by Vaclav Smil
Gas flow depends mainly on reservoir characteristics (above all on the pressure), bore diameter, and age of the well. Daily productivities per well range over four orders of magnitude: they may be as high as 106 m3 and as low as 102 m3. A low-productivity well in Saskatchewan now yields less than 500 m3/day, and the Canadian average in 2012, with about 145,000 operating wells producing 167 Gm3 of gas (CAPP [Canadian Association of Petroleum Producers], 2014), was about 3100 m3/day (1.1 Mm3/year). The US mean for the same year, with about 480,000 operating wells producing roughly 681 Gm3 of dry gas (USEIA, 2014b), was 3900 m3/day. In contrast, new wells in West Siberia’s Urengoy giant field average 500,000–580,000 m3/well a day (Seele, 2012), and 30 wells drilled in the North Dome to supply Qatargas 2 LNG project average 2.75 Mm3 of wet gas a day, while the daily average for 22 production wells supplying Qatargas 1 is 2 Mm3/well, and for 33 wells of Qatargas 3, it is 1.2 Mm3/well (Qatargas, 2014a).
Secular decline of average well productivity is well illustrated by reliable US and Canadian data. In the United States, average well productivity rose from less than 5,000 m3/day during the 1950s to the peak of 12,300 m3/day in 1971, and by 1985, it declined to 4,500 m3/day (Schenk and Pollastro, 2002). Afterward, it kept on declining at a much slower rate to less than 3,300 m3/day by 2005, and it rose slightly to almost 3,900 m3/well by 2012 (Figure 3.3). Both the absolute rates and the pace of decline have been similar for Canadian natural gas extraction: in 1980, the average was 13,500 m3/well, and now, it is just around 3,000 m3/well (CAPP, 2014). Extraction can be also deliberately restricted to prolong a field’s role as a major production: Groningen’s output per well peaked in 1976 at 814,000 m3/day and declined afterward, and between 2010 and 2020, it is capped (with 296 wells and 43.6 Gm3/year) at 400,000 m3/well a day (NL Oil and Gas Portal, 2014). Declining production translates into falling revenues over time unless, of course, there is an intervening substantial price rise.
Figure 3.3 Productivity of US gas wells.
Detailed American data also show that slightly more than 10% of produced gas does not leave hydrocarbon fields (USEIA, 2014a). In 2012, gross withdrawals amounted to 836 Gm3 (43% from gas wells, 17% from oil wells, and the rest from shale and coal bed wells), and while only about 0.7% of all withdrawn gas was vented or flared, 11% were used to repressurize hydrocarbon fields. Once reservoirs begin losing their natural pressure, two major techniques are used to displace remaining liquids toward wellbores and extend the recovery span: flooding with water and injection of gases. Water (fresh, brackish, or ocean) is usually injected into production zones while gas is introduced into reservoir caps to maintain pressure on the underlying liquid. Any readily available gas would do, and some fields use CO2 (an option that has been also promoted as a minor form of future carbon sequestration), but the use of coproduced gas is a convenient, and rewarding, choice because it may raise the overall recovery by 20–40%.
American natural gas wellhead prices (charged by producers selling natural gas into the interstate market) have been always expressed in dollars per thousand cubic feet, abbreviated as $/Mcf (because in the US usage M stands for 1,000, not for 1,000,000 as in the International System of Units). Historical record of average wellhead price goes to 1922 when it was (all prices in current dollars) $0.11/Mcf (USEIA, 2014c). During the crisis years of the 1930s, the price fell to $0.05/Mcf, and since 1938, the Federal Power Commission (FPC) began to enforce the Natural Gas Act that regulated prices of interstate sales of natural gas. In 1954, the Supreme Court ruled that all natural gas sales are subject to the FPC regulation, and hence, all wellhead prices would be regulated, with prices set high enough to cover the cost of production and a fair profit (Foss, 2004; NaturalGas.org, 2014). This system held prices too low, and it created an enormous bureaucratic burden that was not resolved by several attempts at reform. Low prices encouraged higher consumption but discouraged exploration and created supply shortage.
