July 2018

Bonus Report: LNG Technology

Liquefaction technology selection for baseload LNG plants

As global liquefied natural gas (LNG) trade continues to rapidly expand, the challenge of liquefaction process selection—a key element of an LNG project—becomes increasingly important.

Mokhatab, S., Consultant; Messersmith, D., Bechtel Corp.

As global liquefied natural gas (LNG) trade continues to rapidly expand, the challenge of liquefaction process selection—a key element of an LNG project—becomes increasingly important. Selecting more versatile and cost-effective liquefaction technologies that meet stringent environmental emissions standards is a key focus for new projects. 

In recent years, liquefaction process selection studies for baseload LNG plants have grown in scope due to the interest in both larger-train capacities and the potential economies of repeatability, with some large-scale facilities using multiple small-train solutions. As such, this article presents an overview of an LNG production plant, along with an analysis of the most common and proven processes available for the liquefaction of natural gas in baseload LNG plants. 

LNG plant overview.

In a typical LNG production plant, field production fluids are first separated in an inlet reception facility that removes the hydrocarbon liquids (condensate) and aqueous phase and routes the raw gas to the processing circuit. Subsequently, the gas is pretreated to remove any impurities that interfere with the liquefaction process or are undesirable in the final products. The specifications required to be met are hydrogen sulfide removal to under 4 ppmv, carbon dioxide to 50 ppmv, total sulfur under 30 ppmv, water to 0.1 ppmv and mercury to levels of 0.01 mg/Nm3.2 

Natural gas liquids (NGL) are also removed from the dry sweet natural gas using refrigeration to meet BTU requirements of the LNG product and to prevent freezing and subsequent plugging of equipment in the liquefaction unit.

FIG. 1. Traditional LNG production scheme.

The remaining gas, which is made up mainly of methane and contains less than 0.1 mol% of pentane and heavier hydrocarbons, is further cooled in the cryogenic section to approximately –160°C (–256°F) and is completely liquefied (Fig. 1). For feed gas containing nitrogen (N2) levels greater than 1 mol%, N2 will be removed by additional processing in an LNG production plant to help prevent rollover during transport. The specification can be met by adjusting the amount of end flash gas that is produced.

Natural gas liquefaction plants can be classified into large baseload, mid-scale, peakshaving and small-scale plants, depending on their size and function (Table 1). Baseload plants, which typically consist of one or multiple trains, supply natural gas as LNG to consumer nations by ocean transport. The liquefaction train size for baseload plants has steadily increased over the last 40 yr, with capacities of more than 4 MMtpy now being conventional. Single trains of 7.8 MMtpy are in operation in Qatar. 

Large individual LNG trains tend to decrease the overall unit cost of production, helping make both greenfield and brownfield projects more competitive in the marketplace. To fully realize the cost advantages of these large trains, high plant reliability and availability must be maintained; furthermore, the design must be robust and should mirror the equipment count of smaller trains to ensure that economies of scale are realized. 

Natural gas liquefaction.

Liquefaction technology is based on refrigeration cycles, which take warm, pretreated feed gas and cool it through cryogenic heat exchangers into a liquid product. To generate the cold temperatures required for LNG production, work must be put into the refrigeration cycle through compression, and heat must be rejected from the cycle to the environment through air or water coolers. The basic principle of using refrigerants for cooling and liquefying the gas is to match as closely as possible the cooling/heating curves of the feed gas and the refrigerant, which results in a more efficient liquefaction process requiring a lower power consumption per unit of LNG produced.1 Examples of cooling and heating curves for natural gas and refrigerants are shown in Fig. 2. 

FIG. 2. Typical natural gas/refrigerant cooling curves.

Observing the cooling curve of a typical gas liquefaction process (red line), three zones (precooling followed by liquefaction and finishing with subcooling) can be noted. All of these zones are characterized by having different curve slopes, or specific heats, along the process. All LNG processes are designed to closely approach the cooling curve of the gas being liquefied. This is achieved by using mixed (multi-component) refrigerants, a combination of multi-component and single-component refrigerant cycles, or multiple single refrigerants that will match the cooling curve in the different zones/stages of the liquefaction process to achieve high refrigeration efficiency and minimize energy consumption, while keeping the number of refrigeration stages at a reasonable number.

A number of liquefaction processes have been developed, with the main differences seen in the type of refrigeration cycles used. These processes can be broadly classified into two groups: mixed refrigerant processes and cascade liquefaction processes (using pure components as refrigerants). Expansion-based processes are not considered here, as these are typically used for only small-scale LNG, peakshaving and floating LNG (FLNG) applications in the 1 MMtpy–2 MMtpy range. In Table 2, the merits of expansion-based processes are shown compared to other schemes. The lower efficiencies of expansion-based processes generally make them non-viable for larger baseload applications. 

