OTC 18012

Offshore LNG: The Perfect Starting Point for the 2-Phase Expander?
M.A. Barclay, Foster Wheeler Energy Ltd., and C.C. Yang, Foster Wheeler North America Corp.

Copyright 2006, Offshore Technology Conference

Introduction
This paper was prepared for presentation at the 2006 Offshore Technology Conference held in Increasing global demand for natural gas is supporting the
Houston, Texas, U.S.A., 1–4 May 2006.
rapid growth of worldwide LNG production capacity. As
This paper was selected for presentation by an OTC Program Committee following review of
information contained in an abstract submitted by the author(s). Contents of the paper, as
demand continues to grow and the value of natural gas
presented, have not been reviewed by the Offshore Technology Conference and are subject to remains high, the impetus to monetise non-traditional gas
correction by the author(s). The material, as presented, does not necessarily reflect any
position of the Offshore Technology Conference, its officers, or members. Papers presented at resources also grows. Offshore floating LNG production has
OTC are subject to publication review by Sponsor Society Committees of the Offshore
Technology Conference. Electronic reproduction, distribution, or storage of any part of this
generated interest because it offers the potential to:
paper for commercial purposes without the written consent of the Offshore Technology
Conference is prohibited. Permission to reproduce in print is restricted to an abstract of not
• avoid flaring or reinjection of associated gas
more than 300 words; illustrations may not be copied. The abstract must contain conspicuous • monetise smaller or remote fields of non-associated
acknowledgment of where and by whom the paper was presented. Write Librarian, OTC, P.O.
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435. gas
• reduce exposure to public and increase security of
Abstract facilities
Offshore natural gas liquefaction has become increasingly • lower LNG production costs
economically viable as the value of LNG has increased during The realisation of large floating production, storage, and off-
the past few years. This has increased interest in both onshore loading (FPSO) facilities for oil production and LPG
and offshore liquefaction projects. Offshore liquefaction may production, use of barge transport for the Snøhvit LNG
facilitate the introduction of new technologies and processes facility, and other developments demonstrate the potential for
because their project economics favour different criteria. For offshore LNG.
offshore applications, process design criteria such as
compactness, weight, modular design, and process safety Liquefaction Cycle Development (1, 2, 3, 4 ).
become more important. The base-load LNG industry now has over 40 years of
This paper considers the potential process improvements history starting with permanent operations of the Camel plant
that could be realised for mixed refrigerant-based processes in Algeria in 1964. The earliest plants consisted of fairly
using two-phase expander. This type of expander has recently simple liquefaction processes based on either cascaded
been developed for commercial application and operated on a refrigeration or single mixed refrigerant (MR) processes with
small-scale LNG project in Poland. Integration of two-phase train capacities less than one million tonnes per annum
expanders into mixed refrigerant processes offers the potential (MTPA). These were quickly replaced by the two-cycle
to decrease equipment footprint, installed compressor propane pre-cooled mixed refrigerant (C3MR) process
capacity, and flammable liquid inventory but these expanders developed by Air Products and Chemicals Inc. (APCI). This
are still new and have yet to be proven in a large capacity process became the dominant liquefaction process technology
plant. This paper shares the simulation results for dual and by the late 1970s and is stillcompetitive in many cases.
single mixed refrigerant processes using three different Economies of scale, improved process simulation tools and
expansion devices including two-phase expanders. These equipment performance (e.g. liquid expanders and gas turbine
results support a discussion of the potential benefit of drivers) have improved performance and increased the
improved expander operating range on process and facility capacity of liquefaction trains. Recently, three-cycle
design. For onshore liquefaction, expanders would likely be processes such as the AP-X™ and the ConocoPhillips
used to boost efficiency to increased LNG production for a Optimised Cascade have been selected for new projects. The
defined compressor and driver. Offshore however, two-phase third cycle on the AP-X™ process allow onshore train
expanders could enable use of single cycle mixed refrigerant capacities to climb to approximately 7.5-10+ MTPA (5) and
liquefaction processes. Such processes benefit greatly from have thus circumvented the typical process bottlenecks like the
the use of two-phase expansions and may offer the potential to main cryogenic heat exchanger diameter and propane
decrease footprint, weight, complexity, and flammable fluid refrigerant compressor capacity.
inventor. All advantages that are highly relevant to offshore
liquefaction facilities. The results encourage more detailed
analysis of processes incorporating two-phase expanders for
offshore liquefaction.
2 OTC 18012

