Can the world run on renewable energy? An improved negative case.
16.10.12
Ted Trainer, (corresponding author)
Conjoint Lecturer, Social Sciences, University of New South Wales, Kensington 2052. Australia. (F.Trainer@unsw.edu.au.)
The discussion of climate change and energy problems is generally based on the assumption that technical solutions are possible and that the task is essentially to determine the most effective ways. This view relies heavily on the expectation that renewable energy sources can be substituted for fossil fuels. This discussion improves on an earlier attempt to estimate the investment cost that would be involved in deriving total world energy supply from renewable sources. It is concluded that the investment cost would be unaffordable.
It is commonly assumed that greenhouse and energy problems can be solved by intensified conservation and efficiency effort along with a transition from fossil fuels to renewable energy sources. In addition Stern (2006) and others assert that the cost will be easily afforded.
Little attention has been given to the critical assessment of the potential and the limits of renewable energy. Trainer, (2007) attempted what seems to have been the second critical overview, (following Hayden, 2004). The approach taken in this discussion is the same as that in Trainer (2010a) which explored a probable 2050 world energy supply target that might be met by a combination of energy conservation and renewable energy but improves the analysis in some important ways. Better information on some core issues has recently become available, especially through the modelling provided by NREL, (2010, 2011), and the cost estimates reported by Hearps and McConnell, (2010), AEMO, (2010), Lovins, (2011), Lovegrove et al., (2012) and AETA (2012). This information indicated that contrary to the assumption underlying Trainer (2010a) central receivers are likely to be preferable to big dishes, and more recent cost evidence enables revised but more confident conclusions regarding total system capital costs.
After establishing working assumptions, two critical issues are discussed, firstly to do with whether the quantities of alternative energy capacity needed to meet average winter demand could be afforded, and secondly to do with the implications of solar and wind intermittency for plant quantities and total system capital costs.
This analysis attempts to derive a general global case similar to the particular Australian case set out in Trainer, (2012). Consequently it restates some of the assumptions and derivations in that analysis.
Assumptions.
The main purpose of this analysis is to indicate the value of the approach taken to the derivation of an energy budget, so that future studies can refine this when better data becomes available. The assumptions and derivations are transparent enabling the exercise to be reworked using other assumptions.
The probable 2050 energy target.
The IPCC (2011) anticipates a doubling of demand by 2050. Moriarty and Honnery (2009) report several estimates indicating that world energy demand, presently in the vicinity of 500 EJ/y, is likely to approximately double by 2050. Although the demand growth trend has slowed in recent years, presumably due primarily to the GFC, the doubling assumption will be made here as a convenient benchmark for assessing the effects of differing assumptions.
Mioriarty and Honnery also report that the ratio of final to primary energy is .69. The 2050 target will therefore be taken as delivering 700 EJ/y of final energy. It will be assumed that 2050 Òbusiness as usualÓ energy consumption in the electricity and transport sectors will be the same proportions of projected final energy as they are now in Australia, i.e., 21% (i.e., 147 EJ/y) and 33% (i.e., 233 EJ/y) respectively.
Transmission losses.
Very large scale production of renewable energy, especially via solar thermal and PV farms located at the most favourable regions, will involve long distance transmission. European supply from solar thermal fields will probably have to be via several thousand kilometre long HVDC lines from North Africa and the Middle East. Losses in the vicinity of 15% are likely, along with another c. 7% for local distribution. (Mackay 2008, Czisch, 2004, Breyer and Knies, 2009, NEEDS, 2008, Ummel and Wheeler, 2008, and Jacobson and Delucci, 2011.) However, it will be assumed that losses from long distance plus local distribution will be 15%.
Embodied energy costs.
From the gross output figures for a renewable energy device the amount of energy needed to produce it must be deducted. Estimates need to take into account all ÒupstreamÓ costs, e.g., the energy needed to produce the steel works that produced the steel used in plant construction. These factors can double cost conclusions for steel production. (Lenzen, 2008, 2009, Lenzen and Treloar, 2003, Lenzen and Munksgaard, 2001.) Lenzen, 2008, derives an all-inclusive embodied cost of 6.6% for wind, and 33% for PV. (See also Lenzen et al. 2006.) Hall and Pietro (2011), state even higher figures for PV located in Spain, ranging from 33% to 50% of lifetime output. Crawford (2012) believes that he and Lenzen et al. provide the only estimates for PV attempting to take into account all upstream costs. For wind a 5% cost will be assumed here, and for PV only a 15% cost.
The situation regarding solar thermal plant is more uncertain as it does not seem that satisfactory studies have been carried out. The relatively few studies have indicated an up to 11% cost but elements included and assumptions have varied considerably. (Dey and Lenzen, 1999, Weinriebe, Bonhke and Trieb 2008, Norton 1999, and Vant-Hull 2006, Kaneff 1991, Herendeen 1988, Lechon, de la Rua and Saezes, 2006, Lenzen, 2009, p. 117.) No study taking into account all upstream factors seems to have been carried out. (Lenzen, 2012, Crawford 2012.) The unsettled state of this field prohibits the confident assumption of a value for this discussion but a 10% embodied energy cost will be assumed.
Plant lifetime.
A life time of 20 years for wind (Sharman, 2012), and 35 years for PV and solar thermal will be assumed.
Output rate assumptions.
PV.
If 15% efficient PV panels in large power stations are assumed to be located in the worldÕs best regions, such as Central Australia where total global solar radiation in winter is 7 kWh/m2/day on average (ASRDHB, 2006), then the electricity produced would be 1.05 kWh/m2/day, corresponding to a continual 24 hour flow of 44 W/m2. After deducting a 15% transmission loss and the above 15% embodied energy cost a net 32 W/m2 would be delivered at distance.
Biomass.
Some studies conclude that the global biomass potential is very large, for instance 1,548 EJ/y according to Smeets and Faiij (2007) (reported on p.16 Chapter 2 of IPCC, 2011), but the IPCC report points out that these should be regarded as defining theoretical maxima while achievable yields are another matter. It notes that the total net primary productivity of all vegetation on the planet is only about 1,550 EJ/y, so a realistic estimate of the amount that might be harvested for biomass energy is likely to be a small fraction of this.