The classical cascade process reduces irreversible heat exchange losses by utilizing several refrigeration cycles having refrigerants that vaporize at different but constant temperatures. The cascade cycle is flexible in operation, since each refrigerant circuit can be controlled separately; this avoids the need to chase the feed composition with refrigerant composition or risk the curves touching. However, a disadvantages of the cascade technology is the relatively high capital investments related to equipment count. Note: The relationship between equipment count and total investment cost is not always direct. Size and equipment complexity can warp the comparison.

The mixed refrigerant (MR) cycles involve the continuous cooling of a natural gas stream using a carefully selected blend of refrigerants (usually a mixture of light hydrocarbons and N2) that can mimic the cooling curve of natural gas from ambient to cryogenic temperatures. MR technology has been assessed for onshore liquefaction based on both single mixed refrigerant (SMR) and dual-mixed refrigerant (DMR) cycles. The SMR process provides the benefit of operational simplicity and flexibility, in addition to reduced equipment count; however, it comes at the cost of lower efficiency than the DMR cycle, which better matches the overall mixed refrigerant boiling curve to the feed condensation curve. 

Numerous versions of the dual-stage cooling cycles with one or both cycles involving mixed refrigerants have been developed. The propane precooled mixed refrigerant (C3MR) cycle, which is the most widely used liquefaction technology, results in a more efficient plant design and uses less power as compared to SMR and expander-based systems. The downsides of this modification are higher process complexity and higher processing equipment counts, as compared to simpler configurations.

Technologies for baseload applications.

Most baseload LNG plants have two or three refrigeration loops to cool and liquefy the natural gas stream over a wide temperature range. It is realized that for each of these processes, several options exist in the configuration of the process that will influence the capacity and overall attractiveness. These variations within a process result from the particular gas turbine drivers and cryogenic heat exchangers.2

The most well-known/proven gas liquefaction technologies developed for baseload liquefaction plants are described in the following sections. The descriptions do not attempt to disclose the full details of the licensed processes, but rather explain the basic design concepts and design considerations.

Propane precooled mixed refrigerant (C3MR) process.

This liquefaction cycle, which was developed by Air Products and Chemicals Inc. (APCI), is composed of a multistage propane (C3) precooling system followed by liquefaction using an MR system of nitrogen, methane, ethane and propane. The heart of the APCI C3MR process, which features thermal efficiencies of more than 93%, is its proprietary spiral-wound heat exchanger (SWHE). 

Large-capacity trains of more than 5 MMtpy can be designed using a parallel refrigerant compressor/driver arrangement. This arrangement is discussed in the previous section on driver configuration. Although this concept increases equipment count, it also decreases the size of equipment and can, therefore, stimulate competition between equipment vendors.3 In addition, availability benefits exist with parallel compressor streams, and LNG plants can now achieve availabilities of up to 95% using parallel compressor arrangements.

FIG. 3. Train capacity for different baseload liquefaction technologies.

Recent designs can reach up to 7 MMtpy using a single main cryogenic heat exchanger (MCHE), with a parallel refrigerant compressor configuration. In the 2000s, APCI developed an adaptation of the C3MR process to increase the size of a single train to more than 7 MMtpy (Fig. 3). The AP-X process adds a third refrigerant cycle (N2 expander) to provide the LNG subcooling duties subsequent to the SWHE. With this N2 cycle, the size of the SWHE is maintained, with the subcooling duty shared by the N2 cycle that allows lower propane and MR flowrates compared to the C3MR process. This design approach makes liquefaction of 10 MMtpy possible without the development of a larger main heat exchanger. The AP-X process is claimed to achieve high efficiency and low production cost by using all three refrigerant cycles to their best advantage.4

No new AP-X trains have been proposed since the first six AP-X trains were installed in Qatar. The economies of scale are not as transparent for the AP-X process as they are for the large AP-C3MR trains, with ongoing improvements to the size of the SWHE.

Optimized Cascade process.

