Zero Carbon Australia, M. Wright and Hearps, Univ. of Melbourne
Energy Institute, August 2010.
The ZCA report argues that Australia could run entirely on
renewable energy by 2020.
I think this lengthy and detailed report is a valuable contribution to the energy discussion, containing much up to date information and many ideas and proposals that are promising. I believe it heads in the general direction Australia should take. However I think the report is quite mistaken in its optimism, that is in its conclusion that Australia can convert to renewables. This conclusion is based on a number of assumptions, some of which seem to me to be highly challengeable.
My current understanding of the global (not Australian) situation is summarised in “Can renewables etc solve the greenhouse problem — The negative case.” Energy Policy, Aug. 2010, (which I will refer to below as CAN). ZCA has helped me to see some mistakes in my analysis, which will enable improvement of my current attempt to apply the general approach to the Australian situation.
The issue is of course absolutely crucial and ZCA has made a significant contribution to the process whereby we try to sort out the situation. ZCA concludes that there is a neat and simple way growth and affluence society can be made sustainable, quickly and affordably. My view is that global problems cannot be solved within or by a society committed to affluence and growth. The potential of renewables is a central issue in this gigantic debate, (although it is not crucial, that is, even if renewable energy turns out to be sufficient there are many other reasons for scrapping the core systems in this society, especially to do with global economic injustice.
The spectacular conclusions ZCA arrives at are largely due to the energy supply target set, which is very low. Present Australian final or end-use energy consumption is 3900PJ/y, and ZCA says this can be reduced to 1660 and kept there. (Fig 4.1 represents the task as supplying an average of 1317 PJ…35 GW, and the wind and solar thermal plant requirements seem to correspond to this lower figure.)
In recent years Australian energy consumption has been growing at over 2% p.a.;, although ABARE expects this to fall to 1.9% p.a. by 2030. ZCA notes electricity consumption is growing at 3.15% p.a., and transport energy use is growing at a similar rate. A 2% p.a. growth rate indicates that demand will double by 2045. In my view the appropriate target to begin a discussion of the Australian energy problem is the likely 2050 demand. In other words as I see it the discussion should begin by focusing on supplying around 8000+ PJ/y in 2050 if business as usual continues, and then consider how conservation effort and new ways might reduce this.
It therefore seems to me quite inappropriate for ZCA analysis to take their target as meeting present demand of 3900 PJ. Their argument is only that this 2008 demand can be dramatically reduced, and then met by renewables. But what about the forces that are driving demand increase all the time and will continue to do so as there is continued pursuit of economic growth and higher “living standards”… They do not discuss the 2050 task at all, beyond about two sentences, the main one being “Ongoing per capita efficiency gains of 1 – 1.3% p.a. after 2020 keep total demand steady at least to 2040, while allowing or population growth.” (p. 15.) No support is given for this statement. On p.130 they say their plan “…intends to decouple energy use from GDP growth”, with no further comment. The implied assumptions are astronomical, i.e., apparently that the factors presently driving over 3% p.a. growth in electricity and transport energy demand at will cease to operate from now on.
The report assumes that the general efficiency of energy use can be reduced by 20%. I have not been able to find in the international literature more than a few scraps of useful data on what is likely to be achieved re energy efficiency improvements, although there are many unsupported claims. (In CAN I assumed a 33% reduction.) The 2050 task for ZCA therefore seems to me to be dealing with a 6400 PJ/y BAU supply task.
ZCA discusses many valuable ideas whereby this BAU demand might be reduced, such as moving from petrol to electricity for cars. Their major strategy is to assume that the whole economy can be more or less totally shifted to electricity, and that this can come from wind and solar thermal systems. They discuss at length many functions that can be shifted to electricity, but they do not show what proportion of present energy demand can be shifted. This requires good data on the forms and uses of energy in the Australian economy, which I have not been able to get, e.g., how much how goes into water heating, space heating, furnaces… (ABARE doesn’t have such breakdowns, and Ayres, 2008, says they are not generally available.)
At present only 20+% of final energy used is in the form of electricity, so for the ZCA case a great deal depends on how much of the remaining 80% can be shifted, and this is not established. In CAN it was noted that even if 60% of transport is shifted to electricity and electricity was assumed to be 25% of final demand, that left 55% of final energy to be supplied in non-electrical form. CAN assumed 10% of final demand in the form of low temperature space and water heating would come from solar panels without difficulty (although this isnot true in Europe and northern US) and so this was left out of the analysis. (ZCA deals with heating via heat-pumps, which seems sensible but this adds to their electricity task an amount CAN ignored … and they don’t show that heat pumps can do the job in Victoria which sets the main problems via its winter heating demand. (Mackay, 2008, says heat pumps can’t do it in UK, but Victoria sets a less difficult task.)
