Alaska North Slope LNG project considers various CO2 removal processes

June 3, 2002
Several technologies are available for treating Prudhoe Bay gas for LNG production. Considering the huge capital investment involved with such projects, Parsons Energy & Chemicals Inc. conducted a comprehensive process selection study of new and conventional technologies.

Several technologies are available for treating Prudhoe Bay gas for LNG production. Considering the huge capital investment involved with such projects, Parsons Energy & Chemicals Inc. conducted a comprehensive process selection study of new and conventional technologies.

BP Exploration (Alaska) Inc., Foothills Pipe Lines Ltd., Phillips Alaska Inc., and Marubeni Corp. commissioned the study and, as a group, sponsored the Alaska North Slope LNG Project.

Various treating processes will remove bulk CO2, but operators often prefer proven technologies.

While cryogenic distillation and membrane technologies are commercially proven for acid-gas removal, this study shows they do not yet threaten the position of solvent-based processes for base-load LNG applications.

Newer technologies, such as gas-absorption membranes, may hold future promise for cost-effective bulk acid-gas removal.

The wet-absorption process, however, is still the most economical choice for large-scale gas processing.

This will change as newer technologies achieve breakthroughs and advance their competitive positions.

This article provides an overview of acid-gas treating processes evaluated for the ANS LNG project.

Background

The ANS generates large volumes of natural gas as a by-product of oil recovery. About 8 bcf of gas is reinjected every day, an amount exceeding Jap an's daily gas consumption. Recoverable natural gas resources at Prudhoe Bay are around 30 tcf, the energy equivalent of more than 3.8 billion bbl of oil.

In 2001, the ANS LNG project was advancing toward commercial reality through the efforts of the sponsor group companies, which integrated different complementary attributes, including:

  • Gas resource ownership at Prudhoe Bay and other ANS fields.
  • Experience with LNG marketing, transportation, and operations.
  • Expertise in Alaskan operations and permitting issues.
  • Involvement in other synergistic ANS gas monetization options.

The group, formed in 1998, was working toward an internal target of first LNG deliveries as early as 2008, if the market is ready.

The group concluded conceptual and feasibility engineering activities to refine and reduce the project's scope, introduce innovation, and reduce the cost of major facilities.

Click here to enlarge image

Facilities include a gas treatment plant (GTP) located on the ANS, an 800-mile pipeline, an LNG plant (LNGP) and associated marine terminal, and a fleet of LNG tankers. The group had been considering two LNG plant sites and corresponding pipeline routes-one near Nikiski on the Cook Inlet, and the other at Anderson Bay in Valdez Harbor (Fig. 1).

Click here to enlarge image

The GTP for the Prudhoe Bay project would consist of gas-conditioning pro cesses for removing CO2, water, benzene, and heavy hydrocarbons. The treated LNG-quality gas would be pressurized to 2,800 psig, chilled to 30° F., and transported via a buried pipeline routed alongside the Trans-Alaska Pipeline.

Screening studies

Click here to enlarge image

Fig. 2 shows a block diagram of the gas pretreatment scheme. Table 1 shows the feed gas composition and treated-gas specifications. The GTP consists of facilities for CO2, benzene, and heavy hydrocarbon removal; sales-gas compression and chilling; CO2 compression; and dehydration and re-injection.

Various studies helped establish the best process for each function: BASF AG activated MDEA (aMDEA) for CO2 removal; compound-bed molecular sieves for dehydration, benzene, and heavy hydrocarbon removal; and triethylene glycol (TEG) for CO2 dehydration.

This article focuses on CO2-removal technologies.

Parsons performed coarse screening studies of the following technologies:

  • Membranes.
  • Morphysorb.
  • Selexol.
  • Ifpexol.
  • Ryan-Holmes.
  • aMDEA.

Gas-separation membranes

During the past 2 decades, gas separation using membranes has become an important technology. Polymeric membranes made from rigid glassy polymers remove CO2 from natural gas using a spiral wound configuration of membrane sheets.1 2

The more-permeable CO2 gas passes rapidly through the membrane into the permeate spacer and concentrates there as a low-pressure gas stream. These single-stage membrane systems, however, suffer from high hydrocarbon losses; a multistage1 system is often used to reduce losses.