Expansion Devices (6). and are thus only considered suitable for small fields (11).
Liquid refrigerant-based liquefaction processes until the These factors have all been examined in various studies such
1990s were based exclusively on isenthalpic expansion as Project Azure and the Shell development work on floating
through JT valves. These valves expand a liquid stream to a and offshore concepts (10, 12).
two-phase stream at reduced pressure and temperature. Since
they feature no work recovery, the expansion is irreversible The Present Work.
and pressure availability is destroyed through viscous This paper shares the simulation results for dual and single
dissipation such that the end state for the fluid is either at mixed refrigerant processes using the three different expansion
higher temperature and/or vapour fraction. devices described above. These results support a discussion of
During the past decade liquid expanders have been the potential benefit of improved expander operating range on
introduced and become widely accepted in both new process and facility design. Included in the analysis is one
liquefaction plants as well as a common retrofit of existing process, the single MR with the use of 2-phase expansion, that
facilities starting with MLNG Dua in 1996. These expanders has not been rigorously examined for off-shore liquefaction.
are essentially a pump run backwards that allow a subcooled The development of the two-phase expander has enabled more
liquid to be expanded almost to its bubble point isentropically. effective expansion of large-phase envelope fluids such as
The work of the expansion is extracted from the fluid as shaft those required to cover the entire temperature range for a LNG
power than may be used to drive a centrifugal compressor or process.
electrical generator elsewhere in the process. The recovery of
power is valuable, especially offshore, but more importantly, Simulation Work
when the fluid does work (by expansion) it lowers the fluid The dual MR and single MR (SMR) processes were simulated
temperature. This is particularly important when the fluid, because they feature flashing liquid expansion. Each MR
either refrigerant or LNG, is at very low temperature as more process was simulated with three different types of fluid
power is required to reject heat at a lower temperature. By expansions to generate a total of six liquefaction process
doing work at a very low temperature, it is the most efficient simulations. Simulated expansion options were expansion
way of rejecting heat. valves only (isenthalpic), liquid turbine expansion followed by
Recently two-phase or flashing expanders have become expansion valve (isentropic /isenthalpic), or two-phase turbine
available. These expanders are able to isentropically expand a expanders (isentropic). This generated a total of six different
liquid into the vapour dome and maximise work recovery and liquefaction processes that were simulated using Hysys
the reversibility of the expansion. Such expanders result in process simulation software to benchmark relative
expanded fluids at either lower vapour quality or lower performance.
temperature and thus can serve as the basis for more efficient The dual and single MR processes were chosen because
liquefaction. The two-phase expanders have not yet been they offer good potential for offshore liquefaction and the
proven at large scale. In all cases, where liquid or two-phase potential to realise some benefits offered by two-phase
expanders are used in the production of LNG, they have a JT expansion. The dual mixed refrigerant process has been
valve by-pass. previously identified as offering promise for off-shore
liquefaction because it offers efficiencies comparable to the
Offshore Liquefaction. C3MR but decreases the large propane inventory required for
Many studies have discussed process requirements for the multiple stage propane refrigeration system that precools
offshore LNG. Offshore liquefaction processes must be more the MR. The dual MR also allows balancing of the
compact and light weight, support modular design, and offer compressor and driver loads to maximise LNG production and
higher inherent process safety than traditional onshore increase train capacity. Simplified process descriptions and
processes. Offshore processes must also consider deployment diagrams of both the SMR and dual MR processes follow
and operation in a marine environment where vessel motion, starting with the simpler SMR.
ease of operation, low equipment count, quick start-up,
process simplicity, and high availability are important (7). Single Mixed Refrigerant Process.
Early studies on the offshore LNG were conducted over 25 SMR processes have not received much attention for large-
years ago (8). N2 expander cycles for offshore liquefaction scale natural gas liquefaction because they are less efficient
were discussed and studied by Foster Wheeler and others in than their two-cycle (i.e. the C3MR) counterparts. The SMR
the 1980s (9) and recent studies have considered both process was selected because it required a wide phase-
expander-based as well as dual MR processes. Two processes envelope working fluid and thus is expected to show
that have been previously identified as offering potential for dramatically improved performance with a two-phase
offshore liquefaction are nitrogen expander cycles and dual expander. This process is seen below in Figure 1.
mixed refrigerant cycles (10). The C3MR cycle that has
dominated onshore applications is generally discounted
because of the large propane inventory and weight associated
with the propane precooling system. Nitrogen expander
processes attract attention because they offer the potential for
an extremely safe easy to operate liquefaction process that can
be effectively modularised and is indifferent to orientation.
These expander cycles generally suffer from low efficiency
OTC 18012 3

