Power Density

by Vaclav Smil

  Geothermal Energy

  The sources of the Earth's huge heat flux are yet to be accurately apportioned: they include basal cooling of the Earth's primordially hot core and, above all, heat-producing isotopes of 235U, 238U, 232Th, and 40K in the crust (Murthy, van Westrenen, and Fei 2003). At the ocean bottom the heat flux is as low as 38 mW/m2 through very old (at least 200 million years) floors to more than 250 mW/m2 for ocean floors younger than 4 million years. On land, the youngest crust has an average heat flow of 77 mW/m2, the oldest continental shields average less than 45 mW/m2, and the weighted global mean is 87 mW/m2 (Pollack, Hurter, and Johnston 1993).

  Typical terrestrial heat flows are an order of magnitude lower than the flows commonly produced by the photosynthesis of natural ecosystems, too low to be converted to useful energy at the Earth's surface. Drilling gets to progressively higher temperatures: the geothermal gradient (the rate of temperature increase with depth) is usually 25-30°C/km, but the extremes range from less than 15°C/km to more than 50°C/km in volcanic and tectonically active regions, particularly along the plate conversion and subduction zones surrounding the Pacific Ocean (Smil 2008). In selected locations in these regions, hot (above 100°C) pressurized water can be tapped by drilling often no deeper than 2 km and used (directly as steam or indirectly in binary systems where another liquid is heated by hot water in an exchanger) to produce electricity in turbogenerators (Dickson and Fanelli 2004).

  In vast continental regions, geothermal surface and near-surface fluxes are too small and their temperature is too low for rewarding conversion to electricity, but in many locations these flows can supply significant volumes of hot water for household, commercial, and industrial uses, and shallow wells can be used for heat pumps. Of course, high-temperature water can be recovered anywhere after drilling sufficiently deep into the Earth's crust, but in regions with a normal geothermal gradient it would require wells at least 7 km deep, where injected water would contact dry hot rocks. Such enhanced geothermal systems (EGS) would be the most intensive way of geothermal capture as a set of deep wells reaching hot rocks would be connected by circulating injected water that would be heated to more than 200°C and withdrawn for electricity generation (Tester et al. 2006).

  The commercial exploitation of this option is yet to come, and all existing geothermal plants use hot water withdrawn from wells that are relatively shallow (less than 1 km at Italy's Larderello and New Zealand's Wairakei) or of a medium depth (2-2.5 km in Iceland and California); these wells are sometimes recharged not only with fresh surface water but also with treated wastewater. Geothermal electricity generation began at Italy's Larderello field in 1902; New Zealand's Wairakei came only in 1958, followed by California's Geysers in 1960 and Mexico's Cerro Prieto in 1970. Most of these pioneering installations were gradually enlarged, and other countries began to develop their considerable geothermal potential. By the year 2010 the worldwide geothermal electricity-generating capacity had reached 10.7 GW, with nearly 3.1 GW installed in the United States, 1.9 GW in the Philippines, 1.2 GW in Indonesia, 958 MW in Mexico, 843 MW in Italy, 638 MW in New Zealand, and 5 75 MW in Iceland. The total annual generation was about 67 TWh, resulting in a capacity factor of just over 70% (Bertani 2010; IGA 2013b).

  Wells in geothermal fields could be spread over areas of 5-10 km2, but (much as in the case of hydrocarbon extraction) well sites will claim only about 2% of the field's area, and this claim can be further minimized with multiple wells drilled directionally from a single pad. Many gathering pipelines transporting hot water and steam are mounted on high supports and do not preclude grazing (or even crop cultivation) and the movement of wild animals, and the plants themselves (including generator halls, cooling towers, auxiliary buildings, and switchyards) are fairly compact: the holding ponds for temporary water discharges during drilling and field stimulation may be the largest land claim component. But pipelines and roads may cause some deforestation, and they obviously contribute to fragmentation of habitat or (in some locations) to increased erosion and a higher likelihood of landslides.