The IPCC (2011, c. p. 13) reports estimates of Òtechnical potentialÓ for plantations on arable and degraded land, plus crop, forest and urban wastes/residues averaging around a total of 400 EJ/y. However there are several reasons why the amount likely to be available will be well below this figure.
á The difference between ÒtechnicalÓ potential and a realistic figure which takes into account all the social, economic, political and ecological limiting factors is typically large. For instance Field, Campbell and Lobell (2007) conclude that a global supply of only 27 EJ/y can be obtained, under 2% of the Smeets and Faiij figure.
á The IPCC figures explicitly assume (p.13) no increase in the proportion of forest, grass and crop land taken for the production of food, fibre etc. This is virtually certain to be incorrect given the current and increasing food crisis. It is commonly assumed that food output will have to double. The IPCC stresses that its estimates assume very favourable future conditions for food production, and considerable agricultural technical advance. Smeets and Faiij conclude that under plausible unfavourable future conditions there would be no global biomass energy potential at all.
á Large quantities of carbon would be removed from soils and ecosystems. Patzek (2007) argues that over the long term carbon should not be removed and if it is soils inevitably deteriorate. (See also Pimentel and Pimentel (1997).) In the coming era of probably severely limited availability of petroleum and fertilizers it is likely that agriculture will have to focus more intensively on the organic factors contributing to yields, as distinct from external and artificial inputs, meaning that maximum retention of soil carbon and therefore maximum recycling of crop ÒwastesÓ is likely to become crucial.
á Biomass energy conclusions depend greatly on the assumed biomass growth yield. It would seem that the common biomass energy yield per ha assumption of c. 13 t/ha/y, (evident in the IPCC discussion) is unrealistic as an average for very large scale production. It is easily achieved in good conditions, such as willows on cropland, or forests on good soils with adequate irrigation and fertilizer applications, but very large scale biomass energy will have to use large areas of marginal and/or damaged land. World average forest growth is only 2 - 3 t/ha/y. A more realistic biomass-energy yield figure might be 7 t/ha/y. Even if 13 t/ha/y is assumed, i.e., 234 GJ/ha/y, a 250 EJ/y harvest (the average estimate the IPCC reports for technical potential from arable land) would require more than 1 billion ha, which is far more than is likely to be accessible.
For these reasons it is probable that only a relatively small amount of land should be put into global biomass energy production. It is therefore anything but clear how much biomass energy we should attempt to produce, but it would seem that the figure would be a small fraction of the average of the estimates the IPCC reviews.
Easily overlooked is the fact that the estimates are for primary biomass energy. FultonÕs review (2005) concluded that the net yield of ethanol from cellulosic inputs is likely to be c. 7 GJ/t. Farine et al. (2011) report 6.5 GJ/t. Foran (2008) reports the belief among researchers that future yield could be in the region of 9 GJ/t. However the potential is debated. Patzek (2007) says only two plants are in operation in the world, performance is not made public, and energy analysis indicates that the process will not be viable. For the purposes of the following derivation it will be assumed that biomass ethanol is be produced at a net rate of 7.5 GJ/t, and therefore 48.75 GJ/ha, from c.700 million ha (which is regarded here as unrealistically high.) It will also be assumed that an additional about one-third of biomass energy inputs can come from wastes. The total would be equivalent of a 1 billion ha harvested area. Again this is probably much too optimistic an assumption. The biomass ethanol assumption will therefore be 50 EJ/y.
Hydroelectricity.
It will be assumed that the proportion of world energy supply from hydro electric sources will remain about the same as it is now, and therefore that the 2050 supply will be c. 30 EJ. This is likely to be optimistic in view of the warming and drying effects of the greenhouse problem. In recent drought years Australian hydroelectricity production has fallen more than 50%. Because of the low embodied energy cost of hydro electricity and the usually closer proximity to demand no deduction will be made for these factors.
Wind.
LenzenÕs review (2009) concludes that wind is not likely to be able to contribute much more than 20% of the electricity required within a system, because at higher penetrations integration problems rapidly increase. In unusual situations such as that of Denmark (where large neighbouring countries can absorb surpluses) higher penetrations might be achievable. In this discussion 25% will be assumed.
Not taken into account here will be the fact that globally a large scale use of wind energy would have to assume much off-shore capacity, which is around 2.5 times the cost of on-shore capacity (IPCC, 2011, Annex 111, p. 9, Lenzen, 2009, p. 97.)
The world average wind capacity factor is in the region of 0.23. (Smil, 2011 estimates 0.2.) Mainly because in winter winds are stronger than average the figure assumed here for the discussion of the winter supply task will be 0.38, meaning that a 1.5 MW turbine would generate on average 570 kW. Applying the above 5% energy cost for wind makes the net figure 542 kW. Wind farms are more easily located closer to demand than are large scale solar supply systems which would need to be in deserts to enable a reasonable winter contribution. The combined loss due to transmission and local distribution will be assumed at 10%. Therefore a delivery rate per turbine of 487 kW in a winter month will be assumed.
Solar thermal.
Because solar thermal systems can store energy as heat and thereby overcome to a considerable extent (but not entirely; see below) the intermittency and storage problems which most renewables involve, they are likely to be major contributors. However the (limited) technical and climate data accessible indicates that even in the best locations such as the Sahara and Central Australia winter output will be problematic. Troughs are not likely to be viable given their typically low winter to summer ratio of output, in the region of ¼. (NREL, 2010,2011, Odeh, Behmia and Morrison, 2003, Trainer, 2011.)
The preferable option would seem to be central receivers, although unfortunately commercial operators of central receivers do not make performance data public. The evidence available to Trainer (2007) and (2010a) indicated that Big Dishes with heat storage via ammonia dissociation would be preferable. This now seems to have been mistaken. The estimates in the recent NREL Solar Advisory Model (2010, 2011) seem to show clearly that the best option will be central receivers. This is confirmed by Wood et al. 2012. More importantly, Trainer (2010a) assumed that Big Dish storage via ammonia dissociation but more recent evidence on the very high embodied energy cost of the storage containment ( Dunn, Lovegrove and Burgess, 2012) seems to clearly disqualify this option. (See Trainer, 2011b.)
