An indication of the difficulties.
Ted Trainer.
22.8.2024
The potential and limits to renewable energy are debated and far from settled. Many people take it for granted that it can meet all our energy needs, and numerous impressive agencies and technical reports say this. However until the last few years when a few simulations based on detailed weather data have been carried out nearly all pronouncements have been little more than speculation and most have simply selected bits of evidence to confirm preferred beliefs. I have examined the about ten simulation studies for the Australian situation and have published various analyses showing how inconclusive they are. There are now several analysts and studies casting doubt on the 100% renewable claim. (E.g., Moriarty and Honnery 2010, Clack et al., 2017, Hienberg and Fridley 2016, Heard et al. 2017, Friedmann, 2016, de Decker, 2017, Floyd and Palmer, De Castro and Capellan-perez.) Heard et al., (2017), recently published a scathing review of claims.) For a list of 84 papers see RE100%Doubters.html. About 25 of these were written by me, mostly published in academic energy journals
Just as you could run your house on torch batteries, it is technically possible to run everything on renewables … but that doesn’t mean you could afford to do it. The crucial question is what would it cost to have enough renewable generating capacity to meet demand at any time despite the intermittency of renewable energy?.
What about the claim that wind and PV can produce a kWh more cheaply than coal now? That could be true, but it is quite misleading. It doesn’t tell us much about what a kWh would cost if it came from a system with so much redundant renewable capacity and storage to be able to supply that kWh at any time. Again, that system would have to contain a lot of generating capacity and a lot of storage capacity to guarantee supply through periods when there was little wind or solar input, and that would mean high costs.
How often are solar and wind sources both low?
This is the crucial real-world question. How much generating and storage capacity would we need to get through the bad weather periods that actually occur? I published an analysis of this based on wind and solar generating records for the Australian electricity system. (Trainer, 2023.)
Here is an illustration of the approach used in the study.
This is for a day in which the rate at which the input of wind and of solar generation had fallen to around 6% of maximum output capacity. Wind tends to fall to below this level one or more times a month, sometimes to3%. (See Aneroid records. The solar value averages the low daytime input for June 11 -24 over 24 hrs.) Let’s see how much generating capacity would be needed to meet demand without using batteries.
Australian electricity demand is about 25 GW, or 600 GWh per day. To meet this via wind and solar input when their capacity factor was down to 6% would require 416 GW. The capital cost of generating plant capable of producing 1 kW is about $1,500. So the total cost for this capacity would be about $625 billion. If it lasted 25 years the cost p.a. would be about $25 billion p.a. The annual output would be about 8760 x 25 GWh = 210,240 GWh = 210,240 million kWh. So the the capital cost per kWh produced would be 2500 billion cents/210 billion kWh = 12 cents.
The corresponding capital cost for power generated by coal-fired power stations would be about 0.7 cents. Assumptions; Capital cost of a 1 GW plant, $2 billion, so if it lasts 40 years capital cost p.a. is $50 m. At the typical capacity factor of 0.8 output would be 0.8 x 1 million kW x 8,760 h = 7,008 million kWh/y = 7 billion kWh/y. So cost per kWh would be (5 billion cents)/(7 billion kWh) = 0.7 cents.
Compare the costs of the renewable system and the coal system when fuel is included. For the former the fuel, sunlight and wind, costs $0. For coal fired generators ½ kg of coal produces 1 kwh. At $100/t the cost of 1 kg of coal is about 10 cents, so the amount to generate 1 kWh would cost 5 cents.
So the cost of capital + fuel for the renewable system, 12 cents/kWh, is around twice the cost of capital + fuel for the coal powered system.
However the capital cost of producing electricity is a small fraction of the total production cost. Unfortunately the other factors are not easy to quantify. For renewables we would have to include the cost of land, access roads, wages, insurance, losses in transmission, maintenance and interest on capital borrowed, and company. profits. The energy needed to produce the generating capacity and all these other inputs would have to be subtracted from output; they might add to 5% of gross output. The cost of transmission lines would be much higher than for coal-fired electricity because they have to connect many widely spread generating sources, which is not the case with a few big coal-fired power stations. The greater amount of lines would mean greater transmission losses. Transmission costs presently makes up 40% of the retail bill for electricity.
