Field tests of a scalable, microscale, N2-rejection unit have demonstrated the effectiveness of the portable and cost-effective NRU to upgrade low-pressure (<100 psig) and low-volume (≈ 100 Mcfd) low-btu gas to pipeline quality.
The NRU was designed, constructed, and tested1 at Elmdale field (Chase County, Kan.) under a grant from the Stripper Well Consortium (SWC, Pennsylvania State University).
The project was a joint effort between the Kansas Geological Survey (University of Kansas) and the American Energies Corp., Wichita.
Low-btu gas
Pipeline specifications for natural gas in the US vary, but a heat content of 950 btu/cu ft is required for sale and transport. US reserves of natural gas are about 238 tcf.2 "Subquality" gas may constitute as much as 17.5 tcf of reserves in the Midcontinent, 9 tcf in the Rocky Mountain region,3 and 60 tcf in the US.4
Overall, 33% of the 1,253 gas analyses5 recorded in Kansas in the last 50 years are low-btu (<950 btu/cu ft). Low-btu gas is more common in Permian and Upper Pennsylvanian reservoirs in Kansas, but it is found in reservoirs of all ages and lithologies.5 In the Midcontinent, the presence of N26 7 often results in natural gas being low-btu, while nationwide 17%4 of known gas reserves are "subquality" due to N2 contamination. Thus, N2 is a major target for removal so that low-btu natural gas can be upgraded to pipeline quality.Much of today's gas production is from large fields where low-btu gas can be processed in central upgrading facilities that use cryogenic separation,8-11 conventional pressure swing adsorption,12-15 or lean oil absorption (Mehra process),16 but such economy-of-scale is not feasible for smaller and often isolated fields.
A significant portion of N2-induced low-btu gas is found in modest to small fields that are owned and operated by small independent producers.4 At present, most of this low-btu gas is either shut-in behind pipe or simply abandoned because it cannot be burned locally or blended with readily available higher-btu gas.
Development of inexpensive N2-rejection technology, designed for low-volume, low-pressure gas wells, can significantly increase the contribution of marginal low-btu gas to the nation's gas supply.
NRU process
Pressure swing adsorption is a proven commercial technique that is widely used in many industries to separate gases by use of an adsorbent bed. This bed selectively adsorbs one or more components from the feed gas under pressure while allowing other components to pass through unadsorbed. The adsorbed components are then recovered by reduction of pressure on the bed, and the reactivated bed reused for adsorption in the next cycle.
The demonstration NRU discussed here uses the process of vacuum swing adsorption,17 a variation of PSA in which the bed is desorbed under subatmospheric pressure (i.e., vacuum). The NRU uses readily available and non-patented activated carbon (made from coconut husks) as an adsorbent bed to adsorb methane and heavier hydrocarbons under pressure while rejecting the entrained N2 as a vent stream.
Adsorbent beds vary between PSA plants. Some, as the NRU discussed here, have fixed pore openings (as in activated carbon or molecular sieve), while others have custom-designed pore openings (as in molecular gate) to trap targeted gas molecules.
In general, PSA processes differ by the number of towers and the number of stages in each tower required to separate the feed hydrocarbons from the entrained N2.
This NRU has two towers and uses three stages:
1. Adsorption under pressure.
2. Venting to 2 psig.
3. Desorption under vacuum.
Between the adsorption and desorption stages, many PSA systems employ the product (upgraded) gas to purge each tower of residual unadsorbed gas. Instead of this purge stage, which complicates plant controls and increases compression costs, the demonstration NRU vents the unadsorbed gas from each tower at the end of adsorption. The degree of separation13 achieved in PSA process depends on feed-gas composition, charge and vent pressure, and system temperature.
The NRU was designed to have the following characteristics:
• Nonpatented processes and off-the-shelf equipment to minimize construction costs.
• Easily obtained adsorbent bed of inexpensive, non-patented activated carbon.
• Skid-mounted modular units that provide mobility and scalability.
• Small environmental footprint (≈ 400 sq ft).