Trainer (2011b) takes cost and output conclusions mainly directly from the NREL (2010, 2011) example theoretical modelling. This yields the fairly confident conclusion that the 24 hour flow rate of electricity delivered over long distance in winter from the best sites (e.g., Central Australia where winter DNI averages 5.7 kWh/m2/d), net of a 10% embodied energy cost and a 15% transmission loss, but including interest charges, 6 hour storage and dry cooling dollar and energy costs, is likely to be around 23 W/m2 of collection area.
Other renewables.
It will be assumed that although other renewable energy sources might in future become significant contributors, at this stage that seems unlikely (Trainer 2007, 2011a, IPCC 2011, Greenpeace 2010, Jacobson and Delucci 2011, and Harvey, 2011.)
Electric transport.
It will be assumed that around 60% of the 233 EJ/y business-as-usual 2050 transport energy demand assumed here could be shifted from fossil fuels to electricity by use of battery powered cars (if long distance car travel can be included.) Sea transport, heavy road vehicles and aircraft are not likely to be powered by electricity. Rail can be electrified but it accounts for a small proportion of transport energy.
The energy efficiency of electric cars is commonly claimed to be in the region of 4 to 5 times as great as for petrol driven cars. However such figures typically apply to Òbattery to wheelsÓ and do not include losses due to distribution, transforming from 240 volt to 12 volt, battery charging and discharging, discharge from idle batteries, the embodied energy costs of batteries and cars (claimed by Mateja ,2000, to be high), battery lifetime and replacement multiple per car life, and global supply of the relatively scarce element Lithium. If vehicle batteries are intended to store energy for later supply to the grid, equipment for reconversion from 12 volt to mains voltage would impose additional costs and losses. Especially problematic are the dollar and energy costs of Lithium-ion batteries, estimated by Smil (2010) to cost $35,000 and to last around three years. (Jacobson and Delucci 2011 believe future battery cost will be half this sum, and that batteries will last the life of a car.) In view of these uncertainties it will be assumed that the present energy efficiency of cars can be trebled. Bossell (2004) argues that it can only be doubled.
Consequently it will be assumed that the 60% of the Òbusiness as usualÓ 233 EJ/y transport energy budget, i.e., 140 EJ/y, will require 46 EJ/y. Another 40%, i.e., 93 EJ/y, will be required for transport in non-electrical form.
Low temperature heat.
In the absence of clear data it will be assumed that 10% of final energy demand, i.e., 70 EJ/y, will be in the form of low temperature heat supplied by passive solar etc. means, and therefore will not add to the need for electricity generation.
Energy conversion.
Discussions of the potential of renewable energy sources usually tend not to take into account the need to convert energy from forms that are available to forms that are needed, or that can be stored. This would not be so relevant if large scale direct storage of electricity was available. Conversion is typically quite energy-inefficient, meaning that much more primary energy needs to be generated than might appear to be the case. For instance according to Bossel (2004) fuelling transport by hydrogen produced from electricity would require generation of about 4 times the amount of energy that is delivered to wheels. (This aligns with the figures in Harvey, 2011, p.458.)
It will be assumed that where conversion is necessary it will take place via the generation of hydrogen with an overall energy efficiency of 0.5, taking into account losses in electrolysis, compression, pumping and distribution (Harvey also states this figure, 2011, p. 459) but not including the embodied energy costs of the required plant
Energy conservation effort.
Significant reductions in energy supply are likely to be achieved by future improvements in energy use efficiency and conservation. Estimates are varied and uncertain and an attempt to arrive at a confident figure for 2050 is beyond the scope of this discussion. However the evidence discussed in Trainer 2011a indicates that for present purposes a working assumption of a 33% improvement in energy use efficiency would be reasonable, for all but the 60% of transport energy assumed here to be converted to electricity for which a 67% reduction will be assumed.
Summarising the demand situation.
The foregoing assumptions and conclusions are summarised in Table 1. This enables others to assess the derivation of conclusions, and to consider the effects of differing assumptions.
Table 1. Annual Supply and Demand Assumptions.
Demand.
Primary energy demand, 2050 1000 EJ/y
Final energy demand, 2050, assuming a .7 ratio 700 EJ/y
Demand for low heat, temperature (e.g., space and water) 70 EJ/y
Demand for transport, 33% of total 233 EJ/y
Demand for electricity
Direct (21% of final) 147 EJ/y
Transport (assuming 60% of the 233 EJ/y transport
energy) 140 EJ/y
Remaining demand for energy, i.e., other than the above
direct electricity, transport electricity and low temperature
heat; 700 – (147 + 140 + 70) = 343 EJ/y
Demand after applying conservation assumptions:
Direct electricity, 147 EJ/y reduced by one-third to 99 EJ/y
Transport;
Electricity, 60% of transport, i.e., 140 EJ/y, ,
reduced by two-thirds to 46 EJ/y
Remaining demand, i.e., 343 EJ/y, reduced by one
third to 229 EJ/y
-----------------
Therefore totals needed:
Electricity, 99 EJ/y direct + 46 EJ/y transport 145 EJ/y
Remaining, non-electrical energy 229 EJ/y
Totals needed after allocating hydroelectricity, (30 EJ/y)
and biomass (50 EJ/y),
Electricity, 145 EJ/y - 30 EJ/y 115 EJ/y
Non-electrical energy, 229 EJ/y – 50 EJ/y 179 EJ/y
Capital cost assumptions.
Evidence and claims regarding the present and likely long term future construction costs of PV, wind and solar thermal technologies vary considerably and estimates cannot be taken with confidence. Use will be made of reviews studies by (or reported by) the IPCC, (2011; Annex III), Hearps and McConnell,(2011), AEMO, (2010), Jacobson and Delucci, (2011), Hinkley, (2012), Lovins, (2011), Lovegrove et al., (2012) and AETA, (2012). On average these loosely anticipate a 30% fall in capital costs for PV and solar thermal, and a 20% fall for wind, by 2050.
Easily overlooked is the fact that all these cost estimates are based on present materials, construction and energy costs, and in future materials and energy inputs are likely to be considerably more expensive than they are now. Clugston (2012) reports marked recent price rises up to, and during the GFC. Lovegrove et al. (2012, p. 190) note that steel prices rose 75% in the 6 years to 2011. Given that all inputs into production involve energy it would probably be easy to underestimate the total multiplier effect on renewable plant cost that might be brought about by significant increase in energy costs.
