LPG-recovery processes for baseload LNG plants examined

Nov. 24, 1997
With demand on the rise, LPG produced from a baseload LNG plant becomes more attractive as a revenue-earning product similar to LNG. Efficient use of gas expanders in baseload LNG plants for LPG production therefore becomes more important. Several process variations for LPG recovery in baseload LNG plants are reviewed here. Exergy analysis (based on the Second Law of Thermodynamics) is applied to three cases to compare energy efficiency resulting from integration with the main liquefaction
Chen-Hwa Chiu
Bechtel Corp.
Houston
With demand on the rise, LPG produced from a baseload LNG plant becomes more attractive as a revenue-earning product similar to LNG. Efficient use of gas expanders in baseload LNG plants for LPG production therefore becomes more important.

Several process variations for LPG recovery in baseload LNG plants are reviewed here. Exergy analysis (based on the Second Law of Thermodynamics) is applied to three cases to compare energy efficiency resulting from integration with the main liquefaction process.

Process-design considerations of various degrees of LPG extraction need to be investigated not only for the new grassroots baseload LNG plants but also for the debottlenecking and revamping of existing baseload LNG plants.

The extraction and export of LPG product from a baseload LNG plant may achieve a higher profitability through faster rates of return. These LPG products are often unusual because they are not extracted from gas associated with oil production. The high on-stream availability of the baseload LNG plant also ensures a dependable source of LPG product.

Extraction in a baseload plant

In baseload LNG plants, hydrocarbon refrigerants are usually recovered from the feed gas. Some baseload LNG plants need to import refrigerants if the feed gas is too lean or the process itself dictates.

For a lean feed gas, fractionation for the recovery of the heavier components is difficult. Process-design variations can assist heavy components scrubbing and demethanization.

For a heavier feed gas, recovery of LPG by distillation is necessary because the amount of LPG that can be reinjected into the LNG stream is limited by the heating-value specifications of LNG. This is an area that challenges the process designer to devise a favorable process scheme for achieving better energy efficiency.

Another design aspect associated with LPG recovery is the need for aromatics removal to prevent freezing in the liquefaction equipment. Aromatics such as benzene, toluene, and xylenes can pose freezing problems in low-temperature processing equipment. They can impose extra burdens on the LPG-recovery process.

In a baseload LNG plant, the feed gas goes through acid-gas removal, dehydration, and mercury removal before being precooled. After precooling, heavy components removal for refrigerant make-up and aromatics removal is performed before the feed gas can be liquefied. If the feed gas is lean or refrigerants are imported, there is no need for substantial LPG recovery.

In many baseload LNG plants, LPGs are recovered to provide refrigerant make-up. Excess LPGs are reinjected into the feed-gas stream to be liquefied to the extent allowed by the LNG heating-value specification.

Some baseload LNG plants also recover and export LPG. In these plants, the LPG-recovery process may become more elaborate and may involve some parallel LPG plants, with utilities essentially independent of the LNG facilities.

Fig. 1 [50,944 bytes] illustrates a typical LPG-recovery process.1 2

The dried sweet gas is first precooled before entering the scrub column where distillation performs the first step of heavy-components removal. The overhead product of the scrub column is further cooled and liquefied in the main heat exchanger (MHE) to provide the LNG product.

The bottoms product enters the fractionation train, which consists of de-ethanizer, depropanizer, and debutanizer. The products of the fractionation train are fuel gas, ethane, propane, butanes, and condensates. LPG or condensate storage tanks and unloading facilities are usually provided.

The LPG export facilities at the Arun LNG plant, the North West Shelf (NWS) LNG plant, and the Badak LNG plant have been documented.3-6

In 1996 at its Burrup Peninsula LNG plant, the NWS Project commissioned additional facilities of LPG separation, storage, and loadout jetty to provide a nominal export capacity of 800,000 metric tons/year (tpy) of LPG.

The Badak LNG plant also exports LPG and is expanding its LPG facilities.