• Few moving parts (outside engine and compressor) to minimize maintenance costs.
• Powered by solar panels and low-btu feed gas for operations at remote locations outside electric grid.
NRU layout
Fig. 1 shows the compact layout of the NRU.18 Feed gas enters the plant through a 2-in. line, passes through a scrubber for removal of entrained moisture, and then passes through a flow meter into the adsorption/desorption towers. Each tower, made of carbon steel, has a 48-in. diameter, is 8 ft tall, and is designed to handle feed rates of around 100 Mcfd.
The modular design allows additional sets of towers to be added or removed to handle increases or decreases in feed volumes. Electronically controlled solenoid valves allow feed gas to flow into one tower for adsorption while isolating the other tower for desorption under vacuum. A small fraction of the N2-rich vent (waste) gas is used as instrument gas to operate the pneumatics of the control panel.
Ports at the base of each tower provide access for removal of spent bed materials and tower cleanup. Off-the-shelf activated carbon made from coconut husks (Fig. 2a) was used to charge the towers (Figs. 2b and 2c). Each tower was charged with about 2,200 lb of this material, purchased in 1,100-lb bags at a cost of 7¢/lb.
A 6-cyclinder, 50-hp VGG-330 engine (Fig. 1), operating on the low-btu feed gas, drives the compressor, designed for vacuum service, which pulls a vacuum on each tower during its desorption cycle. Desorbed methane flows through a 2-in. line to a gas scrubber and then to the compressor via a 3-in. line. N2-rich vent gas from each tower travels to a flare tower by a 2-in. line. The compressed (upgraded) gas passes through a condensate-removal tower (Fig. 1) before flowing into a surge tank (5 ft diameter and 25 ft long) where it mixes and attains uniform composition before being discharged to the nearby pipeline via a sales-gas meter.
Anticipating a maximum operating pressure of 75 psig, the NRU was successfully pressure tested at 105 psig to check for leaks. Thereafter, it was tested by pulling a vacuum of 26 in. mercury that was held over a 2-day test period.
Operational stages
1. Low-btu feed gas (broken red line, Fig. 3) is fed into the bottom of Tower 1, charging it to the requisite pressure that depends on the feed composition (i.e., N2 and heavy hydrocarbon content) and btu requirements of the pipeline company.
Tower 2, in desorption mode, is evacuated from the bottom to vacuum (22-28 in. mercury) during this stage. The charging time for Tower 1 depends on the feed flow rate and pressure and tower fill-up volume.
During this charging period, hydrocarbons are preferentially adsorbed in the bed of activated carbon inside Tower 1. The free gas remaining in the space between the carbon particles and in the dead volume below the metal grate supporting the bed is rich in N2.
2. The unadsorbed N2-rich free gas in Tower 1 is vented to the atmosphere through the flare tower (broken blue line, Fig. 4) until the tower pressure reaches 2 psig. Tower 2 at this time is kept under vacuum. The length of the venting period is proportionate to the magnitude of the Tower 1 charge pressure.
3. Tower 1 is connected to the compressor from the bottom (red line, Fig. 5) and desorbed under vacuum, while the desorbed Tower 2 is connected (from bottom) to the low-btu feed stream (broken red line) for charging up to the same pressure as Tower 1 in Stage 1. Reduction of pressure in Tower 1 from 2 psig to vacuum (22-28 in. mercury) results in desorption of the adsorbed hydrocarbons (in Stage 1).
The hydrocarbon-rich desorbed gas leaving Tower 1 will be of pipeline quality (≥950 btu/cu ft) when the plant settings (i.e., charge-up pressure and vent pressure) are optimally set for the composition of the feed.
NRU performance
Feed gas (average: 715 btu/cu ft and C2H6+/CH4+ ≈ 7.9%) consisted of commingled production from several wells, some of which produced slugs of water along with gas. These conditions, along with valve adjustment at the upstream central manifold to maintain minimum flow rates and pressures, resulted in variation in feed-gas composition. It is not uncommon for the feed-gas composition to fluctuate under real-life operating conditions of marginal gas wells.