Multistage schemes compress and process the permeate stream in a second membrane stage. Recovered methane recycles to the first stage for further purification, and CO2 is removed as second-stage permeate. Although operators have applied this approach to small-scale operations, it is not economical for processing large volumes of gas.

One significant disadvantage of membranes in this application is a penalty associated with CO2 and permeate reinjection. The large permeate flow incurs a significant cost penalty for CO2 recompression. This drastically increases costs for the membrane option.

While membrane technology for CO2 and hydrocarbon removal is already commercialized, it still needs technological developments to reduce membrane area and make large-scale commercial application feasible. One difficulty is that membranes do not inherently offer the same economy of scale as other processes due to their modular nature.

For example, the cost of a solvent-based process has a capacity ratio exponent of 0.65, whereas a membrane system has an exponent of around 1.0. Membrane technology also needs further developments to improve CH4and CO2 selectivity to reduce CH4losses.

Morphysorb

A new process developed by Krupp Uhde GMBH in cooperation with the Institute of Gas Technology,3 Morphysorb uses a physical solvent of morpholine derivatives, N-formyl-morpholine (NFM) and N-acetyl-morpholine (NAM), and can meet feed-gas specifications for LNG.

Combining two physical solvents of similar chemical structure allows one to optimize the special advantages of a physical solvent (i.e., high gas loadings at high partial pressures and low temperatures). The absorption capacity of acid gas improves considerably.

Compared to other physical solvents, the Morphysorb solvent coabsorbs fewer heavier hydrocarbons and is also suited for simultaneous water removal from feed gas. Parsons decided to evaluate this new technology because of the unique advantages, even though it has no commercial applications.

Although the Morphysorb circulation rate was about 20% less vs. the Selexol process, it was still more than twice that of the BASF aMDEA process. Even at a higher absorber pressure (1,500 psig) the circulation was 28% higher than aMDEA.

Selexol

A pure physical solvent process, Selexol is a mixture of dimethyl ethers of polyethylene glycol, capable of removing H2S, CO2, and organic sulfur compounds. A potential advantage of a physical solvent such as Selexol for this application is that it removes CO2, sulfur compounds, benzene, and heavy hydrocarbons in one step.

The solvent will also dehydrate the treated gas to a low level, which significantly reduces downstream dehydration load. A preliminary evaluation, however, showed that the required circulation rate was about three times as much as BASF's aMDEA process.

Parsons evaluated acid-gas partial pressure sensitivity by increasing absorber pressure to 1,500 psig and reducing circulation rate. Although the increased CO2 partial pressure led to a 30% reduction in circulation rate, it was still about twice as much as the aMDEA process.

Ifpexol

Introduced in 1991, Ifpexol uses methanol as a physical solvent.4 Two processes are used independently or in combination: IFPEX-1 for removing condensable hydrocarbons and water and IFPEX-2 for acid-gas removal.

IFPEX-2 removes CO2 in a gas-liquid contacting column. The dry, dew-point-adjusted acid gas passes up a column where it is contacted with a cold descending methanol stream. External refrigeration improves acid gas absorption and reduces methanol losses.

Multiflash regeneration with or without thermal stripping desorbs acid gas from the methanol, which is recovered for recycle to the IFPEX-2 contactor. Regenerated methanol yields a dry, 200-psig CO2 stream. Disadvantages are methanol losses and contaminated product streams, inherent in methanol injection systems in general.

IFPEX-2 reduces CO2 to 1,000 ppm; this limitation imposes a need for a second treating process (e.g., solvent based) downstream to meet the 50-ppmv CO2 specification.

The short list

Parsons short-listed the Ryan-Holmes5 and BASF aMDEA processes for further evaluation. We prepared detailed heat and material balances with utility summaries for each technology. We sized major process equipment and estimated total capital cost based on Parsons' in-house database and extensive North Slope experience. Vendors provided budget prices for packages and specialized equipment. We factored the major equipment to get the total installed cost.