The Arun LNG plant is an example where major volumes of LPG and condensate are exported.3-5 It is producing approximately 120,000 b/d of stabilized condensate. The average LPG production is about 14,000 b/d propane and 11,000 b/d butane.

The Arun LNG plant has four condensate tanks with a capacity of 530,000 bbl and four LPG tanks (three tanks with a total capacity of 525,000 bbl, and one tank with 350,000 bbl).

There are one LPG berth for LPG loading, one SPM (single point mooring), and one MBM (multiple-buoy mooring) for condensate loading. The condensates from the LPG fractionation trains are combined with the condensate from a front-end stabilizer and sent to the condensate tanks.

Arun LNG units are now being operated in integration with the LPG units, thus yielding improved plant energy consumption and efficiency. This reduces any operational problem such as the interruption of the LNG trains if one of the LPG units or LNG units trip.

The excess gas-turbine capacity in the LPG units can be used to increase the pressure and lower the temperature of residue gas entering the LNG liquefaction units.

Extraction requirements

Heavy liquid removal is mainly dictated by the heating-value specifications of the LNG products. The typical heating-value specification for LNG product is around 1,050-1,190 BTU/std. cu ft higher heating value.

Additional benefits are that the LPG extraction also provides a quality control of LNG-product density. This is desirable to avoid stratification of LNG in storage and to prevent potential rollovers.

Although the heavy alkane solubilities in LNG may be of some concern, experimental data indicate that the solubilities are quite high, relative to aromatics.8 If the condensates are not removed in a front-end stabilizer, however, they need to be removed along with the heavy alkanes that are removed before final liquefaction.

The common refrigerant components are nitrogen, methane, ethane or ethylene, propane, and in some cases butane.

The refrigerant losses occur from the seals and flanges of the refrigerant compressor systems, also from valves to flare and others, over a period of time and during upsets in process operation or short-term shutdowns. To meet the refrigerant make-up requirements, it is necessary to import or to recover the refrigerant components during plant operation.

Recovered propane and ethane need to be sufficient for the refrigerant make-up requirements. These recoveries depend upon the proper design and operation of the scrub column and the fractionation train. Usually the refrigerant make-ups are recovered during normal operation and stored in refrigerant storage tanks.

Methane make-up is not a problem, and nitrogen is derived from the in-plant nitrogen generator or nitrogen-storage system.

Aromatics have limited solubilities in LNG,8 much less than the heavy alkanes. The aromatics usually found in feed-gas streams are benzene and toluene and to a lesser degree xylenes. Solubilities of these aromatics are in the parts-per-million range at LNG temperatures of approximately 2240 to 2264° F.

Aromatics are usually removed in the scrub column along with the heavy components. Preventing freeze-up problems during precooling and liquefaction requires removal of the heavy ends from the feed gas to less than a few ppm aromatics and less than 0.1 mole % of the heptane plus fractions.

Fuel-gas requirements can limit design of the scrub column and the de-ethanizer. The de-ethanizer may withdraw ethane several trays below the top tray, while gas from the column top is sent to fuel.

Nitrogen rejection after the MHE, if required, will limit the amount of methane-rich fuel from the top of the de-ethanizer. This is because most baseload LNG plants produce no fuel gas for sale.

Fuel production is therefore limited to what can be consumed in the plant itself. The Woodside LNG plant in Australia is an exception.

Thus, this will put a special requirement on the amount of methane and ethane that can be sent from the bottoms of the scrub column to the fractionation train.

Process-recovery parameters

The most important process parameters in design for LPG recovery are the operating pressure of the scrub column, the feed-gas composition, and the reflux temperature of the scrub column.

The liquefaction pressure is critical to the energy efficiency of LNG liquefaction.9 10 Liquefaction efficiency improves with increasing pressure, up to around 1,000 psig. On the other hand, the recovery of the heavy hydrocarbons by the scrub-column operation becomes more difficult with increasing pressure.