subcools the incoming high pressure, dry natural gas. The

JT Valve LNG
warmed, vaporized, LP MR then leaves the MCHE and returns
EXPANSION
OPTIONS
EXPANSION
to the first stage of the refrigerant compressor to complete the
MR Flash Gas
Liquid Expander 2-Phase Expander EXPANSION cycle.
The LNG leaving the MCHE is at high pressure and is
G G
LNG expanded to a pressure suitable for transfer and storage in an
end flash. This expansion occurs through one of the three
previously explained expansion configurations and then flows
to a separator. The liquid fraction (~90%) feeds the LNG
transfer pump that sends the product to storage. The vapour is
enriched in N2 and is warmed in a separate exchanger against a
portion of the feed gas before it joins the boil-off gas (BOG)
Feed Gas from storage and offloading operations and is compressed and
used as fuel gas. The heat recovery exchanger is not shown in
Figure 1. Single Mixed Refrigerant Liquefaction Process the schematic (Figure 2) for simplicity.
The process will briefly be explained starting with the
stream leaving the compressor suction scrubber. A warm, low Dual Mixed Refrigerant Process.
pressure MR stream consisting of nitrogen (N2), methane (C1), The dual MR process simulated is similar in concept with
ethane (C2), propane (C3), and normal butane (nC4) is the SMR and will not be explained in great detail. The dual
compressed in two stages. This compression requires an MR features a high temperature (HT) and low temperature
intercooler and aftercooler, shown pictorially as air coolers, to (LT) refrigeration circuit that each operates over a smaller
reject the heat from the liquefaction cycle to the environment. temperature range and thus requires a different refrigerant.
The high pressure MR is partially condensed in the aftercooler The HT circuit consists of a refrigerant including nC4, C3, C2,
before flowing into the spiral wound main cryogenic heat and C1 and operates at temperatures between the hot heat sink
exchanger (MCHE). In this exchanger this stream is and approximately -45oC. This circuit is identical to the SMR
continuously cooled and condensed tube-side against the cold, except that the exchanger has an additional pass that cools the
LP mixed refrigerant stream. Once condensed the cool high LT refrigerant. Most of the heat of the liquefaction cycle is
pressure MR is expanded in MR expansion. Three options rejected through the HT MR circuit. The HT MR circuit also
where considered for expanding the cool high pressure MR. continuously cools the process gas.
The MR could be expanded through a JT expansion valve,
the oldest, most robust, and cheapest expansion device that is JT Valve EXPANSION
LNG
EXPANSION
OPTIONS
simulated as an isenthalpic expansion such that the stream LT MR Flash Gas
Liquid Expander 2-Phase Expander EXPANSION
does work on itself and pressure availability is converted to
heat through turbulent dissipation. This offers the least G G Low LNG
efficient process. Temperature
MR Circuit
The second option considered is to isentropically expand
the liquid to the bubble curve. Liquid expanders are
commonly used on new build and many retrofit large scale
LNG facilities. Up to 2 MW of power is often recovered in
these turbines in electrical generators. The electrical power
recovery is a benefit but is small compared to the decrease in
compressor power input and liquefaction load. The extracted HT MR
EXPANSION
power is a reduction in load at very low temperature from a
liquefaction process with a finite efficiency. For instance, High
1 MW of power extracted from a LNG stream at -160oC may Temperature
MR Circuit
decrease refrigerant driver power by a further 5 MW.
Recovered power is limited by the degree of subcooling since
currently used liquid expanders cannot tolerate vapour
forming and care must be taken to avoid entering the vapour
envelope of the LNG. The expansion is finished in a JT valve
designed for flashing flow.
Feed Gas
The third expansion option considered is two-phase
expansion using a new liquid exducer type expansion engine. Figure 2. A Dual Mixed Refrigerant Process
These expanders can isentropically expand a liquid into the
vapour dome thus recovering the work for both the subcooled The LT circuit is precooled by the HT MR circuit and is
liquid expansion as well as the phase change. again similar to the SMR except that it rejects most of the
Returning to the process description, the cold expanded process heat to the HT MR refrigerant and cools, liquefies, and
MR returns to the MCHE as the cold stream and continuously subcools a precooled process gas.
cools the warm refrigerant stream and cools, condenses, and
4 OTC 18012