Ryan-Holmes

The Ryan-Holmes process is a cryogenic distillation process that can save significant capital and operating costs for this application. The process eliminates the need for circulating solvent systems and it can achieve low hydrocarbon dew points.

Other advantages are that CO2 is removed at high pressure, reducing reinjection compression horsepower, and CO2 dehydration is not required. Typically, this process involves two key separations:

  • Ethane from CO2 uses an alkane to suppress CO2 freezing.
  • CO2 from ethane uses the alkane to break the CO2-ethane azeotrope.

The proposed Ryan-Holmes scheme is a two-column system, consisting of a demethanizer followed by an ethane-recovery column. An ethane-propane cascade refrigeration system is used for the cryogenic columns.

Demethanizer overhead is 50-ppmv CO2 product sent to sales-gas compressors and the bottoms are fed to the ethane recovery column. Ethane recovery column overhead vapor is sent to CO2 low-pressure (LP) compressors, and the bottoms are recycled as the alkane additive to the demethanizer's top section.

Because the Ryan-Holmes process requires substantial refrigeration, extensive heat integration is needed to reduce the refrigeration load. This increases flow scheme complexity and adds more equipment. Normally the refrigeration requirement can integrate with the main liquefaction cycle of the LNG plant. It was not an option for this project, however, due to the physical separation of the GTP and the LNG plants. We can integrate the refrigeration duty with sales gas chilling duty in this case.

Activated MDEA

The aMDEA process is well suited for CO2 removal for the specified feed-gas conditions. The tertiary amine has a high acid-gas loading capacity and, due to its quasi-physical absorption behavior, a low regeneration energy requirement.

The solvent includes small amounts of an activator, which drastically increases CO2 absorption kinetics. Activator content can vary to suit the application and provide a customized process design.

Parsons evaluated several process configurations including multistage flash regeneration and lean, semilean, and well-stripped schemes.6 We found that the well-stripped scheme with a single-stage flash was the optimum scheme for meeting the 50-ppm CO2 specification.

Economics

Since absolute performance data for each technology is proprietary to the respective licensors, we cannot publish such data without violating secrecy agreements. We can, however, use public domain information to indicate relative performance to illustrate how some processes differ in their inherent characteristics and performance.

Click here to enlarge image

Table 2 shows an economic comparison of Ryan-Holmes and aMDEA processes. We compared capital cost (±25%) and operating cost. Capital cost estimate is based on equipment prices obtained from vendors and Parsons' in-house database, as required. We adjusted cost estimates for modular North Slope installation.

Economic comparisons are limited to acid-gas removal, CO2 LP compression and dehydration, and feed gas dehydration only, because the other unit operations are common to both technologies.

The Ryan-Holmes process requires lower CO2 compression horsepower because CO2 is removed at a higher pressure (335 psig) compared with aMDEA (15 psig). The Ryan-Holmes process also eliminates the need for CO2 dehydration.

Ryan-Holmes requires a smaller feed-gas dehydration mole sieve unit due to upstream TEG dehydration. (Water content of treated gas from TEG dehydration is 150 ppm vs. 700 ppm from the aMDEA process.) Nevertheless, Ryan-Holmes' overall capital cost is 54% more than aMDEA, primarily due to more equipment and a large refrigeration system.

Operating costs include power and fuel consumption for each process. Power consumption consists of pumps and air coolers, and fuel-gas consumption consists of gas turbines, reboilers, and mole sieve regeneration heaters.

The aMDEA power consumption is higher, partially because the circulation pumps did not use hydraulic power recovery, which represents about half of the total power requirements. Fuel-gas consumption is higher for the Ryan-Holmes process, mainly due to the large gas-turbine-driven refrigeration compressors.

GPA sponsored research

This project's process design benefited from GPA research programs. Areas of the facility that benefited most were acid-gas removal and dehydration. We used hydrocarbon solubility data for various generic amines to verify commercial simulator results. We were interested in hydrocarbon solubility because it affects the sales gas' higher heating value and is used as an indicator of individual solvent performance related to hydrocarbon pickup.