The separation pressure for the scrub column is an important design parameter. If this pressure is close to the mixture's critical pressure, scrub-column operation can become difficult. If the pressure is higher than the mixture's critical pressure, fractionation at the scrub column is impossible.

The mixture's critical point is a strong function of the feed-gas composition. Therefore, no single arbitrary criterion is available for the selection of the scrub-column pressure.

Use of a heavy alkane recycle, such as the C4 fraction, to the scrub-column overhead, however, can raise the critical pressure of the mixture and thus the operating pressure for the scrub column. This problem will be discussed next.

The mixture's critical pressure is a strong function of the heavy component in the LNG feed stream. This is because the locus of the mixture critical point is highly nonlinear and is not a simple molar average of the critical pressures of the constituent components.

The critical pressure is also a strong function of the methane concentration.

On the other hand, the addition of the heavy component to the mixture to be separated has a significant effect on the critical pressure. This is illustrated in Fig. 2 [55,725 bytes] in which the addition of C4 will increase the critical pressure of the methane and propane mixture.

This principle has been practiced in baseload LNG plant operations by use of a C4 recycle to assist the scrub-column operation. This can increase LPG production and decrease the higher heating value of the LNG.

A lower reflux temperature favors the scrubbing action. Low-level propane can provide refrigeration to around 238° F. Lower reflux temperatures of near 2100° F. can be provided by mixed refrigerant, usually requiring an additional bundle in the MHE.

The lower the reflux temperatures, however, the higher the reboiler temperatures will be. The reboiler temperature will be limited by the mixture condition at the reboiler.

Extraction-process variations

Various LPG-extraction processes are available. Since the fractionation train itself varies little, only the scrub-column portion will be discussed.
  • A scrub column with reflux and reboiler is a typical, commonly used process. The operation pressure is usually limited by its proximity to the critical pressure of the mixture because the mixture cannot be separated into vapor and liquid above and beyond the critical pressure. The reboiler temperature can also become a limiting factor, however.

If the reboiler temperature is close to or greater than 120° F., the column bottoms may become flooded and separation is difficult. This is when a C 4 recycle to the scrub column overhead is helpful.

The C4 recycle helps the start-up operation and shortens the time of lining out the scrub column and the fractionation train. Depending on the degrees of LPG extraction desired, either a partial recycle or a total recycle of C4 recovered from the debutanizer can be performed.

There are also process variations in which only propane refrigeration is sufficient for the reflux condensing and a two-bundle MHE can be used.

  • Use of a scrub column with a demethanizer is equivalent to separating the bottom section of the scrub column into a demethanizer, except that the demethanizer is running at a lower pressure that favors easier separation. This scheme will also produce fuel gas from the top of the demethanizer. The scheme may not be suitable if there is a nitrogen-removal column, which also generates additional fuel gas. If this excess fuel is not posing a problem, then this arrangement may offer more flexibility in operation.
  • Instead of letting down a high-pressure feed gas through a control valve, there are several processes that use the high feed-pressure potential to incorporate an expander and a booster compressor around the scrub column.
This is to use an expander to let down the high-pressure feed gas, especially when the pressure is greater than the critical pressure, before entering the scrub column. The isentropic expansion through an expander can achieve a lower temperature than an isenthalpic expansion through a valve.

There are two major process variations:

One process version, Expander-1, requires a lower temperature mixed refrigerant in the reflux condensing step and requires an additional bundle in the MHE.12Another process version, Expander-2, uses only a higher temperature propane refrigerant for reflux condensing and more heat integration through a feed intercooler.13

This has an advantage of using a two-bundle MHE. Furthermore, if the feed gas is greater than the critical pressure of the mixture, the expander outlet will provide enough cold liquid to the scrub column, and a reflux condenser is not required.

The overhead product from the scrub column is compressed by a booster compressor, which can be driven by the expander, thus providing a feed to the MHE to be liquefied at a higher pressure than the pressure used in the scrub column.