  The Geothermal Energy Association put the aggregate claims for the thirty-year operation of a typical geothermal plant at 404 m2/GWh, that is, 283 W/m2 (GEA 2013). DiPippo (1991) put the land requirements of a 110-MW geothermal flash plant (excluding wells) at nearly 14 ha, or almost 800 W/m2 in terms of installed capacity; for a smaller (20-MW) binary plant (also excluding wells), his rate was about 700 W,/m2, and the power density for a 56-MW flash plant, including all other infrastructures (wells, pipes, roads), was about 135 W,/m2. Finally, McDonald and co-workers (2009) put the highest power density (most compact projects) of geothermal power plants (based on American data) at 113 W;/m2 and the least compact at just 8 Wi/m2.

  As with other kinds of electricity-generating stations, the land claims of geothermal plants are readily assessed by accessing satellite images; in addition, we have detailed land-use studies for some major geothermal fields. California's Imperial Valley has some of the most compact facilities located amid crop fields: a 40-MW Heber binary plant (using hot water to heat another working medium, pressurized liquid with lower boiling point to run the turbine) claims 12.15 ha (for a power density of about 330 Wi/m2) and a nearby 47-MW Heber double-flash plant (with hot water vaporized in sequence under low pressure) occupies just 9.5 ha, for a power density of almost 500 W;/m2 (Tester et al. 2006).

  Koorey and Fernando (2010) examined New Zealand's geothermal projects (14 plants, with a total installed capacity of 890 MW) and found the lowest power density of about 70 W;/m2 at Ngawha 2008 and the highest one at Poihipi at 670 W;/m2. Wairakei-the country's largest installation, with 165 MW, using flash steam with a binary cycle-rated 133 W;/m2; and the national mean was nearly 200 W;/m2. All of these rates are for directly affected areas, that is, the land exclusively occupied by the powerhouse, other plant installations, and pipe routes. The addition of affected areas, arbitrarily defined by the authors as all land 100 m out from the development area, lowers Wairakei's power density to just 40 W;/m2 and the national mean to 45 W;/m2, but this count exaggerates the actual impact, as about 40% of affected areas are used as pasture and 14% are forested, an excellent example of concurrent land use.

  Iceland's largest geothermal development, the Hellisheidi combined power plant (303 MWe) and heating plant (133 MW,, to be expanded to 400 MW,), taps into the Hengill volcanic system with wells up to 4 km deep (Gunnlaugsson 2012; fig. 3.8. The generating plant is supplied by 21 wells, each one supporting 7.5 MWe, and four drill holes are on a platform of just 1,200 m2. The output density of hot water is thus 25,000 W,/m2, but considerable infrastructure is required to turn that flux into electricity. The entire project-including access and service roads, hot water production and freshwater injection wells, hot water, water, and steam pipes, steam separator stations, powerhouses, cooling towers, steam exhaust stacks, water tanks, discharges, and a switchyard system, injection areas, and connection to the power grid-covers 820 ha, and with 2.3 TWh of annual generation, its capacity factor is nearly 87%. When only actual electricity generation is considered, Hellisheidi's power density is just 32 We/m2; adding the available 400 MW, of heat raises that figure to about 660 MW of useful energy (electricity and heat) and implies an overall power density of about 80 W/m2.

  My final example is the world's largest concentration of geothermal plants, in the Geysers area of the Mayacamas Mountains north of San Francisco, where 22 sites with a total capacity of 1.61 GW tap hot water from the productive area of 7,690 ha (BLM 2010). This implies an average power density of just 21 W;/m2, but most of the land is undisturbed mountains, forests, and meadows. The Calpine Corporation operates 15 of the Geysers projects with about 700 MW of installed capacity, and its infrastructure includes nearly 240 km of steam and water injection pipelines and about 270 km of access roads (Calpine Corporation 2013b). Even with generous average claims (3 ha/plant, 5-m ROW for lines and 10-m ROW for roads), the overall claim comes to about 1,250 ha, and the power density rise
s to 55 W;/m2, very similar to that of New Zealand's Wairakei plant.

  Figure 3.8

  Hellisheidi geothermal plant in Iceland. UN Photo/Eskinder Debebe.

  In the future enhanced geothermal systems-where high-pressure injections of water and chemical into vertical wells (in a process akin to hydraulic fracturing in oil and gas production) create new fractures in hot rocks-will be able to either boost the productivity of some existing fields or allow the production from otherwise uneconomical reservoirs (USDOE 2014). Eventhe lowest power densities, those between 20 and 50 W;/m2, are of the same magnitude as those of alpine hydropower stations and an order of magnitude higher than those of the best large wind farms. Of course, these comparisons ignore different qualitative aspects of these three kinds of claims. In addition, and as in the case of underground coal mining (to be noted in the next chapter), there may be also surface impacts following continuous withdrawals of hot water and reservoir recharging. With outflows much larger than recharges, surface subsidence takes place in many geothermal formations: in parts of the Wairakei field it was as much as 45 cm/year (Allis 1990). Subsidence and landslides should not be a problem with EGS operations, but some of them may experience induced seismicity.