Stated cost figures assume construction at relatively convenient locations, but most solar thermal and PV ÒfarmÓ level installations will be in very remote desert regions, such as central and north western Australia, the US southwest, and the Sahara. Lovegrove et al. (2012) say that remote construction will multiply capital costs by 1.3 to 1.4. This factor will not be included in the estimation below.
The following cost working assumptions focus on the quantity of plant needed to provide a unit of energy in winter conditions, net of transmission and embodied energy costs. It has been assumed that these costs include the cost of raising capital although this is usually not clearly stated, and in some cases it is explicitly excluded.
Wind
Because wind power technology might be regarded as relatively ÒmatureÓ, estimates of future cost tend not to be markedly lower than present cost per kWe of capacity. (In fact the AETA estimates for on-shore costs rise.) The estimates of present and 2030 capital cost per kW(p) given by the IPCC 2011, and Hearps and McConnell 2010 for future onshore wind are fairly close to $1,500/kW(peak). No account will be taken of the need to use a large amount of offshore capacity, which at present costs about twice as much as onshore capacity.
Given the above assumptions of a winter capacity of 0.38, and energy costs of 7% for distribution and 5% for embodied energy, a 1.5 MW(p) turbine would deliver 487 kW in winter and would cost $2.25 million, meaning that the net capital cost to deliver electricity at distance would be $4.62/W.
PV.
Unfortunately the estimates accessed for utility scale PV installed systems (as distinct from modules) vary greatly. For present installed utility scale cost the high and low figures found are $8,000/kW(p) and $3,200/kW(p). For future cost the range is from $5,500/kW(p) to $1,060/kW(p). The Wyld Group, (2008) estimates future cost at $2,700/kW and this figure, which is less than the average of the previously stated values, will be used here.
This means that panels costing $2,700 will produce 1kW in 1 kW/m2 global radiation. But the global radiation in Central Australia in winter averaged over 24 hours is 7kWh/24h = 291.7 W. Therefore the area of panels needed to produce the daily equivalent of a constant/average 1 kW output would have to be 3.43 times the area that can be purchased by $2,700, and therefore it would cost $9,256. Taking into account a 15% transmission loss and a 15% embodied energy cost this becomes $12.81/W.
Locating PV on domestic roofs would avoid long distance transmission costs but would mean much lower winter performance for most people in developed countries which are a long way north of the equator. Even the Sydney winter average radiation is under half the Central Australian figure. In addition domestic location involves losses due to poor alignment and maintenance.
Solar thermal.
Trainer (2011b) argues that the most promising solar thermal option will be central receivers and sets out the findings in 8 review commentaries on their likely future capital costs. Despite variation these enable some confidence regarding the figures represented in the NREL Sam example models.
When the NREL figure for present capital cost, $6,580/kW, is reduced by one-third, the most common assumption in Table 1, the cost more or less aligns with those given for future costs by the reviews referred to above. Conveniently, the NREL figures are accompanied by clear information on important factors and assumptions, including field area (1 million square metres), dry cooling, inclusion of interest charges, and 6 hour storage. The above capital cost, the field area and the assumed 33% future cost reduction indicates a future cost per square metre of $460. When combined with the above 23 W/m2 winter output conclusion, the capital cost of the plant required to achieve this output would be $20/W.
Quantities and costs for winter supply.
From Table 1 above the supply task is 115 EJ/y of electrical energy and 186 EJ/y of non-electrical energy, i.e., after the allocation of the available hydro and biomass renewable energy sources. Three possible renewable supply strategies will be explored.
The major problem for renewable energy supply is set by the intermittency of wind and solar energy, leading to the need for large amounts of redundant plant and very high total system capital costs, as will be explained shortly. These problems could be overcome if very large quantities of electricity could be stored. The only plausible strategy for this would involve the production of very large volumes of hydrogen from electrical energy.
From the above discussion of conversion efficiency, to produce the required quantity of non-electrical energy in the form of hydrogen, (186 EJ/y x 2) = 372 EJ/y of electricity would need to be generated. The total amount needed would therefore be 116 EJ/y + 372 EJ/y = 488 EJ/y, corresponding to 15,500 million kW. If divided equally between wind, PV and solar thermal, each would have to provide approximately 5,200 million kW in winter.
Table 2
Meeting Strategy 1 Demand.
Wind:
Required, 5,200 GW
Capital cost assumption for delivery per Watt
in a winter month, $4.62/W
Therefore cost $24.9 trillion.
PV:
Required, 5,200 GW.
Capital cost assumption for delivery per Watt
in a winter month, $12.81/W.
Therefore cost $66.6 trillion.
Solar thermal:
Required in winter 5,200 GW
Capital cost assumption for delivery per Watt
in a winter month, $20/W
Therefore cost $104 trillion
Total cost $195.5 trillion
Average capital cost p.a. assuming 30 year plant lifetime
(...an approximate average for wind, PV and
solar thermal) $6.5 trillion
Percentage of 2012 world GDP, approximately 10%
________________________________________________________________
The 2011 annual investment sum is more than 14 times the early 2000s ratio of rich country energy investment to GDP (i.e., for building and maintaining plant, as distinct from purchase of energy.) (Pfuger, 2003, Birol, 2004.)
There are several major cost factors which have not been included in this exercise but would probably multiply the above cost conclusion a number of times if they could be quantified confidently.
á The above exercise assumes average winter demand but extra capacity would be needed to meet peak demand. This factor can multiply the total system capacity required by 1.3-1.5. (The Australian multiple is 1.8. ABARE, 2010.)
á The exercise has been based on long term average radiation levels but in some winter months solar radiation averages 40% below the long term average for that month. (NASA, 2010, ASRDHB, 2006.)
á The cost of the many long distance transmission lines from deserts where solar plant would have to be located to enable winter supply has not been included. Czisch (2004) estimates that this could add 33% to solar thermal generating plant cost, for relatively short distances such as Egypt to Turkey or Morocco to Spain. Harvey (2011) and Jacobson and Delucci (2011) estimate a global average of c. $.5 per kW-km. However Wood et al. (2012) estimate Australian costs at up to 5.5 times as high. If the $.5 figure is taken then two-thirds of the above 15,500 GW capacity located 2000 km from users would add an approximate $14 trillion for transmission to the above system capital cost sum.