ZCA assumes light and heavy trucks and all cars can be run on electricity. My understanding is that big trucks and long distance trips are not going to be conveniently battery powered, in view of energy density, the recharging task, etc.
ZCA assumes greatly reduced use of cars but nevertheless it seems to me from Mackay’s analysis (2008) that the rate of battery use and therefore lithium per capita would exceed that which all the world’s people could have.
ZCA assumes most travel and freight transport can be diverted to rail. This is problematic. Can rail get individuals and consignments to their unique desired destinations well enough given present expectations re convenience and time. It would take a huge increase in existing infrastructures to make this a satisfactory way for most people to get to where they want to go. ZCA also assumes long distance truck transport will be powered by overhead electricity lines, but the feasiility and infrastructure cost of this is not discussed.
The ZCA travel solution goes beyond the realm of energy technology; it assumes a major change in social attitudes and preferences, i.e., whereby most people would be willing to accept small cars and the greater inconvenience of traveling mostly by rail and bus. The feasibility of these changes can’t be assumed. If for instance people were willing to use community fridges rather than have one in every house we could cut a lot off energy demand, but proposals like that are of a different kind to merely technical proposals like switching cars from petrol to electricity. At least it seems that we should focus first on the technical energy possibilities given the preferences people have at present, then separately deal with how likely it is that we can change preferences how much. (Nicholson and Lang recognize the socio-political problem here, 2010.)
The main problem the transport analysis seems to involve is to do with improved energy efficiency. ZCA claims changing from petrol to electricity would enable a 5 fold improvement but the figures given refer only to what can be achieved by electricity coming out of the battery, and do not include the losses in getting the energy from the windmill into the battery. Bossel (2004) shows that when distribution losses, battery charging etc. are included there is a 50% loss between windmill and wheels. It seems then that when this is combined with a reduced estimate of the proportion of transport that can be electrified a more appropriate gain from electrifying cars would be closer to half that ZCA assumes.
Another problem for electric transport concerns Life Cycle Analysis of the new vehicles. Mateja (2000) says that when the need for strong but light exotic materials is taken into account an electric car takes a surprising amount of energy over its lifetime, even compared to large petrol-driven cars.
Another challengeable claim is that car batteries can be charged when wind and/or sun are at their best. This seems mistaken because car batteries have to be fully charged by the time people want to drive to work or back, whether or not it’s a windy day. Smart grids etc. could get some storage out of the transport fleet but it is not clear how much. (For more detailed discussion see Trainer, 2008.)
Thus ZCA’s transport claims seem questionable. We can surely go a long way in the direction they advocate, but how far? At present transport constitutes about 33% of final energy use. If by 2050 this is still the case, and ZCA proposals can cut it in half, we’d still need 1,335 PJ/y for transport, compared with the 320 PJ/y ZCA concludes.
Issues on the supply side.
Again it is important to note that the ZCA strategy assumes almost all energy can be provided in the form of electricity. Therefore they are in a position to say that wind and solar thermal can provide most of it.
Wind is assumed to provide 40% of energy needed. Lenzen’s review (2009) concluded that in general problems of integration limit wind to a 20-25% contribution. ZCA argue from the Danish situation, and this is not valid. It is well known that Denmark’s situation is unusual, being a very small nation close to large nations and therefore able to export surpluses and import electricity when the winds are down, and to draw on the regions considerable hydro power when necessary. Some good wind regions might be able to integrate much more than 25% of supply from wind, but 40% cannot be assumed as a valid figure for Australia. Denmark produces wind energy equivalent to about 20% of its consumption, but only about 4% of its demand is met by its wind energy (and has a surprisingly high CO2 emission rate.).
Similarly ZCA claims that 15% of wind supply can be regarded as ”firm” or a constant minimum which can be thought of as “base load”. It is difficult to understand how this can be said when the graphs given (e.g., 4.1) show wind’s contribution falling to around zero from time to time, from an assumed national supply system. Several studies have shown that from modeled systems spread over large areas wind’s contribution would often fall to around zero. (Oswald, Raine and Ashraf-Bull, 2008, Coelingh, 1999, Sharman, 2005, and Davey and Coppin, 2003.) That means that no matter how much wind capacity you built you would have to be able to completely substitute for it by other sources from time to time.