During initial testing, the feed gas averaged around 687 btu/cu ft. Under this condition, the plant was optimized to produce pipeline quality gas (>950 btu/cu ft) by charging the towers to 34 psig, then venting unadsorbed N2-rich gas (from the top) to 2 psig, followed by vacuum desorption (25 in. mercury).
These settings (Table 1) resulted in a sales/feed ratio of 0.54 (i.e., 54% of the low-btu feed gas by volume was upgraded to pipeline quality). Thus a feed gas with 63% average hydrocarbon content (CH4+ % mole) was upgraded to a product stream containing around 84% hydrocarbon content resulting in 73.2% of hydrocarbon recovery and 75.7 % of btu recovery.
The btu recovery is calculated as the ratio of the product of total btu coming into the plant (i.e., feed volume times feed btu/cu ft) to that recovered in the sales stream (i.e., sales volume times sales btu/cu ft). Under these settings, the vented gas contained about 63.1% N2 (% mole) resulting in an average N2 rejection efficiency of 76.7%.
The sales/feed ratio critically determines the plant economics. Given similar feed compositions, higher sales/feed ratios result in greater recovery of the hydrocarbons and higher volumes of pipeline-quality gas for sale.
Given unchanging feed composition and bed adsorption characteristics, the sales/feed ratio depends on the differential between the tower charge pressure (34 psig, as stated earlier) and the vent pressure (2 psig), the dead volume within each tower, and volume of gas desorbed from the beds during the venting process.
Variation in feed composition imparted some uncertainty to optimization of plant settings. For example, the pressure differential between tower charge pressure and vent pressure was reduced to 20 psig and 2 psig, respectively, to increase the sales/feed ratio. By the time the plant could be operated under lower tower charge pressure, however, the feed-gas composition changed to an average of 743 btu/cu ft. The plant produced pipeline-quality gas (964 btu/cu ft) at a higher sales/feed ratio of 0.60 (Table 1).
It was difficult to determine, however, if the lower tower charge pressure resulted in slightly higher CH4 recovery efficiencies (75.4%) and slightly lower N2 stripping efficiency (72.6%), or if these were caused by improved feed-gas quality.
Later in the optimization process, the plant was connected to a different combination of wells resulting in a feed with poorer average heat content (622 btu/cu ft vs. 715 btu/cu ft). Also, the ratio of the heavy hydrocarbon fraction (C2H6+/CH4+) decreasing to 3.8% from 7.9% necessitated higher tower charge pressures to produce pipeline-quality gas. The variation in feed btu content was less than 5% during this plant optimization study (Table 2).
When the plant was run with tower charge pressures of 15 and 30 psig and vent pressure of 2 psig (i.e., settings close to that necessary to upgrade feed averaging 715 btu/cu ft with a heavy hydrocarbon fraction ≈ 7.9%), the product gas was of subpipeline quality, i.e., 831 and 881 btu/cu ft, respectively. Raising the tower charge pressure to 70 and 66 psig, followed by venting to 13 and 9.5 psig, increased the heat content of the desorbed gas to around 920 btu/cu ft but also resulted in lower sales/feed ratios, i.e., 45% and 49%, respectively.
Higher tower charge pressures result in greater pressure differential during the vent process and therefore greater loss of hydrocarbons and lower sales/feed ratios.
In the current tower design (Fig. 6a), the grate supporting the adsorption bed was incorrectly designed to be located above the tower access hole resulting in 20 in. of dead volume at the bottom of each (8-ft) tower. This dead volume is not filled with any activated carbon, and N2-rich low-btu gas (at 2 psig) occupies the dead volume at the end of the vent phase when venting is taking place solely from the tower top.
During the desorption stage, this low-btu gas in the dead volume entered the surge tank and lowered the btu of the product (sales) gas. To remediate the problem, attempts were made to see if simultaneous venting from both the top and bottom of the tower would help better vent the residual unadsorbed N2-rich gas.
