & Engineering Chemistry Research

Review

Figure 11. POC process.21

Figure 12. TEALARC process.24

and is ashed across a JT valve on the shell-side of the MCHE.


This stream ows downward to provide cooling duty for the
cold bundle. Then, the vapor and liquid output MR streams are
merged to provide partial cooling duty for the lower bundle.
The overall vaporized MR is compressed up to 4548 bar.
It is cooled and partially liqueed, rst by ambient air and
cooling water, and then by the propane in the precooling cycle.
Precooling is achieved in a kettle-type heat exchanger. The MR
cycle uses a CWHE as the MCHE. Propane compression is
carried out by a centrifugal compressor. In earlier plants, only
centrifugal compressors with steam turbine drivers were used for
MR compression. However, recently constructed plants use axial
compressors for the low-pressure (LP) stage and centrifugal
compressors for the high-pressure (HP) stage, together with gas
turbine drivers.
In the most recently constructed plants, the SplitMR
technology, in which a driver is associated with a set of

also be used to cool and partially liquefy the mixed refrigerant.


To employ propane for cooling the natural gas, the propane is
compressed to a high pressure at which it can be condensed by
ambient air or cooling water.
In the MR cycle, a mixed refrigerant comprising nitrogen,
methane, ethane, propane, and sometimes butane is used in a
single MCHE to liquefy and subcool the natural gas from
typically 35 C to between 150 and 162 C.29,30 The
refrigerant that has been processed in the precooling cycle is
separated in a high-pressure separator. The liquid and vapor
MR streams pass through separate circuits in the MCHE. The
liquid MR stream participates in cooling in the warm bundle
of the MCHE that cools the natural gas, and is ashed across
a JT valve on the shell-side of the MCHE. The liquid MR
evaporates and ows downward to provide cooling duty for
the lower bundle. The vapor MR stream is used to liquefy and
subcool the natural gas stream to 162 C in the cold bundle,
3071

dx.doi.org/10.1021/ie302877g | Ind. Eng. Chem. Res. 2013, 52, 30653088

Industrial & Engineering Chemistry Research

Review

Figure 15. DMR process.33

Figure 13. PRICO process.25

This process uses two CWHEs supplied by Linde or APCI,


and dierent amounts of power are required for each cycle.
This process was selected for the Sakhalin LNG project, in
which each train was designed for 4.8 MTPA LNG production.
Parallel Mixed Refrigerant (PMR) Process.37 Shell developed the parallel mixed refrigerant (PMR) process, which
consists of precooling and liquefaction cycles, as shown in
Figure 16.37 Either propane or MR can be used as a refrigerant

Figure 14. C3MR process.28

propane and MR compressors, is used in the C3MR process.


This conguration allows full utilization of the gas turbine power,
thereby increasing the train capacity for the same number of
compressors and drivers.30
Since the rst LNG plant using the C3MR process was
commissioned in 1972 in Brunei, the capacity of a single train
has increased from less than 0.5 MTPA to about 5 MTPA.
Shell and APCI Dual Mixed Refrigerants (DMR).3336 Shell
and APCI developed the dual mixed refrigerant (DMR) process
to overcome the inherent limitations of compressor size in
using a pure propane refrigerant for the C3MR process. As
shown in Figure 15,33 this process has a conguration similar to
the C3MR process, comprising two separate cycles, precooling
and liquefaction.33 Use of a mixed refrigerant (composed mainly
of ethane and propane) instead of pure propane in the
precooling cycle allows for more exible design while
maintaining the compressor conguration. Natural gas is cooled
to about 50 C in the precooling cycle and then liqueed and
subcooled to about 153 C in the liquefaction cycle using
a mixture of nitrogen, methane, ethane, and propane. LNG is
produced using a liquid expander and end-ash vessel at its
atmospheric boiling temperature of about 161 C.35

Figure 16. PMR process.37

for the precooling cycle. The main feature of the PMR process
is that two MR cycles for liquefaction are congured in parallel,
which reduces the pressure drop in the system and improves
the reliability of the plant, thereby improving process eciency;
the train capacity can reach 8 MTPA using existing
compressors.37
IFP/Axens Liquen.33,34 IFP/Axens developed the Liquen
process with two mixed refrigerant cycles, precooling and liquefaction cycles, as shown in Figure 17.33 The two refrigerants are
composed of methane, ethane, propane, butane, and nitrogen,
but the compositions dier in each cycle. The heavy MR
precooling cycle is used to cool the natural gas, and precool and
liquefy the other mixed refrigerant at three dierent pressure
levels. The light MR liquefaction cycle is used to liquefy and
subcool the natural gas. Both cycles are carried out in PFHE
3072

dx.doi.org/10.1021/ie302877g | Ind. Eng. Chem. Res. 2013, 52, 30653088

Industrial & Engineering Chemistry Research

Review

liquefy and subcool the natural gas. Similar to the single cycle
processes, precooling, liquefaction, and subcooling are carried
out in the same heat exchanger.
APCI AP-X.39,40 The AP-X process was developed from the
C3MR process by APCI. As shown in Figure 19,40 this process
comprises three cycles: a propane precooling cycle, a mixed
refrigerant cycle, and a nitrogen subcooling cycle. A unique
feature of this process is that the LNG is subcooled using a
nitrogen expander cycle rather than a mixed refrigerant cycle.39
Natural gas precooled to about 30 C in the propane cycle
using kettle-type heat exchangers is cooled and liqueed to
about 120 C in the MCHE with a mixed refrigerant. The LNG
is subcooled using cold gaseous nitrogen from the nitrogen
expander. In the nitrogen cycle, nitrogen is compressed to a high
pressure and then cooled to near ambient temperature. The highpressure nitrogen is cooled in a nitrogen PFHE-type economizer
with low-pressure nitrogen returning to the compressor. The
high-pressure nitrogen passing through the nitrogen economizer
is expanded to a low pressure to further reduce its temperature in
the expander. Compared to the C3MR process, the nitrogen
expander subcooling cycle allows the ow of both propane and
mixed refrigerant to be reduced without aecting production,
enabling much higher capacities (approximately 8 MTPA) using
existing equipment.40 CWHEs are used for the MR and nitrogen
subcooling cycles.
Statoil-Linde Mixed Fluid Cascade (MFC).41,42 The mixed
uid cascade (MFC) process was developed by Linde in
collaboration with Statoil and was applied in the Snohvit LNG
project. The capacity of a single train that uses this process
is 4 MTPA. As shown in Figure 20,41 this process is similar
to the cascade process and also consists of the three cycles of
precooling, liquefaction, and subcooling.41 Compared to the
cascade process, the MFC process has higher eciency as it uses
three mixed refrigerants instead of three pure refrigerants. The
mixed refrigerants are composed of methane, ethane, propane,
and nitrogen, but the compositions dier in each cycle. Another
feature is that the power requirements for each cycle are not the
same, unlike the POC process. The precooling cycle uses PFHE,
while the liquefaction and subcooling cycles use CWHE.22
2.2.3. Expander. Single Nitrogen Expander.43,44 Nitrogen
expander processes, which are based on reverse-Brayton and

Figure 17. Liquen process.33

arranged in a cold box. Each cycle is designed to use the same


amount of power so that the same set of drivers can be used for
the compressor across dierent cycles, which translates into
signicant cost savings. In addition, the relatively lower ow
rate of the mixed refrigerant permits a much greater train
capacity to be achieved with existing axial compressors.34
Gaz de France Integral Incorporated Cascade (CII).38 A
new single cycle process named the Integral Incorporated
Cascade (CII) process that was developed by Gaz de France is
shown in Figure 18.38 This process uses a mixture of nitrogen
and hydrocarbons (from methane to propane) as a refrigerant
and consists of three subunits: a compression line, a cycle uid
fractionation unit, and a heat-exchange line. The cycle uid
fractionation unit separates the mixed refrigerant into two
uid types: heavy uid (pentane and butane) and light uid
(nitrogen, methane, and ethane). The heavy uid is used to
precool the natural gas, while the light uid is used to liquefy
and subcool the natural gas. A single heat-exchange line consists
of two large cores of PFHEs and comprises two sections: an
upper section to precool the natural gas, and a lower section to

Figure 18. CII process.38


3073

dx.doi.org/10.1021/ie302877g | Ind. Eng. Chem. Res. 2013, 52, 30653088

Industrial & Engineering Chemistry Research

Review

Figure 19. AP-X process.40

Figure 20. MFC process.41

Figure 21. Single expander process.44

Claude cycles, are used mostly in oshore and small-scale


liquefaction plants. Nitrogen expander processes have been
widely used for cryogenic liquefaction, including LNG peakshaving, and in industrial gas liqueers.43
A simple single nitrogen expander cycle is illustrated in
Figure 21.44 Refrigeration is carried out through compression and
work-expansion using nitrogen as the refrigerant. High-pressure
nitrogen is cooled in the heat exchangers, with low-pressure

refrigerant returning to the compressor. The high-pressure


nitrogen is then work-expanded in the expander to reduce its
temperature. The expander generates simultaneously useful work,
which is usually supplied to the booster compressor. The lowpressure nitrogen from the expander liquees natural gas and
cools the high-pressure nitrogen in heat exchangers. Nitrogen
passing through the heat exchangers is compressed by the main
cycle compressor and booster-compressor. The process eciency
3074

dx.doi.org/10.1021/ie302877g | Ind. Eng. Chem. Res. 2013, 52, 30653088

Industrial & Engineering Chemistry Research

Review

Figure 22. Double expander process.45

Figure 23. Dual independent expander process.48

methane increases the need to ensure adequate spacing between


equipment to prevent jet res and blast pressure damage.48
BHP Compact LNG (cLNG).49 BHP and Linde developed a
compact LNG (cLNG) process similar to the SMR process,
except that pure nitrogen refrigerant is used. As shown in
Figure 24,49 this process is operated under two pressure levels
of nitrogen expansion to improve thermodynamic eciency.
To cool the nitrogen to a low-enough temperature to liquefy
natural gas, the cLNG process uses both self-cooling and turbo
expanders.49 The power generated from the turbo expander is
recovered and used to recompress the refrigerant.
Other Expander Processes.5054 Existing expander processes can be modied to increase their eciency.50 For
example, propane precooling can reduce power consumption
by approximately 20% in conventional expander processes.
Many expander processes have been developed by various
vendors, e.g. EXP by Kryopak,51 MiniLNG by Hamworthy,52
Optimised Expander Cycle by Kanfa Aragon,53 and LNG Smart
Liquefaction Technologies (OCX-Angle, OCX-R, OCX-2 and
NDX-1) by Mustang Engineering.54
2.3. Criteria for Process Selection. We have described
a variety of dierent commercial natural gas liquefaction
processes that are currently used in LNG plants or that have
been proposed for LNG projects. The success of a LNG project
is crucially dependent on the processes selected for natural gas