The dual MR features four compression stages, and three Table 2. Simulation Conditions and Assumptions
expansions. All the expansions were simulated with the three Equipment Value
options previously discussed. The LNG end-flash was Compressor efficiency 82% Adiabatic
simulated as described in the SMR section. Expander efficiency 90% Adiabatic
o
After coolers (discharge) 52 C / 0.7 bar pressure drop
Simulated Conditions. Suction piping 0.5 bar pressure drop
o
As in all complex systems, key liquefaction system MCHE minimum temperature approach of 2.5 C.
performance metrics are strong functions of feed gas, ambient 1% maximum vapour in liquid expander discharge
conditions, and equipment performance. To allow meaningful Neglect heat leak into the system
comparison of the processes, the same set of conditions and Pressures in both circuits kept within 600 # pipe class
constraints were applied to all processes. Electrical generator loaded expanders
Consistent / reasonable pressure profiles for all simulations

Natural Gas Feed and Product Conditions. Simulation Details Value

Details of the feed gas that were used for all simulations Mixture tuning resolution 0.5% molar composition
are seen below in Table 1. Since the simulation focused on Single MR composition N2, C1, C2, C3, nC4
the liquefaction process and the impact of two-phase Dual MR compostion C1, C2, C3, nC4 in HT circuit
expansion, the feed gas is free of water, acid gas components, N2, C1,C2, C3 in LT circuit
and has a tail consistent with a natural gas with the C3+ Simulation package HYSYS Plant v3.2
components recovered in an upstream NGL recovery plant. Equation of state Peng-Robinson (Hysys)
Table 1. Feed and Product Conditions Enthalphy model Equation of State (Hysys)
Vapour fraction in end-flash ~12%
Feed Gas LNG
Most of the constraints and assumptions listed in Table 2
Methane 0.92 0.93 are considered reasonable. Mixed refrigerant processes are
Ethane 0.07 0.06 challenging to simulate because of the multiple degrees of
Propane 0.01 0.01 freedom within the refrigerant composition, added to pressure
Butanes+ 0.00 0.00 and temperatures throughout the system. Optimisation tends
N2 0.00 0.00 to focus on the MCHE minimum temperature approach’s
CO2 0.00 0.00 response to refrigerant composition and flow rate. Mixture
composition is adjusted to avoid a pinch in the MCHE.
H2S 0.00 0.00 Reasonable system pressures were used through the process
H2O 0.00 0.00 considering pipe flange classes limits, volumetric flow in
suction lines, literature, and simulated performance. Care was
Pressure 63 bar 1 bar taken to ensure that the simulation optimisation methodology
o o was used across all six simulations.
Temperature 48 C -160 C
3 The detail present in the simulation was kept to a minimum
Flow rate 2985 Nm /hr ~2.1 MTPA to facilitate rapid evaluation of the six schemes. The same
Vapour quality 1 0 assumptions and basic simulation conditions were maintained
for all six processes to ensure that a meaningful comparison
The simulation’s objective target was to produce
can be made between them. The absolute process
approximately 2 MTPA LNG. This was accomplished in an
performance is a function of many variables including feed
end-flash that produced roughly 10-12% cold vapour for
gas conditions, ambient temperature, and system design. The
recompression. The temperature of the LNG entering the end-
process efficiencies for the dual mixed refrigerant process with
flash was adjusted to ensure that approximately 2.1 MTPA
liquid expansion can be roughly benchmarked against actual
LNG product flow rate was achieved. This meant that the
plant performance to verify the assumptions and degree of
LNG that was finished through a JT valve-based end-flash was
optimisation are reasonable.
cooled to a lower temperature by the liquefaction process.
The relative effects of the two phase and liquid expanders in
the mixed refrigerant circuit and the LNG stream were also Results
Shaft power from the compressor (input) and expanders
quantified and are discussed later in the paper.
(output) are shown as positive numbers in Table 3.
Compressor power is the shaft power required to drive the
Standard Simulation Assumptions.
compressor and is the sum of all energy streams (two
Care was taken to ensure that the comparison was
compressor stages for the single MR and four for the dual
completed on a reasonable and consistent basis. All major
MR). Expander power sum is the shaft power extracted from
constraints and assumptions were kept consistent across the
the process stream in each MR circuit and the LNG end flash.
six simulated processes as seen in Table 2.
Total power is the compressor power required minus the
expander power recovered.
OTC 18012 5