This replumbing met with some success; the product gas was of pipeline quality (at 958 btu/cu ft) when the tower charge pressure was set at 69 psig and vent pressure to 3 psig. This setting resulted in a sales/feed ratio of 0.39. The sales/feed ratio was improved slightly to 0.40 when the tower charge pressure was set to 72 psig and the vent pressure was set at 4 psig.
It is apparent from these results that this NRU can upgrade a feed gas with heat content as low as 633 btu/cu ft and a heavy hydrocarbon fraction around 3.8%. It is critical to note that both the heat content and the amount of heavy hydrocarbons in the feed stream dictate the optimum operational settings for the plant to attain pipeline-quality sales gas. Any deterioration in the quality of the feed will require towers to be charged to higher pressures resulting in higher pressure differentials during the venting process, in greater volumes of gas lost, and in lower sales/feed ratios.
Also, poor quality feed gas has lower amounts of hydrocarbons to recover and thus will naturally result in lower sales/feed ratios. With the poorer quality feed (at 633 btu/cu ft), the btu-recovery (Table 2) efficiency decreased to about 59% as compared with 75+% obtained with a superior feed whose heat content averaged 715 btu/cu ft.
Plant controls
Only two parameters, the tower charge pressure and the vent pressure, are critical to optimizing the plant for upgrading a low-btu feed to pipeline quality. Different combinations of these two parameters must be tested to determine the settings for obtaining pipeline-quality product stream with minimum hydrocarbon loss in the vent stream.
A programmable logic controller pneumatically opens and closes the solenoid valves controlling the flow of gas into and out of the two towers. Charge and vent pressures (or times) are input to the PLC for continuous operation monitored by a once-a-day visit by the plant operator.
The following guidelines will help optimize plant settings due to changes in feed-gas composition:
• If the feed btu and heavy hydrocarbon fraction increase, the towers may be charged to lower pressures to obtain pipeline-quality sales stream. Sales/feed ratios tend to improve with higher quality feed.
• If the feed btu and heavy hydrocarbon fraction decrease, the towers may be charged to higher pressures to upgrade to pipeline quality. Sales/feed ratios will decrease with poorer feed quality.
• With pipeline-quality product stream attained by adjustment of the charge pressure, the vent pressure may be fine tuned to optimize the sales/feed ratio.
Operational problems
During the initial testing period, the sales/feed ratio would suddenly decline despite minor variations in feed composition and unchanging operational settings after a few days of operation. Visual evidence of carbon particles being ejected from the vent tower during each vent phase confirmed that bed blowout was taking place.
This was caused by the absence of any filter in the upper flange connecting the tower to the vent line and by the pressure shock (release) of the venting process. With bed material being blown out, the dead volume increased at the top of each tower, resulting in performance degradation.
Opening the flange atop each tower allowed a visual check for bed blowout and revealed that each tower had lost about 18 in. of bed from the top of the column (Fig. 6a). This problem was solved after the towers were topped off with fresh activated carbon (Fig. 6b) and an appropriately sized screen filter was set below the upper flange.
It was also discovered that the bottleneck affecting the NRU sales (volume) throughput is primarily the time taken to desorb a tower from vent pressure (≈ 2 psig) to 22-25 in. mercury.
The tower evacuation time depends on the tower (or bed) volume and the compressor capacity and is normally longer than the tower charge-up time given sufficient pressure and rate in the feed line. Thus, the tower charging process often had to be slowed to make the charge time equal to the evacuation time for continuous operation.
One important lesson learned from this project is that a strong compressor, capable of evacuating the tower (volume) as quickly as possible, should be employed to reduce process cycle time and increase plant throughput.
Heavy HC adsorption
A mass balance of the heavy hydrocarbons (C2H6+) conducted on the feed and respective upgraded sales gas showed that about 98% of the heavy hydrocarbons entrained in the feed were recovered in the product stream. The bed of activated carbon was therefore efficient in capturing the incoming heavy hydrocarbons and the desorption process was equally effective in recovering the heavy hydrocarbons.