The Optimized Cascade process, solely offered by ConocoPhillips, uses multiple stages of propane, ethylene and methane refrigeration loops to balance refrigeration loads. This process has been designed around a “two-train-in-one” concept to improve reliability. The interesting feature is having parallel lines of compression, with a Frame 5 variable-speed gas turbine. This yields high availability and easier operation; no compressor trip will completely shut down the unit, and the restart of the compressor can be accomplished without loss of refrigerant.5

This process has been successfully proven with decades of plant operation in Kenai, Alaska. Since Atlantic LNG in the late 1990s, the ConocoPhillips process has made successful inroads to the near-monopoly APCI had on baseload liquefaction trains. Trains built to date with this technology are just above 5 MMtpy, and the licensor claims that larger train sizes up to 6.5 MMtpy are possible. Although the reported overall thermal efficiency for the Atlantic LNG plant utilizing this process is approximately 89%,6 recent units are able to operate with higher thermal efficiency exceeding 93%.7

Mixed fluid cascade (MFC) process.

This technology, developed by the Linde/Statoil (now Equinor) technology alliance, is a classic cascade process that precools, liquefies and subcools natural gas by means of three separate MR cycles. Compared to the cascade, the efficiency is higher, as MRs allow a closer temperature approach. However, the power is not the same on all three cycles, unlike with the cascade process. Plate-fin exchangers are used on the first cycle, and coil-wound exchangers are used on the two colder cycles.

This process was pioneered at the Snøhvit LNG terminal on Melkoya Island offshore Hammerfest in the Northern North Sea of Norway. This plant, with a capacity of 4.3 MMtpy, remains Europe’s only baseload export gas liquefaction plant, and the only MFC plant in operation. The plant experienced a challenging startup and first few years of operation. However, a rectification program was put in place, and the issues with Snøhvit have been resolved. The experience gained from implementing the first MFC plant has been translated into proposed designs.8 Linde is now proposing capacities up to 10 MMtpy with this design.

Dual mixed refrigerant (DMR) process.

Dual MR processes are offered by Shell and APCI. The Shell process has been proven in Sakhalin, while the APCI process has been qualified by most major international oil companies. This process uses two separate MR cooling cycles—one for precooling the gas and one for final cooling and liquefaction. The technology, which features a thermal efficiency of more than 93%, was deployed for the first time at baseload scale at the Sakhalin liquefaction plant in eastern Russia. The Sakhalin plant has two 4.8-MMtpy trains that use SWHEs and air cooling enhanced by the cold climate.9

The DMR process configuration is similar to the C3MR process, but with the precooling conducted by an MR (made up of mainly ethane and propane) in an additional SWHE, rather than pure propane in a shell-and-tube exchanger. Using MR with a lower molecular weight on the first cycle allows for a smaller condenser, and also removes the propane compressor bottleneck.5 Even with the use of two MR cycles, the DMR process is very similar in efficiency to the C3MR process when used in tropical climates. The advantages of the DMR process are demonstrated when applied to cold climates, since the precooling MR can be formulated to avoid the pressure limitations associated with propane at colder temperatures. 

Liquefaction process selection.

Selection of an appropriate liquefaction technology for an LNG production plant must be based on technical, economic, commercial and environmental considerations. Other evaluation criteria include technology maturity and plant constructability, operability and maintainability. Technical considerations include process and equipment experience, reliability, process efficiency, turndown, site-specific requirements and environmental impact. Equipment availability and risk factors must also be taken into account. Economic issues include capital, operating and lifecycle costs. All of these aspects must be evaluated to arrive at the optimum solution.1

A “like-for-like” comparison of the different licensed liquefaction processes is difficult to achieve because the detailed technical content is not available in the public domain. The APCI C3MR process is widely used and is accepted as being one of the most cost-effective and reliable baseload LNG processes available. With more than four decades of operating experience and incremental capacity increases up to 6 MMtpy (and up to 7.8 MMtpy using AP-X Technology), the APCI C3MR process is often the first choice for large baseload LNG plants using air cooling in a tropical climate.

The ConocoPhillips Optimized Cascade Process has made successful inroads into the LNG market since the advent of the Atlantic LNG trains in Trinidad in the late 1990s. Locations using this technology include Australia, Angola, Egypt and Equatorial Guinea. The costs and power requirements for the Optimized Cascade process are on par with the C3-MR process. The selection of the main refrigeration drivers has a larger incremental impact on overall plant efficiency than the process selection itself.10,11

The Shell DMR process, which aims to maximize flexibility in harsh environments, is generally selected on the basis of being able to adjust the composition of the precooling cycle to fully utilize the cold available throughout the year. 

Liquefaction process selection is a key activity that starts at an early stage of an LNG project. It should be addressed at the conceptual, feasibility and pre-FEED stages of development, since it has such a large impact on the overall profitability of the project. When a comparison is conducted thoroughly, sufficient process and utility details must be developed to define the capital and operating costs for each licensor. Quotations from the various licensors and main equipment suppliers must be obtained to highlight the differences in the processes, and finally to select an optimized design that will best meet the LNG project owner’s objectives.