is relatively low because a pure gas refrigerant is used over a wide


temperature range. Therefore, the single nitrogen expander
process is only suitable for plants with small capacities.
Double Nitrogen Expander.45,46 The double nitrogen
expander process is a modication of the single nitrogen
expander process. This process has been widely used to liquefy
nitrogen and oxygen for the last few decades. This process
comprises two expander cycles: warm and cold expander cycles,
as shown in Figure 22.45 Both expander cycles enable natural gas
to be liqueed and subcooled at small temperature dierences,
reducing the specic power requirements but increasing the
size of the heat exchanger required. Nitrogen refrigerant can be
substituted with methane refrigerant in the existing process.
The methane refrigerant may reduce the specic power for
liquefaction, but this is outweighed by the safety implications of
using a hydrocarbon refrigerant rather than inert nitrogen.46
Dual Independent Expander.47,48 As shown in Figure 23,48
the dual independent expander process employs two separate
refrigeration cycles, methane and nitrogen cycles, to improve
process eciency by reducing the temperature dierence between
liquefaction and subcooling. While this process requires a larger
heat exchanger than a single expander process, the required
specic power can be reduced compared to that required for
a single expander process. However, the process safety decreases
because of the use of a hydrocarbon refrigerant. The use of
3075

dx.doi.org/10.1021/ie302877g | Ind. Eng. Chem. Res. 2013, 52, 30653088

Industrial & Engineering Chemistry Research

Review

exchangers. Some processes use proprietary equipment. For


example, CWHE used in C3MR, AP-X, and MFC is proprietary
to both APCI and Linde. In addition, some equipment causes
bottleneck problems in terms of the capacity of a LNG plant.
In a propane cycle, the capacity of propane compressors is
limited by the Mach number at the tip of the blades.21 The
pros and cons of equipment used in natural gas liquefaction
processes are listed in Table 3.22,56,57
Other aspects that should be considered when selecting
appropriate processes are reliability, site conditions, safety,
process requirements (NGL or LPG recovery), train capacity,
availability of equipment, space, and total costs (including
capital and operating costs). However, the priority of each item
in this list will dier depending on the characteristics of the
LNG project. In oshore plant projects, energy eciency may
be less important than safety, operability, and compactness
compared to onshore projects. The evaluation criteria for
successful oshore LNG projects are listed in Table 4.4448
The liquefaction processes used for LNG plants in operation
are summarized in Figure 26. Based on existing LNG plants and
other portfolios provided by APCI and Shell,9,58,59 a portfolio for
the selection of a liquefaction process is proposed in Figure 27.

Figure 24. cLNG process.49

liquefaction. We therefore propose a set of criteria for selecting


appropriate processes, taking technical and economic aspects
into consideration.
One key aspect is process eciency. This is not only related
to the thermodynamic eciency of the process, but also to the
eciency of the equipment. Improving thermodynamic
eciency leads to a reduction in energy consumption and
capital costs. A higher thermodynamic eciency can be achieved
by applying composite curves. Typical warming and cooling
curves in liquefaction processes are presented in Figure 25.

3. PATENTING ACTIVITIES FOR NATURAL GAS


LIQUEFACTION PROCESSES
There are a vast numbers of patents related to natural gas
liquefaction technologies. While associated patent publication
activity began before the 1970s, patent activity only started
increasing rapidly in the mid-1970s. Figure 28 shows trends in
patenting activities regarding natural gas liquefaction technologies in the U.S. for the period from 1975 to 2007.60
Patenting activity remained stagnant in the 1990s; however,
there has been a sharp increase in activity in the 2000s due to
the steep increase in demand for natural gas liquefaction plants.
A list of representative patents with a high citation index
and/or of great inuence are summarized in Table 5;32,36,6178
the dominance of leading companies such as APCI,
ConocoPhillips, Exxon Mobile, and Shell is apparent, and the
majority of the patents are focused on process developments
and associated equipment improvements. Two APCI patents
granted in the U.S. (US 4,755,200 and US 4,277,949), which
are among the most inuential patents led in the eld of
natural gas liquefaction technologies, are concerned with a
mixed refrigerant system combined with closed-loop pure
refrigerant cycle.
In general, the protability of natural gas liquefaction plants
is highly dependent on the liquefaction process design and
conguration. Process technologies have evolved from simple
cycles to complex cycles to enhance process eciency in
association with plant capacity, natural gas composition, project
objectives, and site conditions. While natural gas liquefaction
was mainly performed by partial condensation in the early
patents, since the 1990s, the focus has shifted to liquefaction or
condensation of gas or gas mixtures. Recent patent publications
are concerned with complex cycles that have a higher eciency
due to the combination of two or more dierent conventional
single cycles that have synergistic eects. The technical trends
in patents for natural gas liquefaction processes are shown in
the upper portion of Figure 29.60
Equipment technologies for the natural gas liquefaction have
focused mainly on improvements in heat exchangers and
expansion equipment. Early patents were primarily concerned
with cryogenic plate-n heat exchangers and compressors

Figure 25. Typical composite warming and cooling curves.

Mixed refrigerant processes have smaller mean temperature


dierences between the curves in the heat exchanger than pure
refrigerant processes, reecting better heat exchange performance in the mixed refrigerant processes. However, this leads to
require relatively larger heat exchangers than those in pure
refrigerant process. According to Ransbarger,55 the performance
of liquefaction process can be improved by changing the
number of refrigeration stages. This means that it is necessary to
optimize the number of stages to increase eciency.
Congurations of natural gas liquefaction processes are
summarized in Table 2. Each process has its characteristic
refrigeration systems and uses dierent refrigerants and heat
3076

dx.doi.org/10.1021/ie302877g | Ind. Eng. Chem. Res. 2013, 52, 30653088

Industrial & Engineering Chemistry Research

Review

Table 2. Comparisons of Refrigeration Cycle Congurations


refrigerant

heat exchanger

process

precooling

liquefaction

subcooling

precooling

liquefaction

POC
PRICO
C3MR
DMR
Liquen
PMR
CII
AP-X
MFC

propane
MR
propane
MR
MR
propane or MR
MR
propane
MR

ethylene

methane

PFHE or core-in-kettle
cold box (PFHEs)
core-in-kettle
CWHE
PFHE

PFHE

MR
MR
MR
parallel MR
MR
MR

heat
exchanger

type
coil-wound
plate-n

compressor

centrifugal

axial

pros
robust/high
operability
competitive
vendors
low pressure drop
low temperature
dierences
robust/light/simple
design
low manufacturing
cost
high eciency
high compression
ratio

driver

steam turbine

gas turbine

aero-derivate
turbine

electric
motor

several vendors
high reliability and
availability
cost eective/high
eciency
small space/ease of
installation
suitable for high
ow rates

high eciency
high reliability and
availability
ease of installation/
small space
shorter
maintenance
period
high eciency/no
emissions
high operability and
availability
exible
maintenance
simple layout

CWHE and PFHE


CWHE

Table 4. Evaluation Criteria for Oshore LNG Projects4448

Table 3. Pros and Cons of Available Equipment for Use in


LNG Plants22,56,57
equipment

PFHE

CWHE
CWHE
PFHE

heat-exchange line (two PFHEs)


core-in-kettle
CWHE
PFHE
CWHE

N2
MR

subcooling

criteria

cons

thermal eciency
complexity
equipment count
hydrocarbon refrigerant
storage
overall space requirement
compactness and lightness
simplicity of operation
ease of start-up/shutdown
exibility
sensitivity to vessel motion
capital investment

proprietary/more
expensive
careful design
vulnerable to upsets

impossible at high ow
rates
low eciency and
compression ratio
suitable only at high
ow rates
more expensive
vulnerable to FODa
large space
requirements
more expensive

cascade

mixed
refrigerant

expander

high
high
high
large

mediumhigh
medium
lowmedium
mediumlarge

low
low
low
none

high
low
medium
medium
high
medium
high

medium
medium
medium
low
medium
medium
lowmedium

low
mediumhigh
high
high
high
low
low

turbines. The technical trends in patents dealing with equipment


enhancements are shown in the lower section of Figure 29.60

4. RESEARCH ON THE LIQUEFACTION PROCESSES


A number of studies have investigated how to reduce the cost
and improve the eciency of liquefaction processes. These
studies can be roughly divided into those that evaluated the
performance of liquefaction processes, those that investigated
the optimal design and operation of liquefaction processes, and
those that improved the performance of entire LNG plants.
4.1. Evaluation on the Performance of Liquefaction
Processes. The majority of studies that have evaluated liquefaction processes have focused on thermodynamic eciency.
Several of these studies used exergy analysis to evaluate
thermodynamic eciency. Exergy analysis is well-known as a
useful method in studies on the eectiveness of an energy
system, and its fundamental principles and methodology can be
found in literatures.79,80 Kanoglu et al.81 developed relations
that can be used in rst/second-laws analyses of a simple
LindeHampsone cycle used in gas liquefaction systems. They
developed a set of expressions for the minimum work requirement and performed numerical calculations to obtain the
performance parameters. They found that the minimum work
depends only on the properties of the incoming and outgoing
gas streams.
Kanoglu82 performed an exergy analysis of the multistage
cascade refrigeration cycle used for natural gas liquefaction.
He developed a set of equations for exergy destruction and
exergetic eciency of the main equipment, and determined the
relations between the total exergy destruction in the cycle and
the cycle exergetic eciency. He reported that the exergetic
eciency of a multistage cascade refrigeration cycle is only

low thermal eciency


and reliability
intensive maintenance
high CO2 emissions
xed size and optimum
speed
high NOx emissions
intensive maintenance
xed size and optimum
speed

large power plant


starting problems with
large motors

FOD: Foreign object damage.

equipped with relief valves. Since the 1990, patents related to


expansion equipment combined with compressors or pumps,
and coil-wound heat exchangers have appeared. The focus of
recent patents has been alternatives to gas turbines in the
refrigeration cycle and compressors driven by single-shaft gas
3077

dx.doi.org/10.1021/ie302877g | Ind. Eng. Chem. Res. 2013, 52, 30653088

Industrial & Engineering Chemistry Research

Review

Figure 26. Current status of a LNG plant in operation.