And there will be times when the average capacity of the renewables is well under 6%; what would we do then? To keep the lights on we’d need batteries or a fossil fuel supply system capable of meeting much of the 600 GWh/d demand, but sitting idle most of the time, while adding to system cost per kWh.
So at present the capital cost of renewable electricity seems to be about twice al-fired that of coal-fired electricity. From above the total production cost would be much higher (assuming system capacity factor goes no lower than 6%.). The retail price is adding 19 cents to the present coal-production cost. If the energy needed to do the many additional things feeding into production cost is higher, then the cost of doing them would be significantly greater than 19 cents. Let’s assume it adds 50% to the production-retail gap making it 28 cents. The retail rice would then be 42 cents. Add the greater cost of the more elaborate trans mission lines than those adding 40% of 28 cents = 11 cents to the bill now, and the total retail price looks like it would be over 45 cents.
So the cost per kWh for a renewable system is likely to be much higher than the 28 cent average now, even without the cost of dealing with times when system capacity factor is under 6%.
Reduce generating capacity by using batteries for low input days?
This does not reduce capital costs. It increases them significantly. It is always cheaper to minimise use of batteries by maximising use of generating capacity.
Lets take the above arithmetic. If we halve the amount of renewable plant to generate 208 GW costing $312 billion, and provide half the power needed on that day from batteries, we’d need 208 million kW x 24 h = 4,992 million kWh. The present price of big batteries is $440/kWh, so the cost of drawing the necessary amount of power from them would be $2,196,480. We would have saved $325 billion but paid nearly 7 times as much to provide the capacity to supply the other half from storage. It’s much worse than this because the batteries will last only about 10 years compared to 25 for the wind and solar plant.
And this arithmetic greatly underestimates the cost of battery supply; see below on storage options.
To summarise.
These numbers indicate that an Australian electricity supply system dependent entirely on renewable sources is conceivable, but the retail price of the electricity will at least be very high, possibly around twice the present cost.
The outlook is much less optimistic for Europe and North America where winters set much more difficult conditions.
The European situation?
Australia probably has the world’s most favourable conditions for dependence on renewable energy. Europe however can experience continent-wide periods of cold, calm and cloudy weather lasting weeks. Ohlendorf and Shill (2020) report that an independent expert commission of the Bundestag generally assumed no-wind-no-solar periods of two weeks. They refer to these “dark doldrum” periods as ”Dunkelflaute”. “Every year, a period of around five consecutive days with an average wind capacity factor below 10% occurs, and every ten years a respective period of nearly eight days. … The longest event in the data lasts nearly ten days.”
They note that low wind events are more common in summer, but that is when the contribution from northern European hydro sources is at its lowest. In winter solar input is very low or more or less non-existent for weeks.
Ruhnau and Qvist (2021) examine 35 years of German data and find that Germany would need 36 TWh of storage to deal with the longer term tasks. That’s 56 times Australia’s daily consumption, for a population 3.5 times as big.
Smil (2021) points out that between 2000 and 2019 Germany increased renewable capacity by 89%, in a period when electricity consumption hardly increased. Fossil fuelled generation remained about the same, meaning that renewables didn’t replace fossil fuel use, they were added to the existing fossil-fuelled generating system, which had to be largely retained to deal with the intermittency of the renewables. One consequence is that two parallel systems have been built each capable of meeting demand, another is that the price of electricity is almost three times the US price.
It is difficult to imagine how an affordable high renewables penetration energy production system could be established for Europe.
100% renewable supply of total energy demand?
The foregoing discussion has only been about meeting the demand for electricity. Total energy demand in Australia is around 5 times as large as the electricity demand.
There are two difficult issues. The first is how much demand can be transferred from fossil fuels to electricity. Enthusiastic claims are made but the answer is not at all clear. It is highly unlikely that this can be done economically, or at all, for shipping, long-haul air transport, heavy farm equipment, plastics, and big trucks.