A recent project development strategy has been to review the use of multiple small-scale trains for larger export facilities. Many of these project opportunities are located on the US Gulf Coast. The use of multiple small trains provides proponents with the opportunity to scale the facility, to build capacity incrementally to meet LNG demand or to manage capital financing. 

In addition, one of the key drivers for these new projects is the ability to operate across a wide range of capacities to meet swings in market demand without adverse impact on the efficiency of the process. The manufacture of smaller modules and the potential benefit of replication and duplication make the project’s execution model very attractive in certain areas. The ability of small-scale manufacturers to scale their product to support large facilities with multiple trains has yet to be proven over traditional economies of scale. However, it is highly attractive in this era to develop new project opportunities at competitive costs.


Considerable diversification of liquefaction processes has been seen in the last two decades. This increased competition toward cost per capacity has led to increased train capacity, which can result in decreased unit costs, depending on the site location. 

A review of the different baseload liquefaction processes dominating the global liquefaction market suggests that no single process is substantially more efficient than the others in a given situation. Rather, each technology can be competitive within a certain range of train sizes and conditions. The ultimate choice of which process to select will remain dependent on project-specific variables and the development state of novel processes. Maximizing the value for an LNG venture by selecting the optimal configuration to best suit operating and market conditions can only be achieved after a detailed study of all options.

Under present and near-term conditions, innovation and flexibility are key to realizing new opportunities with new suppliers, markets and technologies. Numerous options are now available for owners to develop resources at a rate that suits the market. Project owners and EPC contracting teams must share their experience from previous baseload liquefaction plants, from concept evaluation through startup, and from recent projects using small-scale technologies to ensure that the best solution can be selected with regard to cost, site specifications and constructability. HP


Thanks are due to Isa Mohammed and Philip Hunter of KBR (UK) and Scott Northrop of ExxonMobil (US) for reviewing this manuscript and providing constructive comments and suggestions.

Literature cited

  1. Shukri, T., “LNG Technology Selection,” Hydrocarbon Engineering, Vol. 9, Iss. 2, 2004.
  2.  Vink, K. J. and R. K. Nagelvoort, “Comparison of baseload liquefaction processes,” LNG-12 Conference and Exhibition, Perth, Australia, May 4–7, 1998.
  3. Finn, A. J., G. L. Johnson and T. R. Tomlinson, “LNG technology for offshore and mid-scale plants,” 79th
    Annual GPA Convention, Atlanta, Georgia, March 13–15, 2000.
  4. Roberts, M. J., Y. N. Liu, J. M. Petrowski and J. C. Bronfenbrenner, “Large-capacity LNG process—the AP-X cycle,” Gastech 2002 Conference and Exhibition, Doha, Qatar, October 13–16, 2002.
  5. Martin, P.-Y., J. Pigourier and B. Fischer, “Natural gas liquefaction processes comparison,” LNG-14 Conference and Exhibition, Doha, Qatar, March 21–24, 2004.
  6. Richardson, F. W., P. Hunter, T. Diocee and J. Fisher, “Passing the baton cleanly—Commissioning and startup of the Atlantic LNG project in Trinidad,” Gastech 2000 Conference and Exhibition, Houston, Texas, November 14–17, 2000.
  7. Ransbarger, W., “A fresh look at LNG process efficiency,” LNG Industry, Spring 2007.
  8. Vist, S., et al., “Startup experiences from Hammerfest LNG—A frontier project in the North of Europe,” LNG-16 Conference and Exhibition, Oran, Algeria, April 18–21, 2010.
  9. Dam, W. and S.-M. Ho, “Engineering design challenges for the Sakhalin LNG project,” 80th Annual GPA Convention, San Antonio, Texas, March 12–14, 2001.
  10. Meher-Homji, C., D. Messersmith, T. Hattenbach, J. Rockwell, H. Weyermann and K. Masani, “Aeroderivative gas turbines for LNG liquefaction plants—Part 1: The importance of thermal efficiency,” and “Aeroderivative gas turbines for LNG liquefaction plants—Part 2: World’s first application and operating experience,” ASME Turbo Expo 2008 Conference, Berlin, Germany, June 9–13, 2008.
  11.  Meher-Homji, C., D. Messersmith, K. Masani and H. Weyermann, “The application of aeroderivative engines for LNG liquefaction—Higher plant thermal efficiency, lower CO2 footprint, and modularization capability,” Gastech 2009 Conference and Exhibition, Abu Dhabi, UAE, May 25–28, 2009.

The Authors

Related Articles

From the Archive