Figure 27. Available portfolio for the selection of a liquefaction process.9,58,59

38.5%, indicating great potential for improvement. Remeljej


and Hoadley45 performed an analysis of four small-scale LNG
production processes: SMR, cLNG, and two open-loop processes.
These authors made more general and diversied analyses and
comparisons through exergy analysis, and concluded that there
was still room for improvement of the process eciency within
each of the schemes.
Venkatarathnam83 evaluated the performance of the mixed
refrigerant LNG process through exergy analysis. A number of
mixed refrigerant processes were discussed: a single-stage mixed
refrigerant LNG process without phase separators, a precooled LNG process without phase separators, a single-stage
mixed refrigerant LNG process with a phase separator, a precooled
LNG process with a phase separator, a propane precooled phase
separator (C3MR) process, a mixed refrigerant precooled phase
separator (DMR) process, a LNG process with multiple phase separators (Kleemenko process), a cascade liquefaction process
operating with mixtures, and a LNG process with turbines. The
exergy eciencies of liquefaction processes employed precooling cycle were nearly the same. Further, precooled liquefaction

processes were preferable for large liquefaction systems. On


the other hand, processes with phase separators such as the
Kleemenko process or the PRICO process may be more
suitable for small LNG plants because of their simplicity.
Exergy analysis is a powerful tool with which to obtain useful
information about the maximum performance of an energy
system, the pattern of lost work, and direction for potential
improvements. Furthermore, exergy analysis can be performed
by analyzing the components of a system separately.81
Other studies have compared eciency of processes by
calculating the shaft power requirements of compressors. If the
condition of the feeds and the products and the design of the
liquefaction process are specied in advance, the shaft work is
the main determinant of the operating cost. Specic work
is determined by power consumption per unit mass of LNG
and can be relatively easily calculated. Table 6 summarizes
the specic work values of processes, which were obtained
from various references.13,32,38,39,46 Dierent specic works
were reported by dierent studies, as shown in Table 5. Each
study claimed that the dierent process was the most ecient.
3078

dx.doi.org/10.1021/ie302877g | Ind. Eng. Chem. Res. 2013, 52, 30653088

Industrial & Engineering Chemistry Research

Review

Figure 28. Trend of U.S. patenting activities for natural gas liquefaction technologies.60

Table 5. Major Cited Patents and Inuencing Patents32,36,6178


U.S. patent

year

assignee

3,957,47361
4,445,91662
4,525,18536
4,755,20032
5,791,16063
6,250,10564
6,347,53265
6,658,89066
6,691,53167
6,722,15768
6,725,68869
6,789,39470
7,100,39971
7,234,32272
4,265,30273
4,566,88574
4,880,05575
6,640,58676
7,069,73377
7,266,97678

1976
1984
1985
1988
1998
2001
2002
2003
2004
2004
2006

Exxon Research Engineering


APCI
APCI
APCI
APCI
Exxon Mobil Upstream Research
APCI
ConocoPhillips
ConocoPhillips
ConocoPhillips
Shell Oil

process
process
process
process
process
process
process
process
process
process
process

objective

mixed refrigerant cycle with multiple heat exchanger


separation of heavier hydrocarbons in a scrub column
dual mixed refrigerant cycle with staged compression
precooling cycle with pure or mixed refrigerants
enhanced control system for mixed refrigerant cycle
Dual multicomponent refrigeration cycle
Partial condensation of mixed refrigerant
Open methane cycle employed for natural gas liquefaction
Optimum conguration of drivers and compressors
nonvolatile refrigerant employed in the natural gas liquefaction system
advanced control based on model predictive control

contents

2006
2007
1981
1986
1989
2003
2006
2007

ConocoPhillips
ConocoPhillips
Rosenthal Technik AG
Shell Oil
Sundstrand
ConocoPhillips
APCI
ConocoPhillips

process
process
equipment
equipment
equipment
equipment
equipment
equipment

enhanced operation with reuxed heavies removal column


nitrogen removal from a relatively warm natural gas stream
heat exchanger with a ceramic body
compressor system for gas liquefaction
light and compact impingement plate-type heat exchanger
electric motors employed as compressor drivers
compressor driven by single-shaft gas turbines
vertical heat exchanger comprising one or more core-in-kettle heat exchangers

varied depending on the plant capacity and NGL recovery. Yin


et al.85 performed an economic analysis of a single mixed
refrigerant cycle and nitrogen expander cycle. They concluded
that the SMR process was a more suitable process for smallscale liquefaction because it is comparatively simple and
consumes less energy than the other processes they evaluated.
4.2. Optimal Design and Operation. There are a number
of studies related to the design, simulation, and optimization of
the liquefaction process. Studies that focused on the design of
the liquefaction process in a LNG plant using simulation are
discussed in the next paragraph.
Kikkawa86 studied the design of a precooling mixed refrigerant
cycle and expander cycle by simulation using CHEMCAD.

One possible reason for the dierent results is that comparisons


were made under dierent conditions. In addition, dierent
levels of optimization and the use of dierent equipment and
eciencies in each process could also explain the discrepancies
among studies.
To accurately compare various processes, evaluation criteria
should include not only operating cost, but also capital cost.
However, cost estimation of liquefaction processes is dicult
because of the absence of detailed information. Only a few
studies have incorporated capital cost into economic analyses.
Kotzot et al.84 evaluated the economics of a LNG project based
on the specic plant cost, which they expressed as $/tonne of
LNG production. They reported that the cost of a LNG plant
3079

dx.doi.org/10.1021/ie302877g | Ind. Eng. Chem. Res. 2013, 52, 30653088

Industrial & Engineering Chemistry Research

Review

Figure 29. Technical trends in patents for natural gas liquefaction processes and related equipment.60.

Table 6. Eciency Comparisons of the Liquefaction Process Based on Specic Work13,32,38,39,46


relative specic work (specic work, kJ/kg)
13

process
cascade
SMR
C3-MR
DMR
MFC
single N2 expander
C3 precooled single N2
expander
double N2 expander
a

Dam et al.

32

32a

Finn et al.
(relative to cascade)

Dam et al.
(relative to DMR)

Foerg38
(relative to MFC)

Vink et al.39
(relative to C3-MR)

Barclay et al.46
(relative to C3-MR)

1.00 (1188)
1.25 (1485)
1.15 (1366)

1.39 (1382)

1.155
1.142
1.033

1.156 (1218)
1.189 (1253)
1.000 (1054)

1.000 (1054)

1.06 (1054)
1.09 (1083)
1.00 (994)

1.025 (1080)
1.000

2.00 (2376)

3.10 (3266)
3.32 (3499)

1.70 (2020)
1.70 (2020)

1.35 (1426)

only reported relative values, not detailed values.

Terry87 performed simulation and optimization of a typical liquefaction process in a peak shaving plant using HYSYS commercial
software. Li et al.88 designed small-scale LNG systems that used
a nitrogen expander cycle and mixed refrigerant cycle based on
numerical simulations and system optimization. Cao et al.89
designed and simulated two typical types of small-scale
liquefaction processes in a skid-mounted package. They reported
that the N2-CH4 expander cycle preceded the mixed refrigerant
cycle on the premise of lacking propane precooling. In addition,
it was found that large temperature dierences and high heat
exchange loads were the primary reasons for exergy loss in heat
exchangers.
Jensen and Skogestad90 studied design optimization of the
PRICO process. The process was modeled and optimized using
gPROMS, and Multiash was used for thermodynamic
calculations. The objective function included design parameters
(e.g., heat transfer area) and operating parameters (e.g., ow
rate, pressure, splits) with respect to thermodynamics and cost.
Further, a set of key constraints concerning compressor
feasibility was discussed. Li and Ju91 conducted research on
the design of three dierent liquefaction processes including
SMR, C3MR, and N2 expander for special oshore gases in the
South China Sea. A systematic analysis and comparison of these

processes was performed to select a process suitable for use


in an oshore plant. It was concluded that the N2 expander
process was the most suitable for LNG Floating Production
Storage Ooading (FPSO) due to its inherent safety, ease of
operation, and simple and compact design. The results of
exergy analysis revealed that compression equipment and aftercoolers, expansion devices, and LNG heat exchangers were the
main equipment contributing to the total exergy losses.
Combined multistage BraytonJT cycles were proposed by
Chang et al.92 They applied thermodynamic optimization
theory for three refrigeration systems: a nitrogen Brayton and
ethylene JT (N2C2) cycle, a nitrogen Brayton and propane
precooled ethylene JT (N2C2/C3) cycle, and a nitrogen
Brayton and ethylene propane JT (N2C2C3) cycle. They
found that the N2C2C3 cycle not only had a higher
eciency than the other cycles, but also the potential for higher
capacity than the other cycles. Al-Sobhi et al.93 simulated and
optimized a typical LNG plant by using ASPEN Plus to
determine ows, temperatures, and heat duties for various
equipment and streams. In addition, they performed thermal
pinch analysis to reduce heating and cooling utilities. In a case
study, the result showed a 15% possible reduction in heating
utilities and a 29% reduction in cooling utilities.
3080