Methane was less efficiently captured—only 58-68% of methane in the feed gas was recovered, depending on gas composition. The vent stream was mostly made up of unadsorbed nitrogen and some methane, which limits the feasibility of capturing and upgrading the vent gas to pipeline quality.
Plant economics
The microscale NRU at Elmdale field was designed to handle around 100 Mcfd of low-btu feed gas. A local commercial upgrading plant offered AEC a 51% seller's percentage for low-volume sales less than 450 Mcfd (i.e., AEC gets paid for 51 Mcf of pipeline-quality gas for sale of every 100 Mcf of low-btu gas). Also, the sales contract carried a constraint that the feed could not have nitrogen content in excess of 28%, which, in effect, disqualified the feed from Elmdale field because its nitrogen content was >33%.
Additionally, AEC had to consider the cost of transporting the low-btu gas from the production wells to the commercial plant, provided that a nearby pipeline was available and its operator agreed to transport the low-btu gas. AEC estimated that the transportation would additionally cost about 13% of the volume of low-btu gas that it sold to the commercial upgrading plant. In summary, if AEC were to sell 100 Mcf of low-btu gas to the commercial plant, it would get paid for 38 Mcf of pipeline-quality gas after deduction of the upgrade and transportation costs.
In comparison, if AEC were to use the micro-NRU to treat its low-btu gas on site, it would save the transportation costs. Given the average sales/feed ratio achieved at the micro-NRU, if AEC were to process 100 Mcf of low-btu gas with average heat content of 622 btu/cu ft and 715 btu/cu ft at its own micro-NRU, it would get paid for about 39 and 57 Mcf of pipeline-quality gas, respectively. Thus, the micro-NRU offers competitive value to AEC, particularly for higher quality low-btu feed.
AEC built the NRU with off-the-shelf vessels, pipelines, control valves, engine, and compressor in its workshop with its own maintenance/service crew at a cost of $120,000 in 2008. This achievement highlights the simplicity of the plant design and should therefore provide confidence to other small operators to venture into building a microplant for their needs without relying on expensive expertise from consultants.
Based on average performance (sales/feed ratio), the payout time calculates to be 51 and 35 months, respectively, for feed gases averaging 622 and 715 btu/cu ft, assuming pipeline-quality gas to be priced at $2.00/Mcf and a feed volume of 100 Mcfd.
Lessons
• It is possible to upgrade low-btu gas with a heat content as low as 630 btu/cu ft and a heavy hydrocarbon fraction (C2H6+/CH4+) of at least 3.8% to pipeline quality (>950 btu/cu ft) using a simple, cost-effective microscale NRU with an adsorption bed consisting of readily available nonpatented activated carbon made from coconut husks.
• Dead volume within each tower must be minimized relative to tower volume. Initial operation data indicate that greater bed mass (with minimum dead volume) results in larger volumes of adsorbed hydrocarbons and therefore better sales/feed ratio.
• The off-the-shelf bed of activated carbon is efficient in adsorbing heavy hydrocarbons (C2H6+) from the feed stream and desorbing it under vacuum. Methane removal is less efficient, so that dry-gas feed stock may be difficult to upgrade to pipeline quality.
• Despite the cost of the compressor being one of the major expenses in the building of the NRU, an effective compressor must be used to evacuate towers (to maximum vacuum) in the shortest time so that plant efficiency and throughput are not compromised.
• A constant feed chemistry and pressure require less operational supervision. Plant settings, i.e., tower charge pressure and vent pressure will require reoptimization if feed composition (btu, N2 %, and C2H6+/CH4+ ratio) changes.
• Both nitrogen content and the fraction of heavy hydrocarbons in the feed control the optimum plant settings and determine its efficiency. Greater amounts of heavy hydrocarbons in feed result in higher sales/feed ratio and thus better plant economics.
• A micro-NRU may be a viable option if pipeline transport costs of the low-btu gas to commercial upgrading plants are significant.