á The capital cost of the biomass production system, (and the generators to produce electricity from biomass in the following strategies), have not been included. These would have to be large enough to deliver a quantity of energy in the form of ethanol almost equivalent to current world electricity consumption.
á The cost of the plant needed to provide low temperature heat equivalent to 1.6 times current world electricity supply has not been included.
á The cost of twice the present global hydroelectric capacity is not included,
á The capital cost of the hydrogen production, distribution and storage system, has not been included. This would have to be large enough to deal with almost the equivalent of the present world energy supply. No account has been taken of the increase in these numbers that would be due to the probable need for liquid hydrogen for some forms of transport, notably aircraft. The hydrogen capacity assumed above is just sufficient to produce, convert, store and deliver enough non-electrical energy to meet daily demand. If the hydrogen storage capacity is to be sufficient to meet total energy demand through several days with little wind or solar energy availability, the scope and cost of the hydrogen component would be multiplied in proportion to the period assumed.
á The cost of the re-generating plant needed to convert stored hydrogen into electricity when sun and wind inputs are too low to meet electricity demand has not been taken into account.
á The embodied energy costs of these additional technologies and infrastructures would have to be deducted, e.g., the energy costs of the materials in and the construction of the long distance transmission lines, the biomass generation systems, the hydrogen supply system, etc.
á Much wind capacity would probably have to be located offshore but the much cheaper onshore wind cost has been used.
á Most of the solar thermal and commercial scale PV would be located in remote desert regions, adding 30 – 40% to construction cost according to Lovegrove et al. (2012).
á The above solar thermal costs assume capacity to operate from storage for only 6 hours, but 18 hour capacity is likely to be required in a 100% renewable world. From above this would require three times typical storage tank volume and a solar multiple of approximately 3 - 4, i.e., the collection field would have to be up to 4 times as large as one capable of running the turbine for the daylight hours. These increases would more or less double plant cost.
á There would be a need for a large amount of redundant plant to deal with the intermittency of wind and solar energy. The forgoing derivation assumes availability of wind and solar energy inputs at constant average winter intensities. That is, the amount of plant represented in Table 2 is just sufficient to meet demand every day in winter if solar and wind energy sources are available at their winter average intensities. Half the time they will not be, and for some of the time they will not be available at all. Thus to enable a constant contribution equal to average winter demand far more plant would need to have been built than Table 2 assumes. This is so even when hydrogen storage is assumed, although the minimally costly solution would probably be to significantly increase hydrogen storage capacity rather than wind and solar generating capacity. (The seriousness of this intermittency and redundancy or back-up problem is discussed in more detail below.)
á Although not an up-front capital cost, the lifetime operations and management costs would add to energy price. These are estimated by the IPCC (2011, p. 8) at 25-33% solar thermal construction cost.
If these factors could be quantified reliably the total capital investment cost would be several times the figure derived above.
Since the wind component is much cheaper than the solar thermal or PV components, it might be thought that the whole task should be given to wind. However the average summer wind capacity factor is well below the global annual average, which is in the vicinity of 0.23 or less. If the exercise is reworked assuming wind provides all the electricity required, corresponding to 15,500 GW, and that the summer wind capacity is 15%, then in summer the capital cost of sufficient turbines would be $181 trillion, corresponding to $6 trillion p.a. In other words this wind strategy would not seem to make a significant difference to the cost situation, mainly because windÕs lower peak capital cost is slightly outweighed by its lower summer capacity. (Note also that the foregoing cost analysis assumes a 50% longer life for wind turbines than is valid.)
Thus there is a strong case that the hydrogen path would not be viable.
It is likely that in future the proportion of electricity in the total energy supply will be increased significantly. Let us again assume that conservation effort reduces the 700 EJ/y final demand by one-third by to 466 EJ/y, that biomass provides 50 EJ/y of final energy, hydroelectricity 30 EJ/y, and 70 EJ/y of low temperature heat can all come from solar panels. The remaining energy supply task would be 316 EJ/y, corresponding to a flow of 10,000 GW. If we assume that all of this demand can be met by electricity (which is not plausible), then electricity would constitute 68% of total final energy supply.
Again LenzenÕs review (2009) concludes that because of the difficulties integrating the highly variable wind resource into supply systems wind is not likely to be able to provide more than 25% of demand, possibly only 20%. Lenzen indicates that the PV limit might be somewhat higher. If the 10,000 GW task is allocated 25%, 30% and 45% to the wind, PV and solar thermal sectors the supply from each would have to be 2,500 GW, 3,000 GW, and 4,500 GW respectively. Applying the reasoning in Table 2 to the winter monthly supply task results in a total annual investment cost of $5 trillion, which is 77% of the cost of the previous hydrogen path. Again this does not include the many omitted items noted above except for those to do with hydrogen generation and handling.
At first sight this is counter-intuitive as the task of generating twice as much electricity as is needed in the form of hydrogen has been avoided. However, firstly this only reduces the amount of electricity that needs to be generated by 35%. Secondly strategy 1 assumed wind, the cheapest of the three renewable, contributing 33% of demand (after biomass, hydro and low temperature heat), whereas in strategy 2 it only contributes 25%.
However, strategy 2 is not viable due to problems of intermittency and storage set by the occurrence of Òbig gapsÓ in solar and wind energy availability. Again strategy 1 simply assumed constant winter average contributions, when in fact energy must be provided when it is needed regardless of momentary wind and solar radiation intensities. There can be periods of several consecutive days when there is negligible wind or solar energy or both, anywhere within continental sized regions. For instance Oswald, Raine and Ashraf-Ball, (2008) show that for the first 6 days of February 2006 there was almost no wind energy generated from Ireland to Germany, and one of these days was the coldest of the year in the UK. Although not reported, it would probably also have been a period of negligible solar energy. Similar documentation is given by Soder et al., (2007), Sharman (2005) and Sharman, Leyland and Livemore, (2012) for West Denmark, Flocard and Bach for several European nations, Mackay (2008), for the UK, E On Netz (2004) for Germany, Davey and Coppin, (2003) and Lawson (2012) for Australia, and LenzenÕs review, (2009.)