The second issue is the amount of energy lost when electricity is converted into forms that are needed, especially hydrogen. If it is assumed that 75% of the present non-electrical demand can be converted to electrical drives, and the remaining 25% of present demand is to be met via hydrogen with an overall 30% efficiency when used (after losses due to electrolysis, compression or liquefaction, storage, and fuel cell use or combustion in gas turbines), then the total amount of electricity that would have to be generated would be over 7 times the amount of energy now consumed as electricity. If half of the present energy non-electrical use can be via electricity, then the multiple is about 9. This does not take into account energy embodied in equipment.
The above derivation indicated that the retail price of a kWh of electricity would be at least in the region of 45 cents per kWh. If we needed 7 times the present 219,000 GWh each year the retail cost would be around 219,000 million kWh x 7 x 45 cents = $689,500 million, which is about equal to 41% the GDP. At present we are paying about $60 billion for electricity, $30 billion for petrol and diesel, and $20 billion for gas. But almost half of what we pay for petrol is tax, meaning that our total non-electrical energy expenditure is around $80 billion p.a. and our total energy expenditure is around $140 billion.
So if these numbers are approximately correct it looks like national expenditure on energy would be over 5 times the present figure, and an impossible proportion of GDP.
These have not been highly confident conclusions, some of the numbers are not certain and/or approximate, and the exercise is intended as indicative not precise, but they indicate the difficulties and costs for high penetration renewable energy supply will be very problematic.
Keep in mind that this has been about Australia which has probably the best renewable energy sources of any rich country.
What then should we do?
The answer is, emphatically, move to 100% renewable energy supply as fast as possible! But we cannot function well on renewables unless we undertake an astronomically big De-growth transition to The Simpler Way. That means scrapping capitalism, and the obsession with wealth and affluence, individualism and competition, and instead being content to live frugally and cooperatively mostly in highly self-sufficient local communities. The many people living in Eco-villages today know that these ways can provide us all with a high quality of life on extremely low per capita resource use rates. (See Lockyer’s evidence, 2017.)
No other way can defuse the big global problems, because they are basically due to the quest for affluence and growth and the resulting unsustainable over-consumption. At present effort is going into trying to increase supply, but the problem can only be solved by action on the demand side, that is, by shifting to far simpler lifestyles and systems.
We could make this transition easily and quickly…if enough of us wanted to do it. (For the detailed case see thesimplerway.info/) I think the chances of this path being taken are very poor, but it is the only sensible option, and lots of people are working for it, including the Simplicity Institute (2017.)
Appendix; Problems with storage options?
They are all problematic. Following are brief notes on them; for more detailed discussion see Trainer, 2023.)
Batteries.
There is almost never any reference to the fact that only about 80% of the energy going into charging a battery can be retrieved, or to the fact that 10-20% of that is then lost due to inverter inefficiency in converting 12 volt DC electricity into 240 volt AC power. These two factors mean that the battery performance numbers used above are much too optimistic. They should be reduced by up to 36%, and then the net energy they deliver would cost up to 50% more than it seemed above where they were not taken into account. At present batteries are only built to deliver for up to about 8 hours, maybe 12 hrs. Cost rises rapidly with length of storage time. (Kitchen, 2024, Colthorpe 2020.) It is not clear what this would mean if storage has to be for 72 hours.
At present only very small amounts of power are being supplied by big batteries so long delivery times are not required. In 2024 the 8 biggest batteries in Australia could only provide power at their MW rating for 1.6 hours. (Leitch, 2024.
In addition there is the issue of limits on the rate at which storage can be drawn from. Leitch’s figures show that the MWh value does not indicate the rate at which they can be drawn on to meet demand. Australia’s 8 biggest batteries can hold 825 MWh but can only deliver at a total rate of 530 MW. This means that the MWh storage figure can be a quite misleading indicator of ability to meet demand at a particular time. The Hornblend battery in South Australia for example can store 129 MWh, but could provide power only at the rate of 100 MW for an hour, not 129 MW.
