While the main thrust of the Discussion Meeting Issue on ‘Material efficiency: providing material services with less material production’ was to explore ways in which society's net demand for materials could be reduced, this review examines the possibility of converting industrial energy demand to electricity, and switching to clean electricity sources. This review quantifies the scale of infrastructure required in the UK, focusing on wind and nuclear power as the clean electricity sources, and sets these requirements in the context of the decarbonization of the whole energy system using wind, biomass, solar power in deserts and nuclear options. The transition of industry to a clean low-carbon electricity supply, although technically possible with several different technologies, would have very significant infrastructure requirements.
Industry accounts for roughly one-third of the world's energy consumption ; most of that energy today comes from coal, oil and natural gas. What infrastructure would be required to deliver the same amount of energy from zero-carbon electricity? For a couple of possible power sources, I will compare the increase in infrastructure with existing systems, and describe the approximate land area required. (This simple-minded focus on energy neglects the fact that some of the fossil fuels used by industry deliver not only energy but also chemical services—e.g. coal, converted to coke, acts as a reducing agent in blast furnaces.)
(a) Useful units
For concreteness, I will discuss UK industry, but to make the ideas easily transferrable to other countries, I will frequently express energy consumption and production rates in normalized, per-person units. Thus, I will convert national energy consumption rates, which are sometimes expressed in gigawatts (GW) or terawatt-hours per year (TWh y−1), into per-person units: kilowatt-hours per day per person (kWh d−1 p−1). For the UK, with a population conveniently rounded to 60 million, a national power of 1 GW (per UK) is identical to 0.4 kWh d−1 p−1, and 1 kWh d−1 p−1 is identical to 2.5 GW (per UK). For example, the UK's total primary energy consumption is roughly 310 GW, or 125 kWh d−1 p−1, and the UK's average electricity consumption is roughly 42 GW, or 17 kWh d−1 p−1.
Figure 1a shows the primary energy consumption of the UK, in 2007, broken down by energy source, and the electricity consumption in the conventional national units of terawatt-hours per year (1 GW≃9 TWh y−1; 1 kWh d−1 p−1≃22 TWh y−1); and figure 1b shows the identical facts in per capita units.
Although developed countries with different populations such as Germany, Denmark and Switzerland have incomparable national power consumptions in gigawatts, many developed countries have quite similar per capita energy consumptions. (Figure 2 shows on the vertical axis the per capita consumptions of countries in 2005, and on the horizontal axis their population densities.)
So, I will focus on per capita units, but to visualize the assets we are discussing, we may sometimes wish to talk in national units. The following equivalence may prove handy, especially for British audiences. All three of the following powers are equal (near enough) to 1 GW:
— the average electrical output of the Sizewell B nuclear power station (a standard pressurized water reactor with a maximum output of 1.19 GW);
— the average electrical output of all the UK's onshore wind turbines during 2010 (these turbines were 2747 in number, and were grouped in roughly 273 wind farms whose area on a map is roughly 400 km2; their nameplate capacity was 3.8 GW); and
— the energy consumption rate of one blast furnace, as measured by the heat of combustion of the coal it consumes. (One blast furnace consumes 5.8 Mt of coal per year and produces 2.5 Mt of steel per year, which would be enough to make the steel in the 2.5 million new cars per year that join the road in the UK; the UK has five such blast furnaces today.)
(In saying that these three quantities are ‘equivalent’, I am not necessarily asserting that the blast furnace could be powered by 1 GW of electricity instead of 5.8 Mt y−1 of coal; I am simply pointing out that the three average rates of energy flow are identical.)
(b) Power per unit area of wind and nuclear
When visualizing future low-carbon electricity-generation options for the UK, two technologies with substantial technical potential are wind power and nuclear power. (Other technologies such as tidal power, waste-to-energy, deep geothermal power and photovoltaics may also have useful technical potential, as discussed in MacKay .)
The power per unit area of most large wind farms in the UK is between 1.5 and 4.5 W m−2 (figure 3). Of course, the productivity of wind farms depends on their location; some British farms produce less than 2.5 W m−2, and most Scottish farms produce more—perhaps 3.5–4 W m−2; the four offshore wind farms in figure 3 all deliver about 2.5 W m−2. The power production per unit area of most large wind farms in the USA is between 0.7 and 2.5 W m−2 (figure 4), and there is no obvious trend in the data indicating that power per unit area is increasing with time. In the calculations that follow, I estimate the land required for wind farms assuming a typical power per unit area of 2.5 W m−2.
The power production per unit area of nuclear power stations is roughly 1000–2000 W m−2 when the facility is running. Taking into account the time during which the land lies unavailable during decommissioning, and the land associated with waste storage and reprocessing, appendix B shows that the aggregate power per unit area of the first generation of nuclear power facilities in the UK is roughly 140 W m−2.