This scheme has two advantages. First the pressure potential of the feed stream is converted to refrigeration via the expander, thus saving the refrigerant compressor horsepower. Second, the expander can provide the power required for the booster compressor.

  • A scrub column with double heat integration was developed to heat-integrate the scrub column's feed precooling and reflux condensing with MHE's liquefaction process.14 Precooling of the feed gas to the scrub column is performed in the warm bundle of the MHE. The precooling temperature can be controlled by the bypass around the MHE. The gross overhead is fed to the middle bundle to be condensed and subcooled. A portion of this subcooled liquid is returned to the scrub column for reflux. The rest is sent back to the cold bundle of the MHE to be further subcooled. The advantage of this scheme is that the energy demands for precooling and condensing can be matched nicely with the mixed refrigerant cooling curve on the shell side of MHE, thus optimizing the match between the demand and supply of refrigeration.
  • The scheme of an expander plant followed by liquefaction was developed for high acid-gas content and high feed-gas pressure.15 Essentially, an expander plant is used in the front end simultaneously to remove most of the acid gas and lower the pressure to achieve the heavy removal so that the feed-gas temperature can be dropped to around 2120° F.
The C 2 precooling is followed by a three-component mixed refrigerant (nitrogen, methane, and ethane or ethylene) liquefaction process.
  • The process using a scrub column with Ryan/Holmes process was proposed to remove CO2 along with C2-plus at the scrub column first followed by the Ryan/Holmes process in the fractionation train.11 It reduces the size of the dehydration units because of the elimination of the front-end acid-gas-removal unit which would add moisture back to the feed gas.
This scheme would add a column to split CO 2/C 2 before the de-ethanizer. For high CO 2-containing feed, however, the process scheme becomes more complicated.

Exergy analysis

Energy integration in process design is important. The basic yardstick for comparing the energy efficiency of different process schemes is the exergy analysis based on the Second Law of Thermodynamics. 1 16 17

This is because the fractionation and heat integration are better analyzed by exergy analysis without resorting to more time-consuming total plant simulations.

(Editor's note: "Exergy" is the maximum amount of work potential of a given system or form of energy in relation to the surrounding environment. An "exergy analysis" is performed to pinpoint the losses of work potential in a system.)

Three process schemes discussed previously were analyzed exergetically.

The first scheme is the Base case (Fig. 3 [61,509 bytes]) in which the commonly used scrub-column arrangement with C4 recycle is used.

The second scheme, Expander-1, is shown in Fig. 4 [57,492 bytes], in which the expander and booster compressor are used with the reflux condensing performed with mixed refrigerant in a third bundle of the MHE.12

The third scheme, Expander-2, is shown in Fig. 5 [58,520 bytes], in which no reflux is needed and only propane refrigeration is needed to cool the scrub-column overhead before it is liquefied in a two-bundle MHE. 13

The exergy calculations were performed after computer simulations of the three process schemes. The results are shown in Table 1.

The exergy losses calculated were normalized with respect to the amount of LNG produced, expressed as BTU/lb-mole LNG. The liquefaction exergy losses were only for the feed-gas stream accounted to the exit of the main heat exchanger.

The results shown in Table 1 indicate that Expander-1 case is the most efficient and the Base case is the least efficient. Both the Base case and the Expander-1 case, however, would require a reflux condenser based on mixed-refrigerant cooling and a three-bundle MHE.

The Expander-2 case, requiring a simpler two-bundle MHE, is less efficient than the Expander-1 case.

Acknowledgment

The author acknowledges the valuable comments of Frank Richardson and Fred Staible, both of the Bechtel Corp., Houston.