Table 3. Simulation Results driver capacity. Offshore however, the two-phase expanders
may facilitate the development of less efficient single-cycle
Single MR
JT Valve Liquid Only 2-Phase
processes that are more compact, easier to operate, safer, and
Compressor Power (kW) 108735 88413 84447 more reliable.
Expander Power (kW) 0 2712 6279 Two-phase expanders facilitate the use of a refrigerant
with a wider phase envelope and a larger, more reversible,
Total Power (kW) 108735 85701 78167 temperature drop across the expansion device(s). The process
efficiencies of these single cycle processes is approaching the
Efficiency (kW/TPD) 18.8 14.8 13.5
Added Power Recovered (kW) 2712 3567
JT valve and liquid-expander only process efficiencies of the
dual cycle process. Although a single-cycle process has not
Relative Efficiency 160% 126% 115% been designed for an onshore baseload facility for over 30
Process Improvement 8.8% years, they may be acceptable offshore because of improved
performance and different process selection criteria.
Dual MR Generally speaking, onshore process efficiency and very large
JT Valve Liquid Only 2-Phase
train capacity is more important than process simplicity,
Compressor Power (kW) 81501 76402 75138
Expander Power (kW) 0 2094 7061
footprint, and flammable fluid inventories. The efficiency of
the two-phase expander-based SMR process between that of a
Total Power (kW) 81501 74308 68077 traditional C3MR and a C3MR with liquid expansion is
considered very good considering the increased process
Efficiency (kW/TPD) 14.1 12.8 11.7 simplicity, decreased equipment count, and reduced
Added Power Recovered (kW) 1256 4639
flammable refrigerant inventories this process could offer.
Relative Efficiency 120% 109% 100% The absolute process efficiency for the dual MR is
Process Improvement 8.5% reasonable compared to published baseload process
efficiencies. Onshore baseload liquefaction facility efficiency
values range widely and are not reported on a consistent basis.
LNG liquefaction process efficiency may be expressed as A reasonable range for MR based processes is
thermal efficiency, specific power, or figure of merit. Table 3 11.5-13 kW/TPD of LNG produced (13). The dual MR with
expresses the thermodynamic efficiency as specific power, the liquid expansion efficiency result of 12.8 kW/TPD of LNG is
power consumed per unit of LNG production, in units of approximately 5% less efficient than a published value for the
kW/tonnes per day (TPD) of LNG produced. This metric, Oman LNG C3MR Train 1 of 12.2 kW/TPD (14). This plant
along with some other simulation outputs are also included in (EPC phase completed by Foster Wheeler and Chiyoda)
the table. Added power recovered is the power extracted by featured liquid expansion as well as other design features such
the expander relative to the process with less expansion. For as axial compression boosting efficiency.
example, the Added power recovered of a single MR with The simulation absolute process efficiencies benchmark
liquid expansion is 2.7 MW more than the single MR with JT well against published values but must be used with caution.
valves only due to the expander electricity generation. They are best used to assess the potential of two-phase
The most efficient process was the dual MR process expander-related process improvements and to gain insight
deploying the two-phase expander with a total power into which conditions two-phase expansion will be most
consumption of approximately 68 MW for the simulated beneficial. Process performance is a strong function of
2.1 MTPA production rate. The other process efficiencies are simulation assumptions, pressure profile, equipment
reported relative to this process and range up to 60% less performance, ambient temperatures, feed gas, and other
efficient for the single MR using only JT expansion valves. variables.
The single MR processes are less efficient than the dual MR
processes using the same expansion devices. The efficiency of Conclusions
the single MR process using the 2-phase expander is roughly The analysis supports three basic conclusions relating to two-
between that of a dual MR and dual MR with liquid expander phase expander use in offshore liquefaction processes:
processes.
The use of two phase expanders in either the SMR or dual 1. When proven and optimally designed into liquefaction
MR process result in a power saving of approximately 8.5% processes, the two-phase expander process should offer
over the liquid-only expander. significant efficiency improvements. Further design and
study work is required to quantity the performance of
Discussion two-phase expander processes but the current simulation
The efficiencies of the single MR processes using either liquid results indicate a ~8.5% efficiency improvement relative
expanders or possibly two-phase expanders are considerably to liquid expanders.
higher than the cycle using only JT valve expansion. With an
8.5-8.8% improvement in liquefaction process efficiency, 2. The performance of the single MR processes using either
liquid expanders should be considered, further developed, and liquid or two-phase expanders is sufficiently high that
eventually work their way into baseload liquefaction facilities. they will likely be suitable in some offshore applications
Onshore, this will allow increase process efficiency and a given their simplicity, size, and safety advantages.
resultant increase in LNG production for a given compressor-
6 OTC 18012