This was finally addressed by the Natural Gas Policy Act of 1978 that created a single national natural gas market and allowed market to set the wellhead price of natural gas. Wellhead price, whose average was $0.16/Mcf during the 1960s and close to $0.20 during the early 1970s, rose to $1.18/Mcf by 1979. Deregulation of US natural gas prices began in 1985, and it was formally accomplished with the Natural Gas Wellhead Decontrol Act (NGWDA) in 1989, with all the remaining regulations on wellhead sales to be removed as of January 1, 1993. Finally, in 1992, a federal order completed the unbundling transportation and sale services: pipelines could no longer sell gas sales, leaving customers a choice of selecting sales, transportation, and storage from any provider. During the 1980s and 1990s, wellhead prices fluctuated mostly between $1.5 and 2.5/Mcf and then climbed steeply to the annual record of $7.97/Mcf in 2008 before falling to $2.66/Mcf by 2012 (Figure 3.4).
Figure 3.4 US natural gas wellhead prices.
Finally, a few paragraphs are presented on the power density of natural gas extraction. Both the areas of individual well sites and their density in major gas fields vary greatly. For example, Jordaan, Keith, and Stelfox (2009) found that gas wells in Alberta claim as little as 1.5 and as much as 15 ha (average of 3 ha/well), and while typical well density in conventional US gas fields is just 0.4 wells/km2, the rate is 1.1–1.5 wells/km2 in Barnett shale, and in some shale gas developments, the density will be up to 6 wells/km2 (NYSDEC [New York State Department of Environmental Conservation], 2009). But in general, future well densities should decline because multiple horizontal wells can be drilled more economically from a single well pad: in Marcellus shale, a vertical well may be exposed to no more than 15 m of the reservoir layer, while a horizontal well can reach 600–2,000 m within the targeted formation (Arthur and Cornue, 2010).
Power densities of natural gas production (gross energy flux per unit of surface claimed by wells and associated infrastructure for gathering and field processing of gases) are commonly around 2,000 W/m2 and can be an order of magnitude higher in supergiant fields. Perhaps, no other giant natural gas field is as inconspicuous as Groningen: its extraction proceeds at 29 remotely controlled production clusters, each one with 8–12 wells that are arranged in a strip and connected to associated gas treatment plants (NAM [Nederlandse Aardolie Maatschappij], 2009). Each cluster occupies just 11 ha for the total of about 300 ha, and with annual 2010–2020 production capped at 43.6 Gm3 of annual output, this puts the field’s power density of extraction at roughly 16,000 W/m2 (Figure 3.5)
Figure 3.5 A well cluster (one of 29) in Groningen gas field.
Imagery © 2014, Aerodata International Surveys, Map Data © 2014.
3.1.3 Natural Gas Processing
No processing would be required if raw natural gas were a pure, or virtually pure, methane with a negligible admixture of N2 and CO2. But such natural gases are exceedingly rare, and as already explained, mixtures of gases and liquids that come out of the ground are far too heterogeneous and contain at least two and often half a dozen compounds that should not be present in gas that is transported by pipelines and burned for a variety of industrial, commercial, and household uses. Processing of raw natural gas ensures the delivery of product that is not either pure CH4 or a perfectly standardized homogeneous mixture of allowed components but a fuel (or feedstock) whose composition meets a variety of prescribed conditions and limits.
Processing of natural gas is far less challenging than the processing of crude oil: refining and production of liquid fuels involve high temperatures, high pressures, and catalytic reactions. In contrast, gas processing is largely a matter of separation, of removing liquids and undesirable (corrosive or incombustible) gases in order to meet quality standards for pipeline transportation and commercial use. These standards prescribe maximum concentrations of constituents other than methane: limits are set for CO2, N2, O2, and water vapor present in the gas, and restrictions on the energy density of transported gas also require the removal of NGLs (condensates), mostly
ethane and propane.
Some simple field processing is done right at or near producing wellheads: gravity separation in horizontal vessels (with water settling at the bottom, liquid hydrocarbons above it, and gases on the top) is the simplest and most common field option, and the first oil/gas separator was in operation already in 1863. But in large gas fields, processing has become a major industry, and raw gas from many wells is led through (small-diameter, low-pressure) gathering pipelines to processing plants. Depending on its composition, the raw gas will be subject to at least two and often all four of these purification processes: removal of oil and condensate, separation of valuable NGL, removal of water, and removal of CO2, any sulfur compounds, and, sometimes, N2 (Mallinson, 2004; IHRDC [International Human Resources Development Corporation], 2014; NaturalGas.org, 2014).