An indication of the magnitude of the this intermittency and redundancy problem can be given by reference to two recent analyses put forward by groups who claim that 100% renewable supply is achievable. The study by Elliston, Diesendorf and MacGill (2012) claims to show how 100% of Australian electricity demand could be met by combining varying inputs from renewable. Setting aside other problems evident in the analysis (Trainer, 2012), they conclude that in order to meet an average 25 GW demand the amount of renewable capacity required would be 84 GW, 3.3 times as much coal, gas or nuclear capacity as would suffice.
The second illustration is the modelling study by Hart and Jacobson, (2011) claiming to show that almost all of Californian delivered energy could be produced from renewable sources. However Table 2 on p. 2283 reveals that to meet a 66 GW demand with low carbon emissions no less than 281 GW of capacity would be needed. This would include 75 GW of gas generating capacity which would be required to plug gaps in renewable availability. It would function a mere 2.6% of the time (p. 2283) and it will provide only 5% of annual demand. This means 75 power stations would sit idle almost all the time.
Proposals such as that by Hart and Jacobson typically assume back up by fossil fuel generation This would be far cheaper and more easily organised than back up by renewable sources, which would be necessary if world supply was to be 100% renewable. Also important re long term system capital costs is the fact that coal and gas fired plant lasts twice as long as wind turbines, and using them to back up wind involves stand by idling generating emissions, and ramping up and down which reduces efficiency and increases wear and tear. (Sharman, Layland and Livermore, 2011, p. 8.)
These intermittency and redundancy issues show that the problem is not explaining how renewable sources might meet 100% of demand; the problem is explaining how the amount of plant required could be afforded.
Strategy 3: Use biomass for back up.
The strategy argued by Elliston, Diesendorf and MacGill (2012) for Australia assumes considerable reliance on biomass-gas-electricity generation capacity to plug gaps left by lulls in wind and solar energy availability. This is problematic even though Australia probably has a biomass potential per capita that is 5 times the global average. Thus it is highly unlikely that there is sufficient biomass for this form of back up to be a global solution. The 50 EJ global biomass potential assumed above is in the form of ethanol but probably corresponds to a markedly lower quantity of electrical energy given the combined efficiency of links in the generation chain, such as the production, drying and gasification of the biomass, the generator efficiency, and the return of ash to plantations.
A formidable problem of redundancy would remain. During periods when the sun and wind failed to provide energy sufficient biomass-burning generating plant would be needed to meet total demand, only to remain idle most of the time.
4. Electrify and rely on solar thermal generation and storage.
A fourth conceivable strategy would be to use as much electricity as possible and to solve the intermittency, storage and redundancy problems by relying on the capacity of solar thermal power stations to store heat.
Despite its storage capacity solar thermal generation also suffers an intermittency problem. The discussion of solar thermal potential is typically carried out in terms of annual and at times monthly average levels of solar radiation whereas again what matters most are minima and their frequency of occurrence. Trainer 2011b and 2012 detail evidence on big gap events for solar radiation, e.g., lasting 5 – 9 days (Elliston, 2012), even at the best locations.
At present 7 hour storage is being built into new solar thermal plants, so to get through a four day gap 14 times as much capacity would need to be provided.
The capital cost implications of a strategy of this kind would seem to be clearly impossible. Firstly 6 hour storage capacity makes up about 8 – 10% of plant cost according to NREL, (2010, 2011), and Foran (2008), so four day storage would multiply plant cost by 2.5. More important are the implications of the increased solar multiple required. As has been explained, even enabling 24 hour delivery from a solar thermal plant would require a solar multiple of around 3 - 4, meaning a total plant cost double that of one with 6 hour storage.
The effect on energy price?
The retail cost of electricity in Australian is around 22+ c/kWh. The capital cost of plant plus fuel comes to about 2 cents (AETA, 2012, p.17) and distribution accounts for around half of the retail price. Thus factors such as profit, insurance, taxes and interest charges more or less multiply the capital plus fuel cost of the production of electricity by 6. Expenditure on energy in rich countries comes to about 6 – 9% of GDP, and Murphy (2010) reports that when energy expenditure rises to 10% of GDP recessions occur. The above derivation found that the capital (plus fuel) cost of renewable supply might be 10 times the 2000s cost even though many factors that would multiply this figure several times were not included. If factors such as profit, insurance and taxes further multiplied the figure by 6, the retail price of electricity would be far beyond affordable, and energy would be accounting for far more than 10 % of GDP.
Different assumptions?
The most uncertain assumptions in the above derivation are listed below, along with much more optimistic assumptions.
á The 2050 supply target; assume this is reduced by 25%.
á PV efficiency; assume 20% rather than 15%.
á Future solar thermal cost; assume 50% lower than above, i.e., a fall to 25$ of present cost.
Combining these assumptions would reduce the capital cost conclusion for the hydrogen Strategy by 43%, again not including the many additional cost factors listed above, and not taking into account the problem of dealing with intermittency.
Conclusions.
The total investment sum arrived at above is considerably less than that derived in Trainer (2010a), but the derivation is much more soundly based mainly due to recent access to more confident estimates of output and future capital costs. The general conclusion supported by this discussion is that the capital costs for a totally renewable global energy supply would be far beyond affordable. This means that greenhouse and energy problems cannot be solved by action on the supply side, i.e., by technical developments which promise to provide quantities taken for granted in energy-intensive societies. This general Òlimits to growthÓ perspective is that these and the other major global problems can only be solved by action on the demand side, i.e., by moving to ways, values, institutions and systems which greatly reduce the need for materials, energy and ecological resources.
It should be stressed that the 700 EJ/y supply target would give the worldÕs possible 10 billion people by 2050 a per capita final energy consumption of 70 GJ/y, which is around only one-third of the present Australian level. Thus if renewable sources were to provide all the worldÕs people in 2050 with the present Australian per capita energy consumption, the supply target would have to be three times the quantity taken in this exercise.