The basic electricity supply strategy ZCA offers is to use all the wind energy all the time, then top up with energy from the solar thermal system, drawing on the 17 hour storage built into each plant. It is claimed that very little hydro and/or biomass would be needed to keep supply from these components up to demand.
It is puzzling why CSA proceeds as if 17 hour storage of heat in the solar thermal system is enough. What if the next day is cloudy? What if four days in a row or two weeks are pretty cloudy? (The US stand-alone systems I have read about are designed to cope with six cloudy days in a row.) The main problem for the plan is to provide for Victoria, and the nearest solar thermal plant assumed is at Mildura. Bureau of Meteorology data show that in winter Mildura has more than 11 days a month “cloudy” and only about 7 days a month “clear”. Provision to store for 4 days would require about 5.6 times as much storage capacity as they assume. Even in Central Australia ARDHB (2006) tables show there is a considerable probability of more than a 4 day run of cloud in winter, and longer runs occur. A 4 day task means storage loss would be 4% of energy collected. There are about ten other factors that tend to reduce output which are not apparently included in ZCA’s performance figures (e.g. absorbers being cooled by wind), and it would add to dollar costs. (For my 30 page attempt to determine the potential and limits of solar thermal, see Trainer, 2010.)
Yes greater use could be made of biomass to deal with extended periods of poor wind and sun, but the accounting has not been given. If biomass had to meet the equivalent of demand on 40 days a year given my 2050 estimated target then we’d have to harvest about 14 million ha of woody biomass (which is probably possible.)
A more realistic assessment of the storage problem would play havoc with other aspects of the report, such as the claim that 55 GW has been explained as “firm and “reliable”, and the claimed amount of biomass said to be sufficient.
The second major issue in this supply strategy is the huge over-sizing of the solar thermal system that is assumed. Fig 4.1 makes clear that because the variability of wind is so great, fluctuating between providing 40% of demand and around 0%, a great deal of solar thermal capacity has to be on hand to plug the gaps, i.e., to completely substitute for wind sometimes (because the plot shows that 15% is not ‘firm”.). But that capacity goes on producing energy on the days when wind is contributing 40%, so the solar thermal energy is then dumped. The discussion states that some 30+ of solar thermal energy is “curtailed”, but in Fig 4.1 it looks close to 50%.
System over-sizing makes sense as a way of dealing with variability, but the important point these graphs make clear is the great magnitude required. To cope with the variability of wind in a system where it supplies 40% of electricity demand, you seem to need to have built an equivalent amount of extra solar thermal generating capacity, and you will have to dump half of the electricity it generates (or heat energy it collects.) This has major implications for system plant redundancy and investment cost; below.
It is good that the report refers to the embodied energy cost of plant, or pay-back time, but I don’t think the brief treatment is satisfactory and I think the figures are far too low. I have found the literature on this issue is scant and inconclusive and I feel we are not close to a good understanding of the situation. Fig. 2.27, p. 37 seems to say about 1% of a solar thermal plant’s life time energy output would be needed to produce the plant. Lenzen reports 10.7% for central receivers. ZCA ‘s plot seems to indicate perhaps 1% for wind, but ISA states about 7%. ISA’s (2006) figure for PV is more than 6 times the figure ZCA assumes.
The most important unsettled issue re embodied energy costs is whether a full accounting of all ”upstream” factors has been included. For instance in addition to the energy it takes to produce steel it is important to include the appropriate fraction of the energy it took to produce the steel works, etc. Lenzen and Treloar argue that for steel a full accounting more or less doubles the figure arrived at. For PV panels Lenzen et al. (2006) conclude that a full accounting actually trebles the commonly stated pay-back time, making payback time equal to one-third of plant lifetime. I am pretty sure no one has carried out such a “full accounting” for solar thermal plant, and this suggests that even the 10.7% figure is likely to be much too low.
I would want much more reason to be confident about the solar thermal output figure used. Models and theoretical estimates abound, and one can work these out easily, but I cannot get actual performance figures from operating solar thermal plant. The theoretical discussions almost always deal only with annual average DNI, and almost never throw light on performance in winter. (I do have some data for troughs, and it shows that they are pretty useless in winter.) I have been told that the companies will not make performance data available. There are many reasons why the theoretical derivations can be too high because minor factors have not been included in them, (such as the effect of dust on reflectors, heat losses, breakdowns, cooling by winds; again see the list of 13 factors in Trainer 2010.)