A factor that is often overlooked in these kins of estimates is the energy cost of producing energy, especially that going into producing setting up the solar panels and wind turbines. This might be around 5% of its lifetime energy production. The energy cost of the batteries needs to be taken into account.
The above numbers are only for capital costs and do not include in the other production costs such as maintenance, wages, insurance, storm damage, interest to be paid on borrowed capital, land purchased, access roads. The total would be considerable. For coal-fired power they add to about 8 cents/kWh
Battery costs are very likely to fall ... in the near future, but high demand and scarcity of materials such as Lithium, Nickel and Copper will probably see them rise significantly. Mineral ore grades are falling and costs are rising significantly. NREL expects cost to fall to 40% of present cost by 2050. (Cole and Karmakar, 2023.) Given the many uncertainties, especially the effect of scarcity on materials costs over the next 25 years, little confidence can be put in this expectation.
There is a strong case that there are nowhere enough scarce minerals to enable world wide reliance on batteries. New kinds of batteries are being developed so it is not clear how significant materials scarcities will be.
The above figures indicate that for a world of 8 billion to have a 100% renewable energy supply system using lithium batteries for storage, the amount of lithium metal needed would be over 206 million tonnes. Current world reserves at present are estimated at around 14 million tonnes.
Hydro.
Ordinary “once-through” hydro-electric dams are not likely to take on much more of the storage task as there is limited scope for increasing these, especially in view of environmental considerations. The promising option is Pumped Hydro, pumping water up to high dams and using it to generate later. (Blakers, Lu and Stocks, 2017.) The potential of this approach is far from clear due to uncertainty about the number and cost of small sites. The ROAM (2012) and ENTURA (2018) studies provide similar plots showing possible sites in ascending cost per kWh, indicating that when the total stored reaches c. 320 GWh cost reaches c. $1000/kwh and is accelerating steeply. It is not clear whether this includes the cost of access roads and transmission lines. Parkinson (2020) reports that recently there have been sudden large increases in estimated pumped hydro system costs. He says the technology is regarded as mature and thus no reductions in costs are expected by 2050. The recent elaborate AEMO Integrated System Plan (2022) does not point to the PHS option as a major solution.
Vehicle-to-grid storage.
It is often assumed that the batteries in electric vehicles can make a considerable contribution to the grid storage task. Trainer (2022) points to the commonly overlooked timing problem with this proposal. After morning travel to work places, time would be needed to recharge batteries. Stored energy would then be available to the grid for a few sun-lit hours, but that is when solar energy is maximally available to meet grid demand without drawing on storage. Drawing on batteries would have to cease in time to recharge for the journey home. When they arrived batteries would need to have retained sufficient storage for the journey out next day. It would seem best if the vehicle had just enough storage to enable these steps. Why equip it with additional capacity that it didn’t need, to be used as a store for the grid? Such capacity would be best located in a stationary place, not carried around adding to the weight of vehicles.
Hydrogen.
The viability of this option for large scale storage is problematic, due primarily to the need for energy-intensive compression, cooling or liquefaction, and infrastructure embodied energy costs for instance for pumps, pipes and strong and heavy tanks. Bossel (2006) sets out commonly known figures; electric energy from a wind turbine might be converted to hydrogen at 70% efficiency, compression might take 7% of the energy of the gas compressed, delivery by tanker would involve a 7% loss, and conversion back to electricity via a gas turbine might involve a 40% loss, or a 50% loss for fuel cells. Thus to deliver 1kWh of energy from storage might require 2.5 kWh to be produced, not including any of the significant embodied energy costs of the plant and equipment required for all steps. Several additional difficulties, including leaks, embrittlement of pipes, weight of transport containers, and energy costs of energy embodied in tanks and compressors etc. are considered in Trainer 2022.
Concentrating Solar Thermal generation (CSP).