  While geothermal energy will remain a globally marginal source of electricity-its share was a mere 0.33% of the global supply in 2010 (Bertani 2010)-it can provide significant shares of household, commercial, and industrial heat at local and regional levels. More than sixty countries are now harnessing this resource. The US capacity of direct heat uses more than tripled between 2000 and 2010, from about 3.8 GW, to 12.6 GW,; other large decadal increases have been in the Netherlands (from just 308 MW to 1.41 GW), Sweden (from 377 MW to 4.46 GW), and China (from 2.3 GW, to 8.9 GW,), and the global capacity increased from about 15.1 GW, in the year 2000 to 50.6 GW in 2010, while the actual energy use more than doubled, to 438 PJ, implying an average utilization factor of about 28% (IGA 2013b).

  Household heat pumps-with relatively small individual capacity (on the order of 10 kW of heat for a typical American house) but with large aggregate numbers owing to their increasing popularity in affluent countries (in the United States, their total has surpassed one million)-are the single largest category of these direct uses (Lund et al. 2003; Navigant Consulting 2009). The space requirements of these installations (wells, piping, sometime also water tanks and heat exchangers) for commercial uses are minimal and easily accommodated within the fenced facilities. A typical American house will need three holes (up to 150 m deep) at least 6 m apart, or at least 100 m2 with access space, to produce a heat flux of 100 W/m2. Horizontal closed loops buried at a depth of 1-2 m cover at least 250 m2 (a square of 15.8 m, easy to accommodate in most US suburban house lots but too large for older, densely spaced urban housing) and yield a heat transfer power density of 40 W/m2.

  Fossil fuels are enormously concentrated transformations of biomass, and hence the power densities associated with their extraction are unrivaled by any other form of terrestrial energy. But gathering, preparing, processing, and transporting these energy sources require a variety of permanent and temporary infrastructures that necessarily dilute the power densities of their delivery for final conversion to heat, motion, and electricity.

  All fossil fuels are abundant, and although they are obviously finite (for more than a century the rate of their consumption has been surpassing the rate of their formation by many orders of magnitude), there is no imminent global danger of running out of coal or hydrocarbons. Inevitably, the progressive extraction of fossil fuels led some nations to shift from poorer, and more expensively produced, domestic resources to cheaper imports originating in giant surface coal mines and giant oil and gas fields, often on different continents. Many coalfields, some of them centuries old, have been entirely abandoned as underground mining of relatively thin seams became uneconomical, and an increasing share of hydrocarbons is produced from greater depths, in locations more remote from principal markets or in deeper offshore waters. All varieties of fossil fuels have thus became more expensive, but also more widely available and globally traded on massive scales, and their aggregate extraction is still increasing.

  The genesis of fossil fuels explains the often extraordinarily high densities with which these resources are stored in the uppermost layers of the Earth's crust. Coal seams are lithified and compressed layers of ancient phytomass. As most of the original oxygen and hydrogen in the phytomass were driven away by long spans of pressure and heat processing, generations of those ancient swamp forests were concentrated into layers of carbon. Some have remained remarkably pure, others were later adulterated with incombustible rock, and the younger coals still contain plenty of moisture. Relatively large shares of the carbon that were present in ancient phytomass are preserved in coals, up to 15% for lignites and at least 10% for bituminous coals (Dukes 2003). Actual recovery rates of this ancient carbon depend on the mining techniques (discussed later in this section), but the global mean is around 10%. When inverted, this share means that about 10 units of ancient plant carbon (no less than 5 and as many as 20) yield one unit of carbon in extracted coal.