3. Processes using two-phase expanders should be carefully

optimised to ensure they fully utilise the potential offered
by these newly available machines. These expanders
maximise the utilisation of large phase envelope
refrigerants in simple cycles.

Future work should consider in more detail the relative

contribution of the two-phase expander in the LNG stream and
the refrigerant cycle. This is strongly dependent on the N2
content in the feed gas and the vapour fraction downstream of
the end-flash. Additional study work should focus on the
optimisation of a two-phase expander based process with
possible multiple expansions.

References

1 Shukri, T.: “LNG Technology Selection”, Hydrocarbon

Engineering, (Feb. 2004).
2 DOE/EIA, The Global Liquefied Natural Gas Market: Status and
Outlook, Energy Information Administration, U.S. Dept. of
Energy, DOE/EIA-0637, Dec. 2003.
3 DOE/EIA, Annual Energy Outlook – 2004, Energy Information
Administration, U.S. Dept. of Energy, DOE/EIA-0383(2004),
Jan. 2004.
4 Meyer M.: “LNG Liquefaction Process- Why the Big Fuss about
Selection”, IChemE London SONG Meeting, Nov. 9, 2004.
5 LNG Remains Keystone to Development of Qatari Gas Resources,
LNG Express, Vol. 15, No. 1., Jan 2005
6 Madison, J., Kimmel, h.: “LNG Expander for Extended Operating
Range in Large-Scale Liquefaction Trains”, 5th Topical
Conference on Natural Gas Utilization, AIChE Spring Meeting,
April 2005.
7 Waldie, b.: “Effects of Tilt and Motion on LNG and GTL Process
Equipment for Floating Production”, GPA Europe Annual
Conference, Rome, Italy, Sept. 2002.
8 Kennett, A., Limb, D., Czarnecki, C.: “Offshore Liquefaction of
Associated Gas- a suitable process for the North Sea”, Offshore
Technology Conference, May, 1981.
9 Finn, A.J., Johnson, G.L., Tomlinson, T.R.: “LNG technology for
offshore and mid-scale plants”, 79th Annual GPA Convention,
Atlanta, March 2000.
10 Sheffield, J.A., Mayer, M.: “The Challenges of Floating LNG
Facilities”, Proc. of the GPA Spring Meeting 2001.
11 Barclay, M.A., Denton, N.: “Selecting Offshore LNG Processes”,
LNG Journal, pp. 34-36, Oct. 2005.
12 Faber, F., Resweber, L.R., Jones, P.S., Bliault, A.E.J.: “Floating
LNG Solutions from the Drawing Board to Reality”, Offshore
Technology Conference, Houston, Texas, May 6-9, 2002.
13 Dam, W., Siew-Mung Ho: “Engineering Design Challenges for
the Sakhalin LNG Project”, GPSA Conference San Antonio,
Texas, March 2001.
14 McLachlan, G., Ayres, C., Vink, K., Al Mukhainy M.: “Efficient
Operation of LNG from the Oman LNG Project”, Gastech 2002,
Doha, Qatar, Oct. 2002.