Gas must be dehydrated in order to prevent water condensation and potential corrosion of pipes. This is done mostly by absorption by liquids but adsorption by solids is also used. Liquid desiccants used for absorbing water are either diethylene or triethylene glycol. When in contact with raw gas stream, these hydrophilic compounds will absorb water, and as they get heavier, they sink to the bottom of a contact chamber, and water is then removed by evaporation (glycols have a much higher boiling point, 288°C, than water) and the desiccants are reused. Wet gas is introduced at the bottom of vertical cylindrical (5–8 m tall) columns (contactors), it ascends through glycol solution flowing from the contactor’s top (trays or packing materials are used in order to maximize the contact surface), and dry gas leaves at the top of a contactor.
Water can be also removed by adsorption onto the surfaces of solid desiccants, usually alumina or silica gels placed in adsorption towers. This process is both highly efficient and suited for large volumes of highly pressurized gases and that is why it is used in so-called straddle systems, dehydrators placed on gas pipelines downstream of compressor stations. Dehydration uses two parallel columns in order to allow for the regeneration of the adsorption medium in the column saturated with water. Adsorption is more expensive than glycol absorption, but it is much more effective (it can reach dew points as low as −100°C).
Pipeline gas should be also devoid of any gases whose presence could lead to formation of acids: this requires removal of CO2 and H2S (notable for its “sour” rotten smell) as well as of other sulfur compounds (CS2, COS, mercaptans). The principal process used to remove H2S (“sweetening” the gas) is absorption by amine solutions in a manner similar to glycol dehydration (gas ascending through columns containing either mono- or diethanolamine). Removal by solid adsorbents is also possible, and sulfur is usually recovered by the Claus process (a sequence of oxidation, cooling, reheating, catalytic conversion, and cooling) that produces elemental S to be used in the synthesis of sulfuric acid. Large yellow heaps of the element seen in the Port of Vancouver, ready to be exported to Asia, come from the desulfurization of Western Canada’s sour gases high in H2S (Figure 3.6).
Figure 3.6 Sulfur in the Port of Vancouver.
© Corbis.
Separation of liquid hydrocarbons is an obvious necessity for gases associated with crude oil: some of them have even higher shares of heavier alkanes than nonassociated wet gas rich in higher alkanes. Decreased pressure may accomplish this separation as soon as the mixture of gases and liquids reaches the surface, or the process may be helped simply by gravity as gases and liquids stratify over time inside enclosed tanks. But hydrocarbon recovery and fractionation are needed even for dry gas dominated by methane. There are three main reasons for this: to reduce hydrocarbon dew point and hence to eliminate condensation during the transportation of the gas, to meet the requirements for the prescribed heating value of transported gas, and to separate NGLs and market them as valuable feedstocks (ethane, propane, butane) and fuels (propane, butane).
Low-temperature separators are used to remove light oils and NGLs by taking advantage of different boiling points of alkanes: at atmospheric pressure, methane boils at −161.5°C, but ethane can be separated at −88.6°C, propane at −42°C, and isobutane at just −11.7°C. Removal of light alkanes is done by cryogenic expander process. Pressurized rich (or wet) gas is first precooled by a counterflow of cold methane to about −34°C, and this liquefies all C3+ alkanes; propane (C3H8) and butane (C4H10) are usually marketed as liquefied petroleum gases (LPG) and distributed in refrigerated tank trucks and ships and stored in recognizable large industrial bullet-like tanks and small portable containers to be used as heating and cooking fuels. Both alkanes can be also sold separately as valuable feedstocks (see the next chapter for details). Pentane (C5H12) and any traces of heavier molecules are liquids that are mostly blended into gasolines.
This leaves the mixture of methane CH4 and C2H6 that enters demethanizer column where the cold gas is cooled by using either classical Joule–Thomson valve or a turbo expander: as its pressure drops, its temperature goes below –100°C, lower than the boiling point of ethane, and then gets liquefied (all of it but commonly at least 90–95%). There is also a less efficient absorption process that can recover up to 40% of ethane and 90% of propane as the gas flows upward through a tower filled with absorbing oil; the enriched oil is then led to a distillation column where the NGLs are boiled off and oil is then recycled for repeated use.