  The genesis of hydrocarbons has been entirely different: mixtures of liquid compounds accompanied by lighter gases originated through bacteriogenesis (anaerobic microbial metabolism) and thermogenesis (heat decomposition) of nonhydrocarbon organic molecules (carbohydrates, dominated by cellulose, proteins, and lipids) produced during the eras of high photosynthetic activity and buried in marine or lacustrine sediments; the combination of these kerogen-forming processes can be dated in some shales as far back as 3.2 billion years ago (Rasmussen et al. 2008). Sedimentary source rocks may contain as much as 10% kerogen by mass, but a 1%-2% share is more common. Their prolonged heat decomposition produced liquid hydrocarbons, most of which were subsequently degraded by slowly acting thermophilic bacteria, which convert lighter fraction to denser oils.

  The formation of hydrocarbons preserves much lower shares of the original organic carbon than the genesis of coal, often less than 1% and only exceptionally more than 10% during the formation of oil-bearing sediments. After bacterial and heat processing this figure drops by at least another order of magnitude, and usually only a small fraction of mobile hydrocarbons migrates from the source rock to oil and gas reservoirs, from which anywhere between 25% and 50% of fuels in place could be recovered. Using, again, inverted rates, this means that on the order of 10,000 units of carbon present in the original biomass will be required for every unit of carbon in extracted crude oil, and 12,500 units are needed for every unit of carbon in natural gas (Dukes 2003).

  Some natural gases require minimum processing before combustion (stripping H20, H2S, and other trace gases); others contain a relatively high share of natural gas liquids (ethane to pentane), which are removed and sold separately. Crude oils must be refined to produce the most desirable transportation liquids, including gasoline, kerosene, and diesel fuel. Inevitably, refining further lowers the overall carbon recovery factor. Dukes (2003) summed up this low rate of carbon transfer by noting that every liter of gasoline (that is, about 640 g of carbon) requires about 25 t of initially sequestered marine biomass, or at least 12 t of carbon, corresponding to a transfer rate of a mere 0.005% and the requirement of roughly 20,000 units of initial biomass carbon to produce a unit of carbon in gasoline.

  But as a result of the long periods of ancient biomass accumulations, even these low transfer rates translate into often enormous energy storage in coal seams and in naturally pressurized oil and gas reservoirs, and huge amounts of oil remain bound in their source rocks, mostly in sands and shales. This means that even though some forms of fossil fuel exploitation are relatively space-intensive-notably the surface mining of coal buried under thick layers of overburden and the pumping of oil from large numbers of low-productivity oil wells that also require many access roads and gathering pipelines-the typical power densities of coal and hydrocarbon extraction are high, and the enterprises producing
these fuels claim relatively small amounts of land.

  Coal Extraction, Preparation, and Transport

  Coals form a rather heterogeneous group of fossil fuels, much more diverse than crude oils and natural gases. Considerable ranges apply to the presence of moisture (from less than 1% to more than 40%), ash (incombustible minerals, from less than 1% to more than 40%), and sulfur (from a trace to more than 5%), and even energy density, one of the two key variables that determine the power density of coal resources, has more than a threefold range. Anthracites (essentially carbon with minimal impurities) have an energy density (all rates are higher heating values) of about 30 GJ/t, high-quality bituminous coals range from 24 to 30 GJ/t, subbituminous coal (commonly used for electricity generation) ranges between 18 and 24 GJ/t, and, owing to their high moisture and ash content, lignites (brown coals) span an even wider range, all the way to less than 10 GJ/t (Smil 2008).

  The other key variable in appraising coal's recoverable deposits is the thickness of coal seams, which ranges from less than 30 cm, too thin to be exploited by modern mining methods (but commonly extracted with the help of simple tools by hard and dangerous manual labor in traditional underground mines), to stunning near-surface accumulations of coal, where single seams may, as in Victoria's Latrobe Valley, be more than 150 m thick, and stacked coal seams may add up to 250 m (Australian Government 2012). In contrast, modal values for seams that can be extracted only by underground methods are only between 0.5 and 2 m. For example, the northern Appalachian Pittsburgh coal bed has extensive blocks up to 1.9-2.4 m thick (USGS 2013). The combination of these two variables, energy density and the thickness of coal seams, determines the theoretical extremes of the power densities of coal extraction. A thin, 1-m-thick seam of German Braunkohle, low-quality lignite with an energy density of 8.5 GJ/t and a specific density of 1.2 t/m3, stores just over 10 GJ/m2, while a 30-m-thick seam of high-quality bituminous coal (energy density of 28 GJ/t, specific density of 1.35 t/m3) will contain about 1.1 TJ/m2.