This analysis is not an argument against transition to full reliance on renewable energy sources. It is only an argument against the possibility of sustaining high energy societies on them. Trainer (2010, 2011c) detail the case that the limits to growth predicament cannot be solved by technical reforms to or within consumer-capitalist society and that there must be radical social transition to some kind of ÒSimpler WayÓ. This vision includes developing mostly small and highly self-sufficient local economies, abandoning the growth economy, severely controlling market forces, shifting from representative to participatory democracy, and accepting frugal and cooperative lifestyles. Chapter 4 of Trainer (2010) presents numerical support for the claim that footprint and energy costs in the realm of 10% of those in present rich countries could be achieved, based on renewable energy sources. Although at this point in time the prospects for making such a transition would seem to be highly unlikely, the need to consider it will become more evident as greenhouse and energy problems intensify. It is not likely to be considered if the present dominant assumption that high energy societies can run
ABARE, (Australian Bureau of Agricultural and Rural Economics), Energy in Australia 2007, 2009, 2010. Australian Federal Government, Canberra.
AEMO, (Australian Electricity Market Operator),2010. NTNDP Modelling Assumptions: Supply Input Spreadsheets; http://www.aemo.com.au/planning/ntndp2010consult.html,
ASRDHB, (Australian Solar Radiation Data Handbook), 2006. ANZ Solar Energy Society, April. Energy Partners, 6260 6173.
Birol, F., 2003. World energy investment outlook to 2030, IEA, Exploration and Production: The Oil and Gas Review; Volume 2.
Bossell, U., 2004. The hydrogen illusion; why electrons are a better energy carrierÓ, Cogeneration and On-Site Power Production March – April, pp. 55 – 59.
Breyer, G., B. and Knies. 2009. Global energy supply of concentrating solar power, Proceedings of Solar PACES: Berkeley; Sept., 15 – 18.
Clugston, C., 2012. Ever increasing non-renewable resource scarcity, Mother Pelican, 8, 3, March.
Crawford, R., G. J. Treloar, R. J. Fuller, 2006. Life cycle energy analysis of building integrated photo voltaic (BIPVs) with heat recovery unit. Renewable and Sustainable Energy Reviews; 10, 559 – 576.
Crawford, R., 2011a. Towards a comprehensive approach to zero emissions housing, Architectural Science Review; 54. 4, 277 – 284.
Crawford, R., 2011b, 2012. Personal communications.
Czisch, G., 2004. Least-cost European/Transeuropean electricity supply entirely with renewable energies. www.iset.uni-kassel.de/abt/w3-w/project/Eur-Transeur-El-Sup.pdf
Davy, R., and P. Coppin, 2003. South East Australian Wind Power Study; Wind Energy Research Unit, CSIRO, Canberra, Australia.
Dey, C. and M. Lenzen , 1999. Greenhouse gas analysis of solar-thermal electricity generation, Solar Energy, 65, 6, 353 – 368, p. 359.
Dunn, R., K. Lovegrove and G. Burgess, 2012. A review of Ammonia based thermochemical energy storage for concentrating solar power, Proceedings of the IEEE, 100, 2, Feb., 391-199.
El Bassam, N., 1998. Energy Plant Species; their Use and Impact on Environment and Development, James and James, London.
Elliston B., M. Diesendorf and I. MacGill, 2012. Simulations of scenarios with 100% renewable electricity in the Australian National Electricity Market, Energy Policy, 45, 606 – 613.
Elliston, B., 2012. Analysing power system impacts using solar radiation data for Australia, http://www.ceem.unsw.edu.au/content/userDocs/presentation.pdf
E.On Netz, Wind Report. 2005. http://www.eon-netz.com
Farine D et al., 2011. An assessment of biomass for bioelectricty and biofuel and for greenhouse gas emission reduction in Australia, Bioenergy, doii: 10.111/j.1757-1707.2011.o1115/x
Field, C.B., Campbell, J. E. D. and Lobell B., 2007. Biomass energy; The scale of the potential resourceÓ, Trends in Ecology and Evolution; 13, 2, 65 – 72.
Flocard, H. and J. Perves, 2012. Wind production intermittency cross border compensation: What to expect in West Europe? Analysis of winter 2010-2011.
Foran, B., 2008. Powerful Choices, Dept. Of land and Water Resources, Australian Federal Government, Canberra.
Fulton, L., 2005. Biofuels For Transport; An International Perspective. International Energy Agency.
Greenpeace International and European Renewable Energy Council, 2010. World Energy (R)evolution; A Sustainable World Energy Outlook.
Hall, C. and P. Pietro, 2011. How much energy does SpainÕs solar PV program deliver? Third Biophysical Economics Conference, State of New York; April 15 – 16.
Harvey, L. D., 2010. Caron Free Energy Supply, London, Earthscan.
Hayden, H. C., (2004), The Solar Fraud, Pueblo West, Co, Vales Lake Publishing.
Hearps, P. and D. McConnell, 2011. Renewable Energy Technology Cost Review, University of Melbourne;. http://energy.unimelb.edu.au/index.php?page=technical-publication-series
Herendeen, R. A. 1988. Net energy considerations, in Economic Analysis of Solar Energy Systems MIT Press, Cambridge.
Intergovernmental Panel on Climate Change, Working Group 111, 2011. Mitigation of Climate Change, Special Report on Renewable Energy Sources and Climate Mitigation. June. http:www.srren.ipcc-wg3.de/report, p. 10.
Jacobson, M. Z. and M. A. Dellucci 2011. Providing all global energy with wind, water and solar power, Part 1: Technologies, energy resources, quantities and areas of infrastructure, and materials, Energy Policy; 39, 1154 – 1169.
Kaneff, S., 1991. Solar thermal process heat and electricity generation performance and costs for the ANU big dish technology. A comparison with Luz System costs. Report EP-RR-57, Energy Research Centre, Research School of Physical Sciences and energy, Institute of Advanced Studies, ANU, Canberra.
Lawson, M., 2011. Wind power; Not always there when you want it. On Line Opinion, 18th July.
Lechon, Y., C. De la Rua and R. Saez, 2006. Life cycle envioronmental impacts of electricity production by solar thermal technology in Spain, Solar PACES.
Lenzen, M., 2008. Life cycle energy and greenhouse gas emissions of nuclear energy: A review, Energy Conversion and Management 49, 2178-2199.