2. The energy demand of industry
From the Digest of UK Energy Statistics , 18 kWh per day per person (44 GW) is going into industry. (This quantity is a final energy consumption, not a primary energy consumption. In arriving at this quantity, I have included blast furnaces, but excluded the energy consumption of oil, gas and coal extraction, and of petroleum refineries.) Of that energy demand, electricity is already being used by industry at a rate of 5.2 kWh d−1 p−1, so the non-electrical demand amounts to 12.5 kWh d−1 p−1, which is 31 GW per UK. To put that in context, today's total electricity consumption in the UK is 17 kWh d−1 p−1. So as an approximate estimate, if you wanted to electrify industry without making any change to its efficiency, you would need to almost double electricity production. (I have assumed here a one-for-one substitution of coal, gas and petroleum by electricity; i.e. 1 kWh of chemical energy is substituted by 1 kWh of electricity.) Moreover, to make the 5.2 kWh d−1 p−1 of existing electrical supply to industry all low carbon, assuming that 75 per cent of it is not low carbon today, another 4 kWh d−1 p−1 of low-carbon electricity would be needed (10 GW per UK). So, in total, to decarbonize industry in this way, we would need to supply new low-carbon electricity of 16 or 17 kWh d−1 p−1—roughly 40 GW per UK.
3. Some ways to supply low-carbon electricity
Is this possible? Yes. Referring back to §1a, we can deliver an extra 40 GW by adding an extra 40 Sizewell Bs or an extra 40 replicas of the UK onshore wind fleet of 2010, or any mix of those two actions. (I call wind and nuclear power ‘low carbon’ rather than zero carbon to reflect the small quantity of greenhouse gas emissions that are embodied in the construction and maintenance of the infrastructure.)
Forty Sizewell Bs would be four times the UK's current nuclear fleet, which is technically achievable—France, for example, built more than 50 GW in a couple of decades (figure 5). In per capita terms, France and Sweden are both countries that produce more than 16 kWh per day per person from nuclear power (figure 6). The land area for 40 GW of nuclear power stations and their support facilities would be approximately 290 km2, if the nuclear industry's use of land continued in line with the first generation of nuclear power facilities in the UK; 290 km2 is about one-tenth of 1 per cent of the UK's land area, and is equal, for example, to the sum of the areas of the Balmoral and Sandringham estates.
The pure-wind option would involve roughly 130 GW of offshore wind capacity or 150 GW of onshore wind capacity (assuming load factors of 31% and 27%, respectively). The area of sea or land required for the wind farms would be approximately 16 000 km2, which is 6.5 per cent of the UK's land area, or 77 per cent of Wales; the wind capacity divided by the area of the UK would be roughly 0.57 W m−2—about six times higher than the capacity-to-land-area ratio of Denmark (figure 7a). The pure-wind option would imply that the wind capacity per person was about 2200–2500 W, which is about three times that of Denmark (700 W; figure 7b). So a wind-only mix delivering 16 kWh d−1 p−1 would tread where no country has gone before in terms of wind exploitation.
If one opted for a wind-dominated solution, then electricity-balancing services would be needed to deal with the intermittency of the wind: either interconnectors to countries willing to receive and export many tens of gigawatts, or methods for moving equally large quantities of demand in time, or methods for storing extremely large amounts of electricity. To make up for a missing 40 GW for just a single windless day using storage would require about 1000 GWh of storage per UK (16 kWh per person) which is 100 times the energy-storage capacity of Dinorwig or Cruachan, the two largest pumped storage facilities in the UK; and nationwide near-windless periods of four days are not uncommon.
4. Context and caveats
While asserting that both the fourfold growth in nuclear power and the 40-fold increase in wind power over 2010 levels are technically possible, I am conscious that neither of these developments would be judged easy by politicians or engineers. The social, political and engineering challenges are all the greater when we embed the task of decarbonizing industry within the overall goal of decarbonizing society. In the UK, at least two-thirds of our energy consumption is non-industrial—the biggest sectors are transport, space-heating in buildings and non-industrial electricity consumption. The total energy consumption of the UK, remember, is 125 kWh d−1 p−1. Even with radical improvements in energy efficiency, the UK will need several times the 16 kWh d−1 p−1 we have discussed thus far.
Figure 8 shows, to scale, on a map of the UK, four ways of delivering 16 kWh per day per person. The map shows by 160 grey squares, 100 km2 each, the area of the wind farms discussed in the previous section, and it shows by 24 circular dots sufficient sites for nuclear power stations to deliver 40 GW, assuming roughly two Sizewell Bs per site. The map also shows by seven large polygons, some in the UK, some elsewhere, the area of land (80 000 km2) required to create 16 kWh d−1 p−1 of bioenergy, assuming a net power per unit area of 0.5 W m−2. Each of the three largest squares is 22 588 km2 in size, which is the area of New Jersey, and just slightly larger than the area of Wales. And the map shows by eight hexagons the area required in someone else's desert (2700 km2) to deliver 16 kWh d−1 p−1 from concentrating solar power, assuming a power per unit area of 16.5 W m−2, and allowing for losses of 10 per cent between the Sahara and Surrey. If that power were delivered by overhead high-voltage DC power-lines in a strip of land 750 m wide and 1600 km long, the land area occupied in Spain and France by the power-lines would be about 1200 km2.