dx.doi.org/10.1021/ie302877g | Ind. Eng. Chem. Res. 2013, 52, 30653088

Industrial & Engineering Chemistry Research

Review

energy consumption. Their methodology is based on a mixedinteger nonlinear programming (MINLP) formulation. To
reduce the MINLP model complexity, thermodynamic calculations were carried out using a simplied model based on the
regression of rigorous simulation results. The optimization
problem was solved using the LINDOGloal solver in GAMS.
The optimization results were validated through rigorous
simulation using Aspen Plus. A case study of the C3MR process
showed that the proposed methodology reduced the energy
consumption by 13%.
Aspelund et al.103 developed a methodology for process
synthesis by extending traditional pinch analysis with exergy
analysis, that is, Extended Pinch Analysis and Design (ExPAnD).
ExPAnD showed great potential for improving subambient
processes. Energy requirements (total shaft work) were
minimized by optimizing the compression and expansion
work of process streams, as well as the work required to
generate necessary cooling utilities. The ecacy of the proposed
methodology was veried by applying it to design of LNG
process: exergy eciency increased from 49.7% to 85.7%
compared to standard pinch analysis. Shah and Hoadley104
proposed a shaft work targeting method for multistage gas-phase
auto refrigeration systems. This targeting method demonstrated
the relationship between the expansion/compression ratio and
the heat exchange design parameter delta Tmin. In the case of
natural gas liquefaction, a dual gas-phase system can achieve the
net shaft work requirements with values only slightly higher
than those of single mixed refrigerant systems. Jensen and
Skogestad105 warned that the method using specied delta Tmin
when designing heat exchangers could lead to wrong decisions.
The minimum temperature approach favors a more constant
temperature dierence prole; however, in the case of optimal
operation, it was found that the temperature dierence at one
end was small. These authors proposed a simplied total
annualized cost (TAC) method that considers both operating
and capital costs. In addition, the results showed that there is
a trade-o relationship between heat exchanger capital cost and
compressor power.
More recent studies have discussed mixed refrigerant systems
with a focus on mitigating the energy requirements for
refrigerant compression through optimizing operating conditions
such as pressure, ow rate, and the composition of mixed
refrigerants.
Cammarata et al.106 presented a genetic algorithm (GA)based optimization methodology for liquefaction/refrigeration
systems. Nogal et al.100 also used a GA-based approach for
optimal design of mixed refrigerant cycles. Shirazi and Mowla107
attempted to minimize energy consumption in the PRICO
process designed for a peak shaving plant using a GA
implemented in MATLAB. The set of design variables included
pressure, ow rate, and composition of the mixed refrigerant.
The PengRobinson (PR) equation of state (EOS) was used
for thermodynamic calculations of process streams. The
optimization results showed that the specic work of the
process could be reduced by 3% or 6.5% as compared to the
results of Lee et al.99 Taleshbahrami and Saari108 performed
thermodynamic simulation and optimization of the C3MR
process using a GA approach. Modeling and simulation of the
C3MR process was carried out using MATLAB. The thermodynamic properties were calculated by PR EOS. Minimization of
total compressor power was dened as the objective function.
The optimal power consumption was reduced by 23% as compared
to the base case. Furthermore, the hot and cold composite

In academia, the majority of publications have focused on


determining optimal design and operating conditions. Several
studies have used a systematic approach to design refrigeration
and liquefaction systems.
Barnes and King94 focused on the synthesis of pure
refrigerant cascade systems. They derived and evaluated
numerous design rules, and identied the minimum-cost
refrigeration system conguration using a dynamic programming method. While the approaches they used could handle
detailed equipment cost correlations and thermodynamic
property models, only a limited number of refrigerant levels
and operating temperature ranges were determined by using a
heuristically developed procedure. Cheng and Mah95 developed
an interactive computational strategy for designing cascade
refrigeration systems. Refrigerants were chosen on the basis of
the desired operating temperature range and temperature of the
process streams to be cooled. Vaidyaraman and Maranas96
proposed a systematic methodology to simultaneously select
a set of pure refrigerants for use at each stage and to design the
topology of the refrigeration system. They used mixed-integer
linear programming to minimize the weighted sum of investment and operating costs.
Vaidyaraman and Maranas97 also addressed the synthesis of
mixed refrigerant cascade cycles within an optimization
framework. Both the model equations for conguring the
mixed refrigerant cascade cycle and the design objective of
minimizing the total work input taking refrigerant compositions,
pressure levels, and vaporization fractions in the ash tanks into
account were incorporated into an optimization model that was
formulated as a nonconvex nonlinear program (NLP). Kim et
al.98 proposed a synthesis method for designing and retrotting
industrial refrigeration systems for subambient temperature
cooling, together with exergy analysis. Their proposed method
was based on optimization of the composition of the mixed
refrigerant with the objective of maximal power savings. They
reported that the shaft work requirement could be reduced by
19.5%, compared to the commercial PRICO process. Lee et al.99
developed a systematic synthesis method to determine the
optimal operating conditions of a mixed refrigerant system by
using nonlinear programming (NLP) techniques with a set of
key variables such as pressure level, refrigerant ow rate, and
refrigerant composition. Three dierent types of objective
functionsminimization of crossover, the sum of crossovers,
and shaft work requirementwere employed. They concluded
that approximately 25% of the shaft work requirement could be
reduced in the PRICO process. The limitation of their proposed
method is that it is to unable to simultaneously take pressure
levels and refrigerant ow rates into account during optimization of the refrigerant composition.
Nogal et al.100 proposed a genetic algorithm (GA)-based
approach to optimize mixed refrigerant cycles in terms of
multistage refrigerant compression, full enforcement of the
minimum temperature dierence in heat exchangers, and
simultaneous optimization of variables and capital costs. The
objective function was minimum compressor power as well as
minimum capital cost. Nogal et al.101 also proposed a design
methodology for a cascade mixed refrigerant system with
multistage heat exchangers. Optimization was performed with
a set of key decision variables, namely economic trade-os,
partition temperature, refrigerant compositions, operating
conditions, and refrigerant ow rate.
Wang et al.102 suggested a methodology for synthesizing
natural gas liquefaction systems with the target of minimizing
3081

dx.doi.org/10.1021/ie302877g | Ind. Eng. Chem. Res. 2013, 52, 30653088

Industrial & Engineering Chemistry Research

Review

minimization of Net Present Value (NPV) was optimal. However,


for cases with limited area available, such as oshore production
plants, minimization of the objective function (sum of compressor
power across all compressors plus sum of UA values across all
heat exchangers) was more appropriate.
Hassan et al.116,117 derived optimal compressor operations in
C3MR and AP-X processes to minimize total power consumption. However, one aw of their model was their failure to
account for uncertainties in operating conditions such as
natural gas feed composition, ow rate, pressure, temperature,
and seasonal variations.
There are few publications related to the control of LNG
plants. Singh and Hovd118 studied the eect of simplication
of the heat exchanger model on systematic control structure
design in the PRICO process in order to develop a dynamic
model. The simplied model (heat transfer through conduction
along the longitudinal direction of the metal wall was neglected,
enthalpy was assumed to be a conserved property rather than
internal energy, and all streams were assumed to exchange heat
via a common wall) did not aect the control structure design
and the fundamental limitation of bandwidth in the PRICO
process. Jensen and Skogestad119,120 focused on the operation
of simple refrigeration cycles. They discussed degrees of
freedom, optimality of subcooling, and selection of controlled
variables. Further, Jensen and Skogestad121 proposed a control
structure to minimize the cost during operation of the PRICO
process. Jensen and Skogestad122 focused on determining the
degrees of freedom for optimal operation and plant-wide control.
The steady-state degrees of freedom available for optimization
are very important, because they determine the number of free
variables available to solve the optimization problem as well as
how many steady-state controlled variables must be selected to
operate the process. Methods to determine the potential degrees
of freedom and the actual degrees of freedom for vapor compression cycles were presented. The C3MR and MFC processes were
studied with respect to operational degrees of freedom to solve
the optimization problem related to operation and control. The
three studies above illustrate the eect of operating strategy
(maximum production and given production) on the number of
unconstrained degrees of freedom.
Michelsen et al.123 developed systematic methods for
integrated process and control structure design and applied
them to the TEALARC process design. The proposed methods
could improve both the design and operation of plant processes,
which could translate into potential economic benets. In
particular, these authors showed how a design problem (e.g.,
determination of compressor size) could inuence the control
structure design. Michelsen et al.124 investigated how controlled
variables of the regulatory control layer in the TEALARC
process could be chosen as linear combinations of measurements using self-optimizing control principles. Self-optimizing
control can reduce the need for online reoptimization and may
be used in the process design phase to place measurements
by reducing the maximum candidate set of measurements to
the best possible subset of measurements with acceptable loss.
They proposed a relatively simple method for Successive
Selection (SS) of measurements and compared it to the more
comprehensive Partial Bidirectional Branch-and-Bound (PBB)
method for selecting measurements. Michelsen et al.24
developed a dynamic, control relevant, and mechanistic model
for operability analysis of the TEALARC process. Although the
thermodynamics of this model are simplied, it has sucient

curves were closer together after optimization. Alabdulkarem


et al.109 optimized the C3MR process to reduce power
consumption using Aspen HYSYS for thermodynamic calculations and Matlabs GA optimizer to nd the global optimum.
The propane precooling and mixed refrigerant cycles were
optimized separately. The optimization variables included not
only pressures, refrigerant composition, and refrigerant ow
rate, but also propane ow rate, pressure levels, and the split
ratio. In addition, optimization was carried out while changing
the pinch temperature in heat exchangers to investigate the
eect of pinch temperature on LNG plant power consumption.
Power consumption was signicantly reduced at a pinch
temperature of 1 K as compared to pinch temperatures of
3 or 5 K. However, a low pinch temperature (0.01 K) resulted in
little improvement as compared to 1 K. The power consumptions of C3MR process were decreased by 6.98% and 13.6%
compared to previous publications.
Aspelund et al.110 optimized the PRICO process using an
optimization-simulation model that combined the Tabu Search
(TS) and the Nelder-Mead Downhill Simplex (NMDS)
methods. The local optimal solution obtained from the TS was
ne-tuned with NMDS to reduce the number of simulations.
The optimization model was connected to Aspen HYSYS
through Microsoft Excel Visual Basic for Applications (VBA).
The objective function was to minimize energy requirements
using a set of decision variables consisting of refrigerant ow rate,
refrigerant composition, and suction and condenser pressures.
Morin et al.111 minimized energy consumption for two types of
single mixed refrigerant processes, PRICO and TEALARC, using
an evolutionary search. They demonstrated that an evolutionary
search is a worthy alternative to a sequential quadratic programming (SQP) when the process is complicated.
Paradowski et al.112 carried out a parametric study of the
C3MR process, focusing on pressures, temperatures, and
compressor speeds in the precooling cycle, as well as the
composition of mixed refrigerant. Venkatarathnam83 also
optimized mixed refrigerant processes using the Sequential
Quadratic Programming (SQP) implemented in APSEN Plus
software with a set of optimization variables, including
compression ratio and mixed refrigerant composition, and the
objective function of maximizing cycle exergy eciency. Tak
et al.113 optimized the SMR process using a NLP model to
minimize the power consumption of the compressors. Their
optimization model included the variables of the compression
ratio of the refrigeration cycle, the mixed refrigerant ow rate,
and mixed refrigerant composition. Although they reported
a decrease in power consumption, their optimization model
did not consider feasibility of heat exchangers. Wang et al.114
combined thermodynamic analysis, rigorous simulation, and
optimization to minimize energy consumption. They reported
energy savings as high as 13.7%.
Hatcher et al.115 performed a systematic analysis of
optimization formulations for natural gas liquefaction processes.
They tested eight objective functions and compared them with
each other to identify the most appropriate formulation. Four
objective functions focused on operational aspects, while the
other four functions concentrated on design aspects. They found
that the most eective objective function for operation optimization was to minimize the major operating cost of compressor
power. However, the most eective objective function for design
optimization was dependent on the requirements of those
designing and constructing the process. For the case of no restriction on the area available for LNG plant construction,
3082