Acknowledgments
The authors acknowledge the financial support received from the Stripper Well Consortium (University of Pennsylvania), Alan L. DeGood (American Energies Corp.), and the Kansas Geological Survey (University of Kansas). They also thank Stephen L. Moore (AEC) for technical advice related to the design of the NRU.
References
1. Bhattacharya, S., Newell, K.D., Watney, W.L., and Sigel, M., "Low-cost plant upgrades marginal gas fields," E&P, August 2008, pp. 102-103.
2. US Energy Information Administration, "Natural gas reserves summary as of Dec. 31, 2007," available at: http://tonto.eia.doe.gov/dnav/ng/ng_enr_sum_dcu_NUS_a.htm.
3. Hugman, R.H., Vidas, E.H., and Springer, P.S., "Chemical composition of discovered and undiscovered natural gas in the lower-48 United States," Gas Research Institute, Chicago, Publication GRI 90/0248, 1990.
4. Lokhandwala, K., and Zammerilli, A., "New nitrogen-rejection membrane technology commercialized," GasTIPS, Vol. 12, No. 4, 2006, pp. 16-20.
5. Newell, K.D., "Geochemical trends in gas quality in Kansas," American Association of Petroleum Geologists, 2007 Annual Convention & Exhibition, Long Beach, Abstracts Volume, p. 101 (www.kgs.ku.edu/PRS/publication/2007/2007-8/index.html).
6. Beebe, B.W. (ed.), "Natural Gases of North America," American Association of Petroleum Geologists, Memoir 9, 1968.
7. Jenden, P.D., Newell, K.D., Kaplan, I.R., and Watney, W.L., "Composition and stable-isotope geochemistry of natural gases from Kansas, Midcontinent, U.S.A.," Chemical Geology, Vol. 71 (1988), pp. 117-147.
8. Butts, R.C., Chou, K., and Slaton, B., "Nitrogen-rejection process developed for small fields," OGJ, Mar. 13, 1995, p. 92.
9. Millard, R., and Dreaves, B., "A boost for N2 rejection," Hydrocarbon Engineering, July 2004.
10. Ruehle, S., and Coalmer, L., "New Hugoton gas plant starts up," OGJ, Aug. 17, 1998, p. 99.
11. Emmons, F.R., Hudspeth, L.D., Clancy, J.P., Zornes, D.R., and Philcox, J.E., "Nitrogen management at the East Binger Unit using an integrated cryogenic process," SPE 15591, 61st Annual Technical Conference and Exhibition, Los Angeles, Oct. 5-8, 1986.
12. Mitariten, M., "Economic N2 removal," Hydrocarbon Engineering, July 2004, pp. 53-57.
13. Shirley, A., Porto, J., Ramachandran, R., and Kleinburg, W., "Non-cryogenic nitrogen rejection from natural gas," 75th Gas Processors Association Annual Convention, Denver, Mar. 11-13, 1996.
14. Mitariten, M., "New technology improves nitrogen-removal economics," OGJ, Apr. 23, 2001, p. 42.
15. Willis, J., Shemaria, M., and Mitariten, M.J., "Production of pipeline quality natural gas with the molecular gate CO2 removal process," SPE 80602, SPE/EPA/DOE Exploration and Production Environmental Conference, San Antonio, Mar. 10-12, 2003.
16. Mehra, Y.R., Wood, G.C., and Ross, M.M., "Non-cryogenic N2-rejection process gets Hugoton field test," OGJ, May 24, 1993, p. 62.
17. Ruthven, D.M., Farooq, S., and Knaebel, K.S., "Pressure swing adsorption," Deerfield Beach, Fla.: VCH Publishers Inc., 1994.
18. Newell, K.D., Bhattacharya, S., and Sears, M.S., "Low-BTU gas in the Mid-continent: A challenge to geologist and engineers," OGJ, Sept. 14, 2009, p. 35.
The authors
More Oil & Gas Journal Current Issue Articles
More Oil & Gas Journal Archives Issue Articles