References

  1. Chiu, C. H., "Evaluate Separation for LNG Plant," Hydrocarbon Processing, pp. 266-272, (September 1978).
  2. Chiu, C. H., "LPG Recovery in Baseload LNG Plant," Gastech 96 Conference Papers, Volume 2, Vienna, Dec. 3-6, 1996, and Spring National Meeting, AIChE, Houston, Mar. 10-13, 1997.
  3. Naklie, M. M., Penick, D. P., Denton, L. A., and Kartiyoso, I., "Indonesia's Arun LPG plant production is unique in Far East markets," OGJ, Aug. 3, 1987, p. 46.
  4. Soeryanto, J., and Triyantno, A., "Availability and Capacity Improvement of The Arun LNG Plant," LNG 10 Conference Proceedings, Session II, Paper 1, Kuala Lumpur, May 25-28, 1992.
  5. Soemantri, H., and Untung, N., "LNG Plant Design -- A Wish List From an Operating Company Point of View," LNG 11 Conference Proceedings, Paper 2.12, Birmingham, July 3-6, 1995.
  6. Chalis, H. N., Gafar, A., and Zainuddin, "Commissioning and Start-Up of the Sixth Train of The Bontang LNG Plant," Gastech 94 Conference Proceedings, pp. 399-401, Kuala Lumpur, Oct. 25-28, 1994.
  7. Brehaut, W. J., and Concannon, M. J., "LNG Train Debottlenecking -- The Technology Success," LNG 11 Conference Proceedings, Paper 2.6, Birmingham, July 3-6, 1995.
  8. Kuebler, G. P., and McKinley, C., "Solubility of Solid Benzene, Toluene, n-Hexane, and n-Heptane in Liquid Methane," Advances in Cryogenic Engineering, Plenum Press, Vol. 19 (1974), pp. 320-26.
  9. Chiu, C. H., and Richardson, F. W., "Project Challenges of a Baseload LNG Plant," Gastech 93 LNG/LPG Conference Proceedings, pp. 245-256, Paris, Feb. 16-19, 1993.
  10. Jones, G., and Bowen, R. R., "The Effect of Pressure on Liquefaction Processes," LNG 11 Conference Proceedings, Poster A-5, Birmingham, July 3-6, 1995.
  11. Holmes, A. S., and O'Brien, J.V., "Ryan/Holmes Cryogenic Acid Gas/Hydrocarbon Separations Provide Economic Benefits for LNG Production," LNG 7 Conference Proceedings, Session II, Paper 10, Jakarta, May 15-19, 1983.
  12. Newton, C. L., and Gaumer, L. S., "Process for Manufacturing Liquefied Methane," U.S. Patent 4,065,278, (December 27, 1977).
  13. Newton, C. L., "Process for Liquefying Methane," U.S. Patent 4,445,916, (May 1, 1984).
  14. Chiu, C. H., "Process for Liquefied Natural Gas," U.S. Patent 4,445,917, (May 1, 1984).
  15. Chiu, C. H., "Process for the Liquefaction of Natural Gas," U.S. Patent 4,548,629, (October 22, 1985).
  16. Chiu, C. H., and Newton, C. L., "Second Law Analysis in the Cryogenic Processes," The Workshop on the Second Law of Thermodynamics, Washington, Aug. 14-16, 1979; Energy-The International Journal, Vol. 5, No. 8-9, pp. 899-904, (August-September 1980).
  17. Hurstel, X., Lepetit, P., and Kaiser, V., "Refrigeration schemes serve olefin plant needs," OGJ, Sept. 7, 1981, p. 107-23.

The Author

Chen-Hwa Chiu is a principal process engineer at Bechtel Corp., Houston. Previously he was associated with M.W. Kellogg, Exxon Production Research Co., Air Products & Chemicals, and Lummus. He holds a BS from National Taiwan University, an MEngr, and a PhD from the University of Oklahoma, all in chemical engineering. Chiu is a fellow of the AIChE, a member of ACS, and the Cryogenic Society of America. He is a registered professional engineer in Texas and Pennsylvania.

Copyright 1997 Oil & Gas Journal. All Rights Reserved.