dx.doi.org/10.1021/ie302877g | Ind. Eng. Chem. Res. 2013, 52, 30653088

Industrial & Engineering Chemistry Research

Review

complexity for both steady-state and dynamic operability


analyses.
4.3. LNG Plant Performance. Commercial LNG plants
have evolved along with cost-saving improvements in equipment. Equipment enhancements are also important in the
design of the liquefaction process for operation of an actual
plant. The following studies focused on replacing and
conguring equipment to improve the liquefaction process.
Kanoglu125 investigated various aspects of cryogenic turbines
intended to produce power by replacing the JT valve used in
the LNG expansion. Barclay126 investigated the potential for
improvement in the SMR and DMR processes using expansion equipment. When the JouleThomson (JT) valve was
replaced with a liquid expander or a 2-phase expander, the
process eciencies were improved by 8.58.8%. Mortazavi
et al.127 suggested several options to enhance the energy
eciency of the C3MR process. Four expansion loss recovery
options were created by replacing conventional expansion
processes with expanders. The C3MR process enhanced by
these options was modeled and simulated using Aspen Plus.
Simulation results showed that by replacing the JT valve with
a gas expander, a 2-phase expander, or a liquid turbine, the
amount of compressor power required was reduced, expansion
work was recovered, and there was an increase in LNG
production. By implementing all of these improvements, the
energy consumption per unit mass of LNG was reduced by
7.07% by deducting the recovered power from the total required
power, and 3.68% without this deduction.
Kalinowski et al.128 analyzed the eect of replacing propane
chillers with absorption refrigeration systems powered by waste
heat from the power-generating gas turbine. They demonstrated
that recovering waste heat from a 9 MW electricity generating
process could provide 5.2 MW of waste heat for additional
cooling to the LNG plant and could save 1.9 MW of electricity.
Furthermore, Mortazavi et al.129 investigated enhancements of
the C3MR process using an absorption chiller powered by gas
turbine waste heat. The absorption chillers both reduced
compressor power and gas turbine fuel consumption by 21.32%,
and the capital cost of the gas turbine driver was also reduced
due to reduced capacity. However, this improvement should
recover at least 97% of the gas turbine waste heat.
Only a few studies have focused on performance improvements of entire plants and value chains. Hudson et al.130
investigated the eects of integrating the hydrocarbon removal
step into the LNG liquefaction process. This integration could
potentially produce both LNG and a separate heavier hydrocarbon liquid product using signicantly less energy than a
process comprising separate liquids recovery and liquefaction
processes. Aspelund and Gundersen131,132 proposed a transport
chain for utilization of stranded natural gas for power
production with CO2 capture and storage. In this transport
chain, natural gas was liqueed in an oshore process by
utilizing the cold exergy in Liquid Carbon Dioxide (LCO2) and
Liquid Inert Nitrogen (LIN). An exergy eciency of 87% was
achieved in the optimized oshore process.
With regard to performance and cost, the multistream heat
exchanger (MSHE) is very important in LNG plants.
Therefore, some researchers have focused on developing a
rigorous MSHE model. Hasan et al.133 demonstrated and
predicted MSHE behavior using operational data in a LNG
plant by modeling. A superstructure of heat exchanger networks
and a piecewise quadratic function for calculating the enthalpies
of the various phases were employed. Further, Hasan et al.134

discussed extending the traditional heat exchanger network


(HEN) synthesis to accommodate nonisothermal phase changes.
The limitation of their methodology is that it only generates a
feasible solution when the underlying model is nonconvex and
highly nonlinear. Wechsung et al.135 developed an optimization
formulation for the synthesis of heat exchanger networks. Their
model combines pinch analysis, exergy analysis, and mathematical programming. Kamath et al.136 developed an equationoriented (EO) approach based on cubic EOS to focus on MSHE
modeling in the NLP formulation, with an emphasis on phase
change by piecewise constant heat capacity ow rates. They
divided the mainstream into a set of substreams depending on
state to optimize the SMR process. The composite curves and
dew point and bubble point of each stream changed when
temperature was added as a decision variable.

5. FUTURE RESEARCH DIRECTIONS AND


CONCLUSIONS
Potential improvements in the natural gas liquefaction process
used in LNG plants can be realized by taking the following into
consideration.
Development of advanced liquefaction processes in terms
of the capacities of LNG plants and the characteristics of
natural gas elds: The optimal single train capacity and
the number of trains should be determined based on
consideration of technical and economic facets to achieve
economical LNG production. Furthermore, the process
should be developed and modied according to the
characteristics of natural gas elds.
Single large train/multiple small trains
Large-scale LNG plant/small- or midscale LNG
plant
Conventional onshore LNG plant/oshore LNG
plant (LNG FPSO)
Improvement of liquefaction processes intended for use
in large-scale LNG plants in terms of economy of scale:
Because alternative equipment has been developed and
the eciency of existing equipment has been improved,
the liquefaction process should be designed on the basis
of equipment availability. For optimal design of a
liquefaction process for a large-scale LNG plant, the
methodology for process design must be integrated or
considered simultaneously together with driver selection.
A variety of recent studies have focused on improving
the protability of stranded gas elds. Equipment and
environment restrictions aect the design of the
liquefaction process in small- and midscale LNG plants.
Future studies should focus not only on improving
eciency, but also on reducing the capital cost associated
with the liquefaction process.
Design of the liquefaction process for oshore plants
depends on various additional factors that are dierent
from those for onshore plants because of the completely
dierent environment on the topside of ships. Prior to
considering process eciency, applicable technology
criteria for the liquefaction process design must be taken
into account. The criteria for oshore plants include
compact size, low weight, simple operation, easy startup
and shutdown, inherent safety, sensitivity to vessel motion,
robustness to the marine environment, and exibility to
changes in process conditions. Very few commercial
applications of pretreatment or liquefaction for LNG
3083

dx.doi.org/10.1021/ie302877g | Ind. Eng. Chem. Res. 2013, 52, 30653088

Industrial & Engineering Chemistry Research

Review

Table 7. Comparison of Optimization Methodologies83,89,90,93104,107111,113,114


publication

objective function

optimization formulation

equations of state (EOS)

process

Venkatarathnam83 exergy eciency

SQP (Aspen plus)

Cao et al.89

shaftwork requirement

Aspen Hysys optimizer (original mode)b

PR or LKP (Aspen Hysys)

Jensen and
Skogestad90
Al-Sobhi et al.93

capital and operating costs

gPROMS with Multiashc

SRK (Multiash)

heating and cooling utilities

pinch analysis with heat integration

capital and operating costs

dynamic programming with heuristics

PR and ELECNTRL (Aspen Plus),


KEMDEA (property package)
SRK

Cheng and Mah95 capital and operating costs

dynamic programming with heuristics

Vaidyaraman and
Maranas96
Vaidyaraman and
Maranas97
Kim et al.98
Lee et al.99

capital and operating costs

MILP [GAMS(CPLEX)]

total work input to the system

nonconvex NLP (MINOS)

SRK

shaftwork requirement
(i) crossover (ii) sum of crossover
(iii) shaftwork requirement
capital cost and shaftwork
requirement
shaftwork requirement

pinch analysis and NLP


pinch analysis and NLP

PR

GA

PR (WORK)

Wang et al.
Aspelund et al.103

shaftwork requirement
shaftwork requirement

PR
SRK

Shah and
Hoadley104
Shirazi and
Mowla107
Taleshbahrami and
Saari108
Alabdulkarem et
al.109
Aspelund et al.110

shaftwork requirement

MINLP [GAMS (LINDOGlobal)]


ExPAnD (pinch analysis with exergy
analysis) and heuristics
targeting method

shaftwork requirement

GA (MATLAB)

PR

cascade nitrogen
expander
PRICO

shaftwork requirement

GA

PR

C3MR

shaftwork requirement

GA (MATLAB) linked with Aspen Hysys PR (Aspen Hysys)

C3MR

shaftwork requirement

PRICO

Morin et al.111
Tak et al.113
Wang et al.114

shaftwork requirement
shaftwork requirement
shaftwork requirement

TS with NMDS (Aspen Hysys linked with


Microsoft VBA)
evolutionary search
NLP (gPROMS with Multiashc)
SRK (Multiash)
SQP (Aspen plus)
PR (Aspen Plus)

94

Barnes and King

Nogal et al.100
Nogal et al.101
102

GA

PR (Aspen Hysys)

SMR, C3MR, DMR,


Kleemenko, etc.
SMR, N2-CH4
expander
PRICO
LNG plant
cascade pure
refrigerant
cascade pure
refrigerant
cascade pure
refrigerant
cascade mixed
refrigerant
PRICO
PRICO
cascade mixed
refrigerant
cascade mixed
refrigerant
C3MR
new expander131,132

PRICO, TEALARC
PRICO
C3MR

NLP: Nonlinear Programming; MILP: Mixed-Integer Linear Programming; MINLP: Mixed-Integer Nonlinear Programming; SQP: Sequential
Quadratic Programming; GA: Genetic Algorithm; TS: Tabu Search; NMDS: NelderMead Downhill Simplex; PR: PengRobinson; LKP: Lee
KeslerPloecker; SRK: SoaveRedlichKwong. bAspen HYSYS Optimizer (original mode) includes the following types of algorithms: Function
Setup, Box Method, SQP Method, Mixed Method, Fletcher Reeves Method, and Quasi-Newton Method. cgPROMS is used for process modeling,
simulation, and optimization, and Multiash is used for physical and thermodynamic properties calculations.