These four technologies are not the only low-carbon power sources, though they are among the most promising sources with large potential. All the other renewable sources share the property of wind power that they are relatively diffuse: they deliver a power per unit area in the region of wind's 2.5 W m−2. Solar parks, for example, which are appearing across Europe, deliver an average power per unit land area of roughly 4 W m−2 ; and hydroelectric facilities in Scotland deliver about 11 W per square metre of lake area, and about 0.2 W per square metre of catchment area . So whatever the mix of renewables one develops, the land area or sea area required for 16 kWh d−1 p−1 is roughly as indicated by the area of the wind farms. When I talk of the land ‘required’, of course not all the area is literally used up. In a wind farm, for example, only a tiny fraction of the land area is occupied by turbines, foundations and access roads. The rest remains available for agriculture or other uses.
(a) Public discussion of decarbonization pathways
I have emphasized the area required for energy infrastructure, but this is of course not the only important metric. Cost, resilience and air quality are other metrics that may be important in the public deliberation of energy options. The UK Department of Energy and Climate Change has published an interactive open-source tool, the 2050 Pathways Calculator, which allows the user to explore the effectiveness for the UK of different combinations of demand-side and supply-side actions, and which computes and displays several metrics. The UK government's ‘Carbon Plan’, published in December 2011 , illustrates the magnitude of effort required to achieve the UK's 2050 goal of 80 per cent decarbonization within its own borders. The ‘Carbon Plan’ sketches a corridor of pathways in which: per capita demand in the UK falls by between 31 and 54 per cent; nuclear power generation capacity increases from today's 10 GW to between 16 and 75 GW; renewable electricity-generation capacity increases from today's 10 GW to between 22 and 106 GW; carbon-capture and storage electrical capacity increases to between 2 and 40 GW; and bioenergy use increases from today's 73 TWh y−1 to between 180 and 470 TWh y−1 (21–54 GW).
The authors of the 2050 Pathways Calculator would be delighted to see enhancements made to the Calculator's industry module.
(b) This review has not described how to decarbonize industry
This review has explored only the scale of energy infrastructure required to provide a low-carbon flow of (electrical) energy equal to the current high-carbon flow of (mainly chemical) energy into industry. I have not addressed the questions of whether the energy-consuming industrial processes could in fact be electrified, nor what their efficiencies would be when electrified. Moreover, some industrial processes directly emit carbon dioxide because of chemistry, as well as indirectly from their energy consumption—cement production is the most notable example—and this review has not addressed the challenge of eliminating these chemically driven emissions.
Appendix A. Power per unit area of wind farms
Methodology for figure 3: from the monthly statistics published at https://www.renewablesandchp.ofgem.gov.uk/ and collated by the Renewable Energy Foundation , whole-year average outputs were obtained for each wind farm; areas of wind farms were measured by the author from Ordnance Survey maps, which showed the locations of turbines, as illustrated in figure 9. For isolated turbines, the ‘area’ was deemed to be a circle of diameter five times the turbine diameter. For a large farm, the perimeter of the ‘area’ was sketched allowing a strip around the turbines of width equal to half that farm's typical turbine spacing, or 2.5 turbine diameters, whichever was the larger. These data are recorded in tables 1 and 2. Figure 9 shows Red Tile wind farm, which is typical of the British farms represented in figure 3.
Appendix B. Power per unit area of nuclear facilities
Table 3 shows the energy produced by several nuclear facilities in the UK, and their land areas. The average power per unit area of each site is calculated in two ways: first, during operation alone, and, second, taking into account the duration of decommissioning. The more modern sites in table 3 generated about 2000 W m−2 when operating, or about 500 W m−2 on average if we take account of the time taken for decommissioning; the aggregate power density of all these sites, including Sellafield (the largest site, which has hosted other nuclear functions beside power generation), is 140 W m−2.
Nuclear power also has a footprint where the ore is mined, but this footprint is shared with the mining of other useful minerals. The Olympic Dam Mine in South Australia (opened in 1988) produces much of the world's uranium oxide, along with significant quantities of copper, silver and gold. The site has an area of about 20 km2 and is expected to be able to sustain production of roughly 4000 tonnes of uranium oxide per year for 200 years. If the uranium oxide is used in once-through reactors with an efficiency of 1 GW-year per 191 t, the uranium-related power production per unit area of the mine is roughly 1000 W m−2. If the uranium oxide were used in 60-fold more efficient breeder reactors, the power per unit area would be 63 000 W m−2.
One contribution of 15 to a Discussion Meeting Issue ‘Material efficiency: providing material services with less material production’.
- © 2013 The Author(s) Published by the Royal Society. All rights reserved.