Lenzen, M., 2009. Current State of Electricity Generating Technologies, Integrated Sustainability Analysis,The University of Sydney
Lenzen, M. 1999. Greenhouse gas analysis of solar thermal electricity generation, Solar Energy, 65, 6, 353 – 368, p. 359.
Lenzen, M., 2012. Personal communication.
Lenzen, M. and G. Treloar, 2003. Differential convergence of life-cycle inventories toward upstream production layers, implications for life-cycle assessmentÓ, Journal of Industrial Ecology, 6, 3-4.
Lenzen, M. and J. Munksgaard, 2001. Energy and CO2 analyses of wind turbines – review and applications Renewable Energy 26(3) 339-362.
Lenzen, M. C. Dey, C. Hardy and M. Bilek, 2006. Life-Cycle Energy Balance and Greenhouse Gas Emissions of Nuclear Energy in Australia. Report to the Prime Minister's Uranium Mining, Processing and Nuclear Energy Review (UMPNER). http://www.isa.org.usyd.edu.au/publications/documents/ISA_Nuclear_Report.pdf, Sydney, Australia, ISA, University of Sydney.
Lovins, A., 2011. Reinventing Fire. Rocky Mountains Inst., Colarado.
Mackay, D. 2008. Sustainable Energy – Without the Hot Air, Cavendish Laboratory; http://www.withouthotair.com/download.html
Mateja D., 2000. Hybrids arenÕt so green after all. www.usnews.com/usnews/biztech/articles/060331/31hybrids.htm
Moriarty, P. and D. Honery, 2009. What energy levels can the earth sustain?. Energy Policy, 37, 2469 – 2472.
NASA, 2010. Solar Radiation Data Base; http://eosweb.larc.nasa.gov/cgi-bin/sse/grid.cgi?uid=3030
NEEDS, (New Energy Externalities Developments in Sustainability), 2008. Final Report concentrating Solar Power Plants; http://needs-poroject.org/RSIa/RSIa.pdf
Norton, B. 1999. Renewable energy – What is the true cost? Power Engineering Journal; Feb, 6 – 12.
NREL. System Advisory Model, (SAM), 2010, 2011 https://www.nrel.gov/analysis/sam/
Odeh, S. D., M. Behnia and G.L. Morrison, 2003. Performance Evaluation of Solar Thermal Electric Generation Systems, Energy Conversion and Management 44, 2425-2443.
Oswald, J. K., M. Raine and H. J. Ashraf-Ball, 2008. Will British weather provide reliable electricity?, Energy Policy, 36, 3202 – 3215.
Patzek, T. W., 2007. How Can We Outlive our Way Of Life? SustainableDevelopment of Biofuels; OECD Headquarters, Paris, 11-12 September, http://www.lifeofthelandhawaii.org/Bio_Documents/2007.0346/LOL-EXH-51.pdf
Pimentel, D. and M. Pimentel, 1997. Food, Energy and Society, University of Colorado Press, p. 241.
Pfuger, A., 2004. World Energy Investment Outlook, International Energy Authority, Berlin.
Sargent and Lundy, 2003. Assessment of Parabolic Trough and Power Tower Solar Technology Cost and Performance Forecasts, NREL.
Smeets, E. and A. Faaij, 2007. Bio-energy potentials from forestry in 2050 -- An assessment of the drivers that determine the potentials, Climatic Change, 8, 353 – 390.
Sharman, H., 2005. Why UK wind power should not exceed 10 GW, Civil Engineering; 158, Nov., pp. 161 – 169.
Sharman, H., B. Layland, and M. Livermore, 2011. Renewable Energy; Vision or Mirage. Adam Smith Research Trust.
Smil, V., 2010. Energy Myths and Realities The AEI Press, Washington, D.C
Smil V. 2011. Energy Transitions, Praeger, Oxford.
Soder, L., L. Hoffman, A. Orfs, H. Holttinnen, Y. Wan and A. Tuiohy, 2007. Experience from wind integration in some high penetration areas IEEE Transactions on Energy Conversion; 22, 4 – 12.
Stern, N. 2006. Review on the Economics of Climate Change, H. M. Treasury, UK, Oct., http://www.sternreview.org.uk.
Trainer T., 2007.Renewable Energy Cannot Sustain a Consumer Society, Springer, Dodrect.
Trainer T., 2010a. Can renewables etc. solve the greenhouse problem? The negative case. Energy Policy, 38, 8, August, 4107 - 4114. http://dx.doi.org/10.1016/j.enpol.2010.03.037
Trainer T., 2010b. The Transition to a Sustainable and Just World, Envirobook,
Sydney. Trainer T., 2011a. Renewable energy – Cannot sustain an energy-intensive society. (http://ssis.arts.unsw.edu.au/tsw/REcant.html
Trainer T., 2011b. The potential and limits of solar thermal power, http://ssis.arts.unsw.edu.au/tsw/ST.html
Trainer, T., 2011c. The Simpler Way website, http://socialscience.arts.unsw.edu.au/tsw/
Trainer, 2012. Can Australia run on renewable energy? The negative case. Energy Policy, 50, 306 – 314.
Ummel, K., and D. Wheeler, 2008. Desert power; The economics of solar thermal electricity for Europe, North Africa and the middle East, Centre for Global Development, Dec.
Vant-Hull, C., 2006. Energy return on investment or solar thermal plants, Solar Today, May/June, 13 – 16.
Vitousek, P. R. Ehrlich, A. H. Ehrlich and P. A. Matson, 1986. Human Appropriation of the Products of Photosynthesis, BioScience Vol. 36, No. 6. Jun., 368-373. www.biology.duke.edu/wilson/EcoSysServices/papers/VitousekEtal1986.
Weinrebe, G. M., M. Bonhke and F. Trieb, 2008. Life cycle assessment of an 80 MW SEGS plant and a 30 MW Phoebus power tower, in Proceedings, Solar 98: Renewable Energy For the Americas, ASME International Solar Energy Conference, Alberquerque, NHM, 13 – 18, June.
Wood, A., D. Mulloworth and H. Morrow, 2012. No Easy Choices: Which Way to AustraliaÕs Energy Future? Technical Analyses, Grattan Institute.
Wyld Group, 2008. High temperature solar thermal technology roadmap Report prepared for the New South Wales and Victorian Governments, Sandringham.