major equipment such as MSHEs, drivers, and expanders


would be worthwhile.
In most studies, processes were compared by calculating
the power consumption of compressors, because these
contribute the most to the operating cost. However, this
could lead to wrong decisions as capital cost was not
taken into account. Therefore, processes should be
compared based on both capital cost and operating cost.
The minimum temperature approach is commonly used
for heat exchangers in the design and optimization of the
liquefaction process. However, this approach may lead
to wrong decisions as it does not consider the heat
exchange area. For optimal design of heat exchangers,
research must focus on equipment size and costs as well
as thermodynamic eciency.
Optimization of liquefaction processes has been mostly
performed using either deterministic or stochastic methods
with a set of decision variables including mixed refrigerant
composition, pressure levels, and mixed refrigerant ow
rate to minimize power consumption in the mixed
refrigerant processes. The optimization methodology is

FPSO have been reported. Additional studies in this area


are required to realize eective oshore LNG production.
Use of a dynamic model for eective control design is
highly recommended. If the design constraints of a
process are changed, the entire control design structure
may be aected. Therefore, control structure design
should be considered when process design is changed.
This can yield economic benets through optimal
process and control design. Additional dynamic simulation and optimization studies to improve plant-wide
control are needed.
Few studies have focused on NGL fractionation, material
recovery, or energy recovery in the liquefaction process.
Findings made in these research areas will help maximize
prot and improve the eciency of entire LNG plants
and projects.
Recent studies have been concerned with developing
new LNG processes or optimizing existing LNG
processes. The eciency of commercial LNG plants
has been enhanced due to improvements of equipment
as well as of process. Therefore, further research on
3084

dx.doi.org/10.1021/ie302877g | Ind. Eng. Chem. Res. 2013, 52, 30653088

Industrial & Engineering Chemistry Research

Review

summarized in Table 7.83,89,90,93104,107111,113,114 According to analysis of Table 7, it is expected that stochastic


methods will become of more interest when liquefaction
processes become more complex.
Few studies have focused on modeling or improving the
equipment itself. In particular, a rigorous model for a
MSHE is required to determine optimal solutions. The
other rigorous models should be developed in association
with equipment innovations, such as the use of a 2-phase
expander, liquid turbine, or absorption chiller. New
process congurations and ecient cycles coupled with
improved equipment will yield alternative processes for
new LNG projects.
To increase the accuracy of simulation and optimization,
suitable equations of state (EOS) should be selected to
calculate thermodynamic properties. PengRobinson
(PR) and SoaveRedlichKwong (SRK) EOSs are
widely used in LNG processes (Table 7). However,
Kunz et al.137 recently proposed GERG-2004, which was
specically designed to calculate the thermodynamic
properties of LNG constituents. The deviations of cubic
equations for liquid densities are low compared to those
used in the PR and SRK equations.138 To date, GERG2004 is the most suitable EOS to calculate thermodynamic properties in LNG processes.
Over the past decades, various natural gas liquefaction
processes have been proposed. Although only a few processes
have been successfully applied in commercial LNG plants,
various studies on ways to improve the liquefaction processes
have contributed greatly to technical progress in the past
decade. Most studies have focused on increasing process
eciency without evaluating capital costs, which could lead
to wrong decisions. Further research studies focused on
maximizing prot in combination with process eciency are
required to achieve more economical LNG plants.

(6) Liqueed Natural Gas: Understanding the Basic Facts; DOE NETL,
2005.
(7) The Global Liqueed Natural Gas Market: Status & Outlook; U.S.
Energy Information Administration (EIA), December 2003.
(8) Investors Handbook. http://www.investorshandbook.shell.com/
2010/servicepages/downloads/les/download.php?le=all_shell_
ih10.pdf (accessed September 3, 2012).
(9) LNG Technology. http://www.airproducts.com//media/Files/
PDF/industries/energy-lng-brochure-0408.ashx (accessed September
3, 2012).
(10) Castle, W. F. Cryogenics and Gas Separation Processes. Bull. IIR
2004, 515.
(11) World Energy Outlook 2009; International Energy Agency (IEA),
November 2009.
(12) Kirillov, N. G. Analysis of Modern Natural Gas Liquefaction
Technologies. Chem. Pet. Eng. 2004, 40 (78), 401406.
(13) Finn, A. J.; Johnson, G. L.; Tomlinson, T. R. Developments in
natural gas liquefaction. Hydrocarb. Process. 1999, 78 (4), 4759.
(14) Smith, R. Chemical Process Design and Integration; Wiley:
Chichester, West Sussex, England/Hoboken, NJ, 2005; p xxiii, 687 pp.
(15) Smith, J. M.; Van Ness, H. C.; Abbott, M. M. Introduction to
Chemical Engineering Thermodynamics, 7th ed.; McGraw-Hill: Boston,
MA, 2005; p xviii, 817 pp.
(16) Barron, R. F. Cryogenic Systems, 2nd ed.; Oxford University
Press/Clarendon Press: New York/Oxford, Oxfordshire, 1985; p xvii,
507 pp.
(17) Walker, G. Cryocoolers; Plenum Press: New York, 1983.
(18) Timmerhaus, K. D.; Flynn, T. M. Cryogenic Process Engineering;
Plenum Press: New York, 1989; p viii, 612 pp.
(19) Bosma, P.; Nagelvoort, R. K. In Liquefaction Technology;
Developments through History, 1st Annual Gas Processing Symposium,
Doha, Qatar, January 1012, 2009.
(20) Andress, D. L. The Phillips Optimized Cascade LNG Process: a
Quarter Century of Improvements; Phillips Petroleum Company:
Bartlesville, OK, 1996
(21) Martin, P.-Y.; Fischer, B. In Natural Gas Liquefaction Processes
Comparison, 14th International Conference and Exhibition on Liqueed
Natural Gas (LNG-14), Doha, Qatar, March 2124, 2004.
(22) Shukri, T. LNG Technology Selection. Hydrocarbon Eng.
February 2004.
(23) ConocoPhillips Optimized Cascade Process. http://
lnglicensing.conocophillips.com/EN/Documents/
ConocoPhillipsLNG_Brochure.pdf (accessed September 3, 2012).
(24) Michelsen, F. A.; Halvorsen, I. J.; Lund, B. F.; Wahl, P. E.
Modeling and Simulation for Control of the TEALARC Liquified
Natural Gas Process. Ind. Eng. Chem. Res. 2010, 49 (16), 73897397.
(25) Stebbing, R.; OBrien, J. An Updated Report on the PRICO
Process for LNG Plants. In Gastech 75: LNG & LPG Technology
Congress, Paris, France, September 30October 3, 1975.
(26) Singh, A.; Hovd, M. Dynamic Modeling and Control of the
PRICO LNG process. In 2006 AIChE Annual Meeting, San Francisco,
CA, November 1217, 2006.
(27) Swenson, L. K. Single mixed refrigerant, closed loop process for
liquefying natural gas. U.S. Patent 4,033,735, Jul. 5, 1977.
(28) Schmidt, W.; Kennington, B. Air Products meets requirements
of full range of Floating LNG concepts. LNG J. March 10, 2011.
(29) McKeever, J.; Pillarella, M.; Bower, R. An ever evolving
technology. LNG Industry Spring, 2008.
(30) Pillarella, M.; Liu, Y.-N.; Petrowski, J.; Bower, R. The C3MR
liquefaction Cycle: Versatility for a Fast Growing, Ever Changing LNG
Industry. In 15th International Conference on LNG (LNG-15),
Barcelona, Spain, April 2427, 2007.
(31) Longsworth, R. C. Cryostat with serviceable refrigerator. U.S.
Patent 4,277,949, Jul. 14, 1981.
(32) Liu, Y.-N.; Newton, C. L. Feed gas drier precooling in mixed
refrigerant natural gas liquefaction processes. U.S. Patent 4,755,200,
Jul. 5, 1988.

AUTHOR INFORMATION

Corresponding Author

*Tel.: +82-2-363-9375. Fax: +82-2-312-6401. E-mail: ilmoon@


yonsei.ac.kr.
Notes

The authors declare no competing nancial interest.

ACKNOWLEDGMENTS
This research was supported by a grant from the GAS Plant
R&D Center funded by the Ministry of Land, Transportation
and Maritime Aairs (MLTM) of the Korean government and
also respectfully supported by BK 21 Program funded by the
Ministry of Education (MOE) of Korea.

REFERENCES

(1) The outlook for energy: A view to 2040. http://www.


exxonmobil.com/Corporate/Files/news_pub_eo2012.pdf (accessed
September 3, 2012).
(2) International Energy Outlook 2011; U.S. Energy Information
Administration (EIA), September 2011.
(3) World Energy Outlook 2010; International Energy Agency (IEA),
November 2010.
(4) Natural Gas Issues and Trends; U.S. Energy Information
Administration (EIA), April 1998.
(5) Kumar, S.; Kwon, H. T.; Choi, K. H.; Lim, W.; Cho, J. H.; Tak,
K.; Moon, I. LNG: An eco-friendly cryogenic fuel for sustainable
development. Appl. Energy 2011, 88 (12), 42644273.
3085