The biggest problem with ZCA’s solar thermal discussion is to do with whether plant peak or average output has been used in deriving quantities and costs. It seems that the wrong one has been used.
The need in Fig 4.1 is for enough solar thermal plant to generate a constant 35 GW. ZCA seems to be saying that 156 central receivers each of 217 peak MW and each of 2.65 million square metres collection area, will do this. This would be true if 217 MW is the average output per central receiver, but not if it is the peak output, which clearly seems to be the case.
Sargent and Lundy say that the solar to electricity efficiency of central receivers is under 8% (2003), but they estimate it can be raised to 16% -17.3% in future. (ZCA’s Fig. 2.19 shows 10.4% for the Andersol plant. If we take the 17.3% figure and 2.65 million square metres of reflectors and the average 7.9 kWh/m2/day ZCA assume, then output only averages 143 MW, not 217.
The Mildura plant, where ZCA states winter DNI at only 4.8 kWh/m2/d, would generate around 87 MW, as distinct from 217 MW. So in that region in winter we would need about 3 times as many ST plants as ZCA concludes. (NASA’s radiation data give 3.9 kWh/m2/day for Mildura, not 4.8 kWh/m2/day.) Note that most electricity demand is in the south east of Australia, so most of the plant would have to be in that direction, or significant transmission losses would have to be covered. (ZCA opts not to locate in central Australia in view of the long transmission distances and losses, but most of the locations chosen still involve very long distances and there is no discussion of a loss factor, e.g., Carnarvon, Roma, Charleville, Longreach, Broken Hill and Moree. Over a 3000 km distance losses would probably be 15%; Mackay, 2008.)
Figures on solar radiation from different data sources vary surprisingly. Again the NASA source indicates that for Mildura in winter DNI is only 3.9 kWh/m2, and in addition radiation is not the same every year. NASA says the Mildura figure can be 3.45 kWh/m2/d in winter, some 25% below the figure ZCA takes. In other words the output of the Mildura plant in winter would sometimes be 62 MW, a long way under the 217 MW the report assumes.
The forgoing discussion does not take into account the fact that the winter performance of central receivers compared with their summer performance seems to be even worse than for troughs. What seems to be the only publicly accessible data, from the early SANDIA/NREL projects (e.g., Radosevich, 1988) and Alpert and Kohlb, 1988) shows that in winter output falls to around one-seventh to one-tenth of summer output. (I have been unable to get SANDIA to confirm or esplain these figures.) The geometry of trough and central receivers determines that in winter the average angle between sun, reflector and absorber is considerably worse than in summer. Taking this factor into account would mean that output at Mildura in winter would be even lower than the above estimates indicate. ZCA does not discuss this issue. It means that supply from a solar thermal plant in the South East of Australia to the main demand regions in winter would be around one-quarter of the rate ZCA assues.
Another major area of concern is to do with assumed future costs. Nicholson and Lang (2010) make several important critical points here, firstly that the solar thermal technology assumed is anything but well established and cost reductions should not be assumed, let alone falls of 50% or more. Indeed they provide impressive information on recent large cost rises in alternative technology costs. For instance the cost of wind energy systems has increased markedly in recent years. In the early 2000s around $1500/kW was assumed but ABARE (2010) estimates recent Australian installations are costing $2,900/kW, having undergone a 30% increase in one year. ZCA assumes wind capacity will be about half this amount.
ZCA accept the expectation expressed by Sargent and Lundy (2003) that solar thermal costs will fall to around one-third of the present amount. Such figures can only be educated guesses. Although some supporting reasons are given for them by Sargent and Lundy they are not derived of checkable. The only two commercial plants under construction are estimated by Nicholson and Lang to now be costing $(A)16,400 and $(A)25,000 per kW respectively. They note a large cost increase or Solar Tres. Lang reports the more recent NEEDS study (2009) of solar thermal systems as estimating $(A)17,000 /kW. ABARE anticipates only a 34% fall in solar thermal costs to 2030. ZCA proceeds on the assumption that the cost will fall to c. $4,130/kW overall (or, confusingly, to $3,360/kW if calculated per central receiver, i.e., $739 million/220MW.)
Nicholson and Lang rightly urge caution re adoption of the “learning curve” notion. This applies usually to refinements and increases in scale which come with experience in building more units of an established technology, and neither the 7.5 MW wind or 220 MW solar thermal units ZCA assumes have even been built yet. In the establishment phase of a new technology huge cost over runs are common.