Although the initial screening of various generic amines benefited from these data, we selected a proprietary solvent for final design.

Sales-gas dehydration

The correct equilibrium water content at the natural gas hydrate point has been a source of confusion within the industry for many years. This is especially true when one designs equipment for arctic conditions where low water dewpoints ensure hydrate prevention.

The commonly used McKetta-Wehe chart correctly correlates water content above the freezing point and warmer hydrate temperature. The water content of natural gas in equilibrium with hydrate, however, is appreciably lower than the water content of natural gas in equilibrium with meta-stable liquid water.

Serious deviations of water content read from charts (e.g., McKetta-Wehe) become increasingly severe at lower temperatures. Preventing hydrate formation in arctic conditions, therefore, makes the area of meta-stable equilibrium significant for process design and can result in underdesigned dehydration equipment.

GPA Research Report RR-45, "The water content and correlation of the water content of methane in equilibrium with hydrates," studied water content measurements of a simulated Prudhoe Bay gas. Projected gas transmission lines across Arctic permafrost terrain drove the demand for these studies. Fig. 3 in the report shows corrected values of water content for simulated Prudhoe Bay gas.

CO2 dehydration

In a methane-water system at constant temperature, the equilibrium water content of water-saturated vapor decreases with increasing system pressure, as expected. In a CO2-water system between 300-800 psia and 3,000-5,000 psia, the equilibrium water content increases substantially with increasing pressure at constant temperature.

If CO2 is saturated with water at 600 psia, it can compress to 4,000 psia, cool down to the original saturation temperature, and still be superheated with respect to water. This can result in a significant water dewpoint depression at the higher pressure. This idea was the basis for our evaluation of an inter-stage cooling option as an alternative to conventional TEG dehydration.

Data from GPA Research Report RR-120, "Water content values of a CO2-5.31 mole % methane mixture," clearly show that-although significant water dewpoint depression is possible for pure CO2 (dewpoint of 213° F.)-small amounts of diluents lower the saturation water content significantly.

In the aMDEA gas-treating design, the CO2 injection gas stream contains about 3% methane that would result in a 55° F. equilibrium dewpoint, at best. We therefore discarded the inter-stage cooling option.

Figs. 6, 7, and 8 in the GPA report show the effect of small amounts of diluents that significantly lower the saturation water content.

These examples show that data from GPA-sponsored research can directly apply to process design. Such research offers tangible benefits to the industry and should, therefore, be enthusiastically supported. This type of research can play an ever-increasing role in reducing the risk of incurring costly design deficiencies.

References

  1. Cook, P.J., and Losin, M.S., "Membranes provide cost-effective natural gas processing," Hydrocarbon Processsing, April 1995.
  2. Lokhandwala, K.A., and Jacobs, M.L., "New membrane applications in gas processing," presented to the 79th GPA Annual Convention, Atlanta, Mar. 13-15, 2000.
  3. Kolbe, B., et al., "Acid gas removal," Hydrocarbon Engineering, May 2000, pp. 71-74.
  4. Minkkinen, A., Larue, J.Y.M., Patel, S., and Levier, J.-F., "Methanol gas-treating scheme offers economics, versatility," OGJ, June 1, 1992, p. 65.
  5. Ryan, J.M., and Holmes, A.S., "Cryogenic distillation separation of acid gas from methane, solidification prevention," US Patent 4318723, Mar. 9, 1982.
  6. BASF AS, "The customized method for removing CO2 and H2S from gases: BASF activated MDEA," company brochure.

The author

Click here to enlarge image

Arif Habibullah ([email protected]) is a senior technical director and manager of process technology with Parsons Energy & Chemicals Group, Arcadia, Calif. For the past 25 years, he has been involved in process design of various oil and gas production, gas treating, LNG, and petroleum refining projects. More recently, he has served as process manager on various project development programs for the Alaska North Slope. Habibullah holds a BS and MS in chemical engineering. He is a registered professional engineer in California.