dx.doi.org/10.1021/ie302877g | Ind. Eng. Chem. Res. 2013, 52, 30653088

Industrial & Engineering Chemistry Research

Review

24th World Gas Conference, Buenos Aires, Argentina, October 59,


2009.
(59) Jager, M.; Mahony, S.; Driel, A. v. Success Factors in new LNG
Plants. Hydrocarbon Eng. October, 2004.
(60) Patent Trends of LNG Process Technologies; Korean Intellectual
Property Oce (KIPO), December 2007.
(61) Stark, T. M.; Bodnick, S. Mixed refrigerant cycle. U.S. Patent
3,957,473, May 18, 1976.
(62) Newton, C. L. Process for liquefying methane. U.S. Patent
4,445,916, May 1, 1984.
(63) Mandler, J. A.; Brochu, P. A.; Hamilton, J. R. Method and
apparatus for regulatory control of production and temperature in a
mixed refrigerant liqueed natural gas facility. U.S. Patent 5,791,160,
Aug. 11, 1998.
(64) Kimble, E. L. Dual multi-component refrigeration cycles for
liquefaction of natural gas. U.S. Patent 6,250,105, Jun. 26, 2001.
(65) Agrawal, R.; Daugherty, T. L.; Roberts, M. J. Gas liquefaction
process with partial condensation of mixed refrigerant at intermediate
temperatures. U.S. Patent 6,347,532, Feb. 19, 2002.
(66) Hahn, P. R.; Yao, J.; Lee, R.-J.; Baudat, N. P. Enhanced methane
ash system for natural gas liquefaction. U.S. Patent 6,658,890, Dec. 9,
2003.
(67) Martinez, B. D.; Thakkar, S. R.; Hahn, P. R.; Baudat, N. P.;
Qualls, W. R. Driver and compressor system for natural gas
liquefaction. U.S. Patent 6,691,531, Feb. 17, 2004.
(68) Eaton, A. P.; Martinez, B. D. Non-volatile natural gas
liquefaction system. U.S. Patent 6,722,157, Apr. 20, 2004.
(69) Elion, W. J.; Jones, K. A.; McLachlan, G. J.; Wilson, J. H.
Controlling the production of a liqueed natural gas product stream.
U.S. Patent 6,725,688, Apr. 27, 2004.
(70) Elion, W. J.; Jones, K. A.; McLachlan, G. J.; Wilson, J. H.
Controlling the production of a liqueed natural gas product system.
U.S. Patent 6,789,394, Sep. 14, 2004.
(71) Eaton, A. P. Enhanced operation of LNG facility equipped with
reuxed heavies removal column. U.S. Patent 7,100,399, Sep. 5, 2006.
(72) Hahn, P. R.; Ritchie, P. D.; Yao, J.; Lee, R.-J.; Eaton, A. P.; Low,
W. R. LNG system with warm nitrogen rejection. U.S. Patent
7,234,322, Jun. 26, 2007.
(73) Forster, S.; Kleemann, M.; Krauth, A.; Maier, H. R.; Pohlmann,
H.-J. Heat exchanger. U.S. Patent 4,265,302, May 5, 1981.
(74) Haak, F. W. Gas liquefaction process. U.S. Patent 4,566,885, Jan.
28, 1986.
(75) Niggemann, R. E.; Nguyen, D. C. Impingement plate type heat
exchanger. U.S. Patent 4,880,055, Nov. 14, 1989.
(76) Baudat, N. P.; Martinez, B. D.; Hahn, P. R.; Weyermann, H. P.;
Qualls, W. R. Motor driven compressor system for natural gas
liquefaction. U.S. Patent 6,640,586, Nov. 4, 2003.
(77) Lucas, C. E.; Brochu, P. A.; Rooney, W. C. Utilization of
bogdown of single-shaft gas turbines to minimize relief ows in
baseload LNG plants. U.S. Patent 7,069,733, Jul. 4, 2006.
(78) Eaton, A. P.; Martinez, B. D.; Christian, M. Vertical heat
exchanger conguration for LNG facility. U.S. Patent 7,266,976, Sep.
11, 2007.
(79) Gaggioli, R. A. Available Energy and Exergy. Int. J. Appl.
Thermodyn. 1998, 1 (14), 18.
(80) Dincer, I. b.; Rosen, M. Exergy: Energy, Environment, and
Sustainable Development; Elsevier: Amsterdam/Boston, 2007; p xvii,
454 pp.
(81) Kanoglu, M.; Dincer, I.; Rosen, M. A. Performance analysis of
gas liquefaction cycles. Int. J. Energy Res. 2008, 32 (1), 3543.
(82) Kanoglu, M. Exergy analysis of multistage cascade refrigeration
cycle used for natural gas liquefaction. Int. J. Energy Res. 2002, 26 (8),
763774.
(83) Venkatarathnam, G.; Timmerhaus, K. D. Cryogenic Mixed
Refrigerant Processes; Springer: New York, 2008; p xv, 262 pp.
(84) Kotzot, H.; Durr, C.; Caswell, C.; Borsos, S. LNG liquefaction Not all plants are created equal - the sequel. In Gastech 2009, Abu
Dhabi, UAE, May 2528, 2009.

(33) Boutelant, P. Selecting an LNG Process Not an Easy Task. In


OAPEC-IFP Joint Seminar, Rueil-Malmaison, France, 1719 June,
2008.
(34) Martin, P.-Y.; Pigourier, J.; Fischer, B. LNG process selection,
no easy task. Hydrocarbon Eng. May, 2004.
(35) Dam, W.; Ho, S.-M., Unusual design considerations drive
selection of Sakhalin LNG plant facilities. Oil Gas J. October 1, 2001;
pp 5869.
(36) Newton, C. L. Dual mixed refrigerant natural gas liquefaction
with staged compression. U.S. Patent 4,525,185, Jun. 25, 1985.
(37) van de Graaf, J. M.; Pek, B. Large-capacity LNG Trains - The
Shell parallel Mixed Refrigerant Process. In Business Brieng: LNG
Review, 2005.
(38) Flesch, E.; Raillard, J.-C. CII Liquefaction Process: 2 cascades
into 1. In 12th International Conference and Exhibition on Liqueed
Natural Gas (LNG-12), Perth, Australia, May 47, 1998.
(39) Roberts, M. J.; Agrawal, R. Hybrid Cycle for the Production of
Liqueed Natural Gas. U.S. Patent 6,308,531, Oct. 30, 2001.
(40) Roberts, M. J.; Liu, Y.-N.; Bronfenbrenner, J. C.; Petrowski, J.
M. Reducing LNG Capital Cost in Todays Competitive Environment.
In 14th International Conference and Exhibition on Liqueed Natural Gas
(LNG14), Doha, Qatar, March 2124, 2004.
(41) Bauer, H. A Novel Concept for Large LNG Baseload Plants. In
2001 AIChE Spring National Meeting, Houston, TX, April 2226,
2001.
(42) Berger, E.; Forg, W.; Heiersted, R. S.; Paurola, P. The
MFC(Mixed Fluid Cascade) Process for the First European Baseload
LNG Production Plant; Linde, 2003.
(43) Vink, K. J.; Nagelvoort, R. K. Comparison of Baseload
Liquefaction Processes. In 12th International Conference and Exhibition
on Liqueed Natural Gas (LNG-12), Perth, Australia, May 47, 1998.
(44) Finn, A. J.; Johnson, G. L.; Tomlinson, T. R. LNG technology
for oshore and mid-scale plants. In 79th Annual GPA Convention,
Atlanta, Georgia, March 1315, 2000.
(45) Finn, A. J. Eective LNG Production Oshore. In 81st Annual
GPA Convention, Dallas, Texas, March 1013, 2002.
(46) Finn, A. J. Are floating LNG facilities viable options?
Hydrocarbon Process. 2009, 3138.
(47) Foglietta, J. H. Production of LNG using Dual Independent
Expander Refrigeration Cycles. In AIChE Spring National Meeting,
New Orleans, LA, March 1014, 2002.
(48) Foglietta, J. H., Consider dual independent expander
refrigeration for LNG production. Hydrocarbon Process. January, 2004.
(49) Remeljej, C. W.; Hoadley, A. F. A. An exergy analysis of smallscale liquefied natural gas (LNG) liquefaction processes. Energy 2006,
31 (12), 20052019.
(50) Barclay, M.; Denton, N., Selecting oshore LNG processes.
LNG J. October, 2005.
(51) The Kryopak EXP LNG Process. http://www.lngplants.com/
KryopakTypicalProcessDescription.html (accessed September 3,
2012).
(52) Small Scale and MiniLNG Systems for LNG production and
emission recovery. http://www.hamworthy.com/PageFiles/177/
Onshore%20Brochure%20%20Small%20Scale%20and%20Mini%20LNG%20Systems.pdf (accessed September 3, 2012).
(53) Waldmann, I. B. Evaluation of Process Systems for Floating
LNG Production Units. In Tekna Conference, June 1819, 2008.
(54) Susan Walther, P. E.A new generation of liquefaction processes
for LNG FPSO applications. In International Gas Union Research
Conference, Paris, October 08, 2008.
(55) Ransbarger, W. A fresh look at LNG process eciency. LNG
Industry Spring, 2007.
(56) Finn, A. J. New FPSO design produces LNG from oshore
sources. Oil Gas J. August 26, 2002.
(57) Price, B. C. Small-scale LNG facility development. Hydrocarbon
Process. January, 2003; pp 3739.
(58) Bradley, A.; Duan, H.; Elion, W.; Soest-Vercammen, E. v.;
Nagelvoort, R. K. Innovation in the LNG industry: Shells approach. In
3086