To summarise, my back of the envelope impression is that when the foregoing points are added the ZCA conclusion is out by the following factors:
i. The efficiency gain assumed for electric vehicles should be perhaps halved.
ii. The assumed proportion of travel that can be transferred to electric vehicles is too high, in view of how well people and freight can be got to intended destinations by light vehicles and public transport, and in view of what people will accept.
iii. The embodied energy costs of plant might be much more than 10 times as high as has been assumed.
iv. Far more storage for solar thermal needs to be assumed, perhaps 96 hours, as distinct from 17.
v. The amount of solar thermal capacity might need to be quadrupled in view of the peak vs average capacity issue.
vi. Very optimistic assumptions and estimates have been made throughout, including regarding costs.
Note that all of the above points refer only to achieving the energy supply target ZCA assumes, which I began by arguing is mistaken. If the 2050 supply problem we are heading towards is addressed the (final not primary) target would be 6400 PJ/y, (after taking off ZCA’s assumed general 20% efficiency gain) not 1660 PJ/y (let alone 1330 PJ/y.) That is, the task would be 4 times as great.
Combining clear and confident estimates for all these factors would obviously yield a final multiple of ZCA plant requirements, costs and annual investments that would be many times greater than those ZCA arrives at, and quite unaffordable. The application of the CAN approach to Australia’s situation indicates that even with our good renewable resources we could not afford to depend solely on them. This is not an argument against transition to renewables, which is highly desirable. It is an argument against the possibility of running an energy intensive society on renewables – and therefore an argument for the need to move to The Simpler Way (detailed in Trainer, 2006.)
Alpert, J. L., and G. Kolb, (1988), Performance of the Solar One Power Plant As Simulated by the SOLENERGY Computer Code, Sandia National Laboratoreis, Alberquerque.
Australian Solar Radiation Data Handbook, (2006), ANZ Solar Energy Society, April. Energy Partners, 6260 6173.
Bossel, U., (2004), ‘The hydrogen illusion; why electrons are a better energy carrier’, Cogeneration and On-Site Power Production, March – April, pp. 55 – 59.
Coelingh, J. P., (1999), Geographical dispersion of wind power output in Ireland,
Ecofys, P.O. Box 8408, NL – 3503 RK Utrecht, The Netherlands. http://www.ecofys.com.
Davy, R. and Coppin, P., (2003), South East Australian Wind Power Study, Wind Energy Research Unit, CSIRO, Canberra, Australia.
EPRI (2009): EPRI (2009b). Program on technology innovation: integrated generation technology options; Technical Update, November 2009. http://my.epri.com/portal/server.pt?Product_id=000000000001019539
ISA, (Integrated Sustainability Analysis), (2006) Life-cycle energy and greenhouse gas emissions of nuclear energy in Australia, Dept. of Physics, University of Sydney.
Lenzen, M., (2009), Current state of development of electricity-generating technologies – A literature review. Integrated Life Cycle Analysis, Dept. of Physics, University of Sydney.
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), Internet site http://www.isa.org.usyd.edu.au/publications/documents/ISA_Nuclear_Report.pdf, Sydney, Australia, ISA, University of Sydney.
Lenzen, M and G. Treloar, (2003), Differential convergence of life-cycle inventories toward upstream production layers, implications for life-cycle assessment”, Journal of Industrial Ecology, 6, 3-4.
Mackay, D., (2008), Energy – without the Hot Air. http://www.withouthotair.com/download.html
Mateja, D., (2000), ‘Hybrids aren’t so green after all’, www.usnews.com/usnews/biztech/articles/060331/31hybrids.htm
NEEDS, (2008). Final report on technical data, costs, and life cycle inventories of solar thermal power plants. http://www.needs-project.org/docs/results/RS1a/RS1a%20D12.2%20Final%20report%20concentrating%20solar%20thermal%20power%20plants.pdf
Nicholson, M and P. Lang, (2010), Zero Carbon Australia – A Strategic Energy Plan: A Critique.
Oswald, J.K., M. Raine, H.J. Ashraf-Ball, (2008), “Will British weather provide reliable electricity?”, Energy Policy, 36, 3202 – 3215.
Radosevich, L. (1988), Final Report on the Power Production Pase of the 10 MWe Solar Thermal Concentrating Receiver Pilot Plant, SAND 87 – 8022, NREL, Sandia National Laboratories, Alberquerque.
Sharman, H., (2005), ‘Why UK power should not exceed 10 GW’ ,Civil Engineering, 158, Nov., pp. 161 – 169.
Trainer, T., (2006), The Simpler Way website, http://ssis.arts.unsw.edu.au/tsw