The United States has now about 600 gas processing plant, and many central facilities have been built to treat large volumes of raw gas, both near major field and close to major market centers. For example, St. Fergus plant in Scotland processes about 20% of all natural gas from the North Sea (Total, 2014), and Aux Sable processing plant in Channahon near Chicago removes NGLs from natural gas imported from Alberta where the gas is treated to remove water and acid compounds (Figure 3.7). Importance, cost, and economic impact of natural gas treatment vary depending on the composition of extracted gas. Little has to be done with the gas from the West Sole field in the North Sea as it is a dry gas composed of nearly 95% of CH4 and just 3.1% of C2H6, less than 1% of heavier alkanes, 1.1% N2, and 0.5% CO2 and contains no sulfur. In contrast, Qatari gas has less than 77% CH4, nearly 13% of C2H6, almost 3% of heavier alkanes, and 1% of H2S.
Figure 3.7 Natural gas processing plant, Central Alberta, Canada. © Corbis.
3.2 PIPELINES AND STORAGES
Pipelines are expensive and often surprisingly long-lived structures, usually out of sight and operating with exceptionally high reliability—but their safety requires dedicated management of corrosion risks, and their failures can have locally a truly catastrophic impact (AGA [American Gas Association], 2006). Crude oil, oil product, and natural gas pipelines share many commonalities, starting with construction methods and ending with the need for constant monitoring, but they move two very different substances. Crude oil and refined oil products are virtually noncompressible, while natural gas in its ambient, gaseous state can be compressed. Gas moving in pipelines is pressurized to between roughly 1.4 and 10.4 MPa (200–1,500 lbs/in2), that is, at least about 14 times the ambient atmospheric pressure at sea level (101.325 kPa). This means that natural gas pressurized to 1 MPa will have density roughly 10 times higher than at ambient pressure (8.4 vs. 0.85 kg/m3).
And while oil pipelines are a very important component of global oil industry, gas pipelines are simply indispensable. While the modern oil industry (extraction and transportation of crude oil and production and distribution of refined oil products) is highly dependent on pipelines, it is not only possible but often more economical to move large quantities of liquids by other means. The cheapest way to move crude oil between continents is in double-hulled giant tankers, and both intra- and international deliveries can use trucks, railroads, or river barges. Since 2008, the railroad option has actually expanded in North America as construction of new pipelines has not kept pace with rapid increases of shale oil extraction in North Dakota. In contrast, rail, road, barge, and tanker transport cannot be used to transport natural gas at ambient pressure and temperature. Without large-diameter lo
ng-distance pipelines, there would be no economical long-distance transmission of natural gas, and the fuel could be used only by industries located near its sources or by households, institutions, and enterprises in nearby cities.
Another key difference concerns pipeline function. Oil fields have often extensive networks of gathering pipelines; large-diameter trunk lines move crude oil to refineries, to ports, or directly to foreign buyers; and oil product lines carry gasoline, kerosene, and diesel oil or fuel oil to local storages and distributors. In contrast, small-diameter (0.5–6 in. or 1.22–15.24 cm) distribution lines bringing gas to individual consumers are the most extensive segment of the natural gas transportation system: in the United States, their length is now about 3 million km (1.8 million km of distribution mains leading to industrial consumers and 1.2 million km of service lines supplying businesses, apartments, and households). In 2012, the total length of large-diameter (4–48 in. or 10.16–121.2 cm) onshore natural gas trunk lines was 477,500 km, and field gathering lines added up to less than 17,000 km.
By the 1870s, when modern hydrocarbon extraction became a large commercial enterprise, there were extensive networks of distribution lines supplying manufactured gas to businesses and homes in most major cities, but the absence of long-distance pipelines restricted the use of natural gas not only during the last three decades of the nineteenth century but also during the earliest decades of the twentieth century. Wooden lines that were built until the 1870s by using short pieces of hollowed-out pine trunks leaked and could not withstand much pressure. The earliest (low-pressure) pipelines distributing coal gas were made of cast iron, as were the first lines bringing gas to cities from nearby reservoirs, such as the 25 km line from the Maurice River area delivering gas for street lighting in Trois-Rivières in Quebec in 1853.