dx.doi.org/10.1021/ie302877g | Ind. Eng. Chem. Res. 2013, 52, 30653088

Industrial & Engineering Chemistry Research

Review

(85) Yin, Q. S.; Li, H. Y.; Fan, Q. H.; Jia, L. X. Economic analysis of
mixed-refrigerant cycle and nitrogen expander cycle in small scale
natural gas liqueer. In Cryogenic Engineering Conference, Chattanooga,
TN, July 1620, 2007.
(86) Kikkawa, Y.; Nakamura, M.; Sugiyama, S. Development of
liquefaction process for natural gas. J. Chem. Eng. Jpn. 1997, 30 (4),
625630.
(87) Terry, L. Comparison of liquefaction process. LNG J. 1998, 21
(3), 2833.
(88) Li, H. Y.; Jia, L. X.; Fan, Q. H.; Yin, Q. S. In Numerical
Simulation and Optimazation of Small Scale LNG Plant. In Cryogenic
Engineering Conference, Keystone, Colorado, August 29September 2,
2005.
(89) Cao, W. S.; Lu, X. S.; Lin, W. S.; Gu, A. Z. Parameter
comparison of two small-scale natural gas liquefaction processes in
skid-mounted packages. Appl. Therm. Eng. 2006, 26 (89), 898904.
(90) Jensen, J. B.; Skogestad, S. Single-cycle mixed-uid LNG process
Part I: Optimal design. In 1st Annual Gas Processing Symposium, Doha,
Qatar, January 1012, 2009.
(91) Li, Q. Y.; Ju, Y. L. Design and analysis of liquefaction process for
offshore associated gas resources. Appl. Therm. Eng. 2010, 30 (16),
25182525.
(92) Chang, H. M.; Chung, M. J.; Lee, S.; Choe, K. H. An efficient
multi-stage Brayton-JT cycle for liquefaction of natural gas. Cryogenics
2011, 51 (6), 278286.
(93) Al-Sobhi, S. A.; Alfadala, H. E.; El-Halwagi, M. M. Simulation
and Energy Integration of a Liqueed Natural Gas (LNG) Plant. In 1st
Annual Gas Processing Symposium, Doha, Qatar, January 1012, 2009.
(94) Barnes, F. J.; King, C. J. Synthesis of cascade refrigeration and
liquefaction systems. Ind. Eng. Chem., Process Des. Develop. 1974, 13
(4), 421433.
(95) Cheng, W. B.; Mah, R. S. H. Interactive Synthesis of Cascade
Refrigeration Systems. Ind. Eng. Chem., Process Des. Develop. 1980, 19
(3), 410420.
(96) Vaidyaraman, S.; Maranas, C. D. Optimal synthesis of
refrigeration cycles and selection of refrigerants. AICHE J 1999, 45
(5), 9971017.
(97) Vaidyaraman, S.; Maranas, C. D. Synthesis of mixed refrigerant
cascade cycles. Chem. Eng. Commun. 2002, 189 (8), 10571078.
(98) Kim, J. K.; Lee, G. C.; Zhu, F.; Smith, R. Cooling system design.
Heat Transfer Eng. 2002, 23 (6), 4961.
(99) Lee, G. C.; Smith, R.; Zhu, X. X. Optimal synthesis of mixedrefrigerant systems for low-temperature processes. Ind. Eng. Chem. Res.
2002, 41 (20), 50165028.
(100) Del Nogal, F.; Kim, J. K.; Perry, S.; Smith, R. Optimal Design
of Mixed Refrigerant Cycles. Ind. Eng. Chem. Res. 2008, 47 (22),
87248740.
(101) Del Nogal, F.; Kim, J. K.; Perry, S.; Smith, R. Synthesis of
Cryogenic Energy Systems. In 18th European Symposium on Computer
Aided Process Engineering, Lyon, France, June 14, 2008.
(102) Wang, M. Q.; Zhang, J.; Xu, Q. Optimal design and operation
of a C3MR refrigeration system for natural gas liquefaction. Comput.
Chem. Eng. 2012, 39, 8495.
(103) Aspelund, A.; Berstad, D. O.; Gundersen, T. An Extended
Pinch Analysis and Design procedure utilizing pressure based exergy
for subambient cooling. Appl. Therm. Eng. 2007, 27 (16), 26332649.
(104) Shah, N. M.; Hoadley, A. F. A. A targeting methodology for
multistage gas-phase auto refrigeration processes. Ind. Eng. Chem. Res.
2007, 46 (13), 44974505.
(105) Jensen, J. B.; Skogestad, S. Problems with specifying Delta Tmin in the design of processes with heat exchangers. Ind. Eng. Chem.
Res. 2008, 47 (9), 30713075.
(106) Cammarata, G.; Fichera, A.; Guglielmino, D. Optimization of a
liquefaction plant using genetic algorithms. Appl. Energy 2001, 68 (1),
1929.
(107) Shirazi, M. M. H.; Mowla, D. Energy optimization for
liquefaction process of natural gas in peak shaving plant. Energy 2010,
35 (7), 28782885.

(108) Taleshbahrami, H.; Saffari, H. Optimization of the C3mr Cycle


with Genetic Algorithm. Trans. Can. Soc. Mech. Eng. 2010, 34 (34),
433448.
(109) Alabdulkarem, A.; Mortazavi, A.; Hwang, Y. H.; Radermacher,
R.; Rogers, P. Optimization of propane pre-cooled mixed refrigerant
LNG plant. Appl. Therm. Eng. 2011, 31 (67), 10911098.
(110) Aspelund, A.; Gundersen, T.; Myklebust, J.; Nowak, M. P.;
Tomasgard, A. An optimization-simulation model for a simple LNG
process. Comput. Chem. Eng. 2010, 34 (10), 16061617.
(111) Morin, A.; Wahl, P. E.; Molnvik, M. Using evolutionary search
to optimize the energy consumption for natural gas liquefaction. Chem.
Eng. Res. Des. 2011, 89 (11A), 24282441.
(112) Paradowski, H.; Bamba, M. O.; Bladanet, C. Propane
Precooling Cycles for increased LNG train capacity. In 14th
International Conference and Exhibition on Liqueed Natural Gas
(LNG-14), Doha, Qatar, March 2124, 2004.
(113) Tak, K.; Lim, W.; Choi, K.; Ko, D.; Moon, I. Optimization of
mixed-refrigerant system in LNG liquefaction process. In 21st
European Symposium on Computer Aided Process Engineering, Chalkidiki,
Greece, 2011; Vol. 29, pp 18241828.
(114) Wang, M. Q.; Zhang, J.; Xu, Q.; Li, K. Y. ThermodynamicAnalysis-Based Energy Consumption Minimization for Natural Gas
Liquefaction. Ind. Eng. Chem. Res. 2011, 50 (22), 1263012640.
(115) Hatcher, P.; Khalilpour, R.; Abbas, A. Optimisation of LNG
mixed-refrigerant processes considering operation and design
objectives. Comput. Chem. Eng. 2012, 41, 123133.
(116) Hasan, M. M. F.; Karimi, I. A.; Alfadala, H. E. Optimizing
Compressor Operations in an LNG Plant. In 1st Annual Gas Processing
Symposium, Doha, Qatar, January 1012, 2009.
(117) Hasan, M. M. F.; Razib, M. S.; Karimi, I. A. Optimization of
Compressor Networks in LNG Operations. In 10th International
Symposium on Process Systems Engineering, Salvador, Brazil, August 16
20, 2009.
(118) Singh, A.; Hovd, M. Model Requirement for Control Design of
an LNG Process. In 17th European Symposium on Computer Aided
Process Engineering, Bucharest, Romania, May 2730, 2007.
(119) Jensen, J. B.; Skogestad, S. Optimal operation of simple
refrigeration cycles. Part I: Degrees of freedom and optimality of subcooling. Comput. Chem. Eng. 2007, 31 (56), 712721.
(120) Jensen, J. B.; Skogestad, S. Optimal operation of simple
refrigeration cycles. Part II: Selection of controlled variables. Comput.
Chem. Eng. 2007, 31 (12), 15901601.
(121) Jensen, J. B.; Skogestad, S. Single-cycle mixed-uid LNG
process Part II: Optimal operation. In 1st Annual Gas Processing
Symposium, Doha, Qatar, January 1012, 2009.
(122) Jensen, J. B.; Skogestad, S. Steady-State Operational Degrees of
Freedom with Application to Refrigeration Cycles. Ind. Eng. Chem. Res.
2009, 48 (14), 66526659.
(123) Michelsen, F. A.; Halvorsen, I. J.; Lund, B. F. The impact of
process design decisions on operability and control of an LNG
process. J. Nat. Gas Sci. Eng. 2010, 2, 183191.
(124) Michelsen, F. A.; Lund, B. F.; Halvorsen, I. J. Selection of
Optimal, Controlled Variables for the TEALARC LNG Process. Ind.
Eng. Chem. Res. 2010, 49 (18), 86248632.
(125) Kanoglu, M. Cryogenic turbine efficiencies. Exergy, Int. J 2001,
1 (3), 202208.
(126) Barclay, M. A. Oshore LNG: The Perfect Starting Point for
the 2-Phase Expander? In Oshore Technology Conference (OTC),
Houston, Texas, May 14, 2006.
(127) Mortazavi, A.; Somers, C.; Hwang, Y.; Radermacher, R.;
Rodgers, P.; Al-Hashimi, S. Performance enhancement of propane precooled mixed refrigerant LNG plant. Appl. Energy 2012, 93, 125131.
(128) Kalinowski, P.; Hwang, Y.; Radermacher, R.; Hashimi, S.;
Rodgers, P. Application of waste heat powered absorption refrigeration
system to the LNG recovery process. Int. J. Refrig. 2009, 32 (4), 687
694.
(129) Mortazavi, A.; Somers, C.; Alabdulkarem, A.; Hwang, Y.;
Radermacher, R. Enhancement of APCI cycle efficiency with
absorption chillers. Energy 2010, 35 (9), 38773882.
3087

dx.doi.org/10.1021/ie302877g | Ind. Eng. Chem. Res. 2013, 52, 30653088

Industrial & Engineering Chemistry Research

Review

(130) Hudson, H. M.; Wilkinson, J. D.; Cuellar, K. T.; Pierce, M. C.


Integrated liquids recovery technology improves LNG production
eciency. In 82nd Annual Convention of the Gas Processors Association,
San Antonio, Texas, March 11, 2003.
(131) Aspelund, A.; Gundersen, T. A liquefied energy chain for
transport and utilization of natural gas for power production with
CO(2) capture and storage - Part 1. Appl. Energy 2009, 86 (6), 781
792.
(132) Aspelund, A.; Gundersen, T. A liquefied energy chain for
transport and utilization of natural gas for power production with
CO(2) capture and storage - Part 2: The offshore and the onshore
processes. Appl. Energy 2009, 86 (6), 793804.
(133) Hasan, M. M. F.; Karimi, I. A.; Alfadala, H. E.; Grootjans, H.
Operational Modeling of Multistream Heat Exchangers with Phase
Changes. AICHE J. 2009, 55 (1), 150171.
(134) Hasan, M. M. F.; Jayaraman, G.; Karimi, I. A.; Alfadala, H. E.
Synthesis of Heat Exchanger Networks with Nonisothermal Phase
Changes. AICHE J. 2010, 56 (4), 930945.
(135) Wechsung, A.; Aspelund, A.; Gundersen, T.; Barton, P. I.
Synthesis of Heat Exchanger Networks at Subambient Conditions with
Compression and Expansion of Process Streams. AICHE J. 2011, 57
(8), 20902108.
(136) Kamath, R. S.; Biegler, L. T.; Grossmann, I. E. Modeling
Multistream Heat Exchangers with and without Phase Changes for
Simultaneous Optimization and Heat Integration. AICHE J. 2012, 58
(1), 190204.
(137) Kunz, O.; Klimeck, R.; Wagner, W.; Jaeschke, M. The GERG2004 Wide-Range Equation of State for Natural Gases and Other
Mixtures; VDI Verlag GmbH Verlag des Vereins Deutscher Ingenieure
(Publishing House of the Association of German Engineers):
Dusseldorf, 2007.
(138) Dauber, F.; Span Embedding a Thermodynamic-Property
Model for Natural Gases in a Simulation Tool for Modelling LNG
Evaporation and Liquefaction Processes with Higher Accuracy. In 5th
CAPE-OPEN European Conference, Cambridge, UK April 34, 2008.

3088

dx.doi.org/10.1021/ie302877g | Ind. Eng. Chem. Res. 2013, 52, 30653088

Anda mungkin juga menyukai