CSP systems can store energy in the form of heated oil in tanks. These are designed to sustain electricity generation for up to 15 hours after sunset but are not envisaged on a scale that would last for several days. More importantly, recent evidence indicates that CSP has poor performance in periods of low solar energy. (Trainer, 2014.)
De Castro and Capellan-Perez (2018) undertook a detailed study of actual production data for CSP units, including all 50 operating in Spain, and concluded that performance is markedly lower than that claimed by the many theoretical studies and industry pronouncements, indeed often under half those expectations. They are not alone in this finding. The EIA’s review (2017) found that “… the actual production of electricity of ISEGS, SEGS, Solana and Crescent Dunes has been much less than the planned values, … Additionally, the costs have been much larger than what was planned. The real-world experience thus casts considerable doubts on the number being proposed for the CSP ST technology by expert panels and the literature.”
De Castro and Capellan-Perez point out that CSP summer/winter output ratios are much higher than for PV or wind sources, meaning that its capacity to deliver in winter is relatively handicapped. Several studies show that average performance in mid-winter months is quite low, often around 5% of capacity. (Danelski 2015, Danko 2015, Dietrich 2016, Weißbach et al. 2013, IEA 2010.) The capacity factor during the worst few days in the year is likely to be well below these monthly averages.
Biomass.
The relatively few estimates of potential Australian biomass energy production differ considerably and viability is controversial. Electricity production from biomass via direct combustion or by conversion to gas for turbine use has low reported efficiencies. Rana et al. (2020) report EROI (Energy Return On Investment) ratios as low as between 6:1 for wood inputs and 2.12 and 3:1 for other types. In their CSIRO study for AEMO Hayward (2013) says the energy efficiency of gas production from biomass is under 50%. Thodey (2013) puts it in the range of 13-20%. The generators must be large in scale to maximise efficiency, which means longer trucking distances, adding embodied transport energy costs to those of plant.
These difficulties have led various reviews to see the technology as not being viable. Weisbach et al. (2013) say, biogas-fired plant are “…clearly below the economic limit with no potential of improvements in reach.” Syngas Technology (2012) says “...no viable technology has been available to produce refinery grade syngas from biomass.” (See also Wood et al. 2012.)
In addition to these production issues there are uncertainties to do with future biomass availability for power generation given the competing demand for land for food and materials and the difficulties likely to arise to do with climate and water in an era when accelerating dryness is predicted. These are serious concerns for Australia. There is also the argument that biomass should be reserved for transport fuel especially aviation, not allocated to electricity generation.
This brief review indicates that all the major options for the storage of very large quantities of electrical energy for periods of several days involve major difficulties.
----------------
‘
Aneroid, (2022), Wind energy. https://anero.id/energy/wind-energy/2022/january (Other dates and solar data are accessible at this site.)
Aneroid, (2022), Solar energy. https://anero.id/energy/solar-energy
Capellan-Perez, I., C Castro and L. Gonzlez, (2019), “Dynamic Energy Return on Energy Investment (EROI) and material requirements in scenarios of global transition to renewable energies’ Energy Strategy Reviews, Volume 26, November. 100399 https://doi.org/10.1016/j.esr.2019.100399
Clack, C. et al., (2017), “Evaluation of a proposal for reliable low cost grid power with 100% wind, water and solar”, PNAS, June 27, 114, 26, 6722 – 6727.
Cole, W., and A. Karmakar, (2023), Cost Projections for Utility-Scale Battery Storage: 2023 Update, National Renewable Energy Laboratory
Crawford, D., T. Janovic, M. O”Connor, A. Herr, J. Raison and T. Baynes, (2012), Potential for electricity generation in Australia from Biomass in 2010, 2030, and 2050, AEMO 100% Renewable Energy Study, 4 Sept. CSIRO Report EP – 126969.
Elliston B., Diesendorf, M.,and I. MacGill, (2012). “Simulations of scenarios with 100% renewable electricity in the Australian National Electricity Market”, Energy Policy, 45, 606 – 613.
Friedmann, A., (2016), When Trucks Stop Running, Dordrecht, Springer.
Hall, C. A. S and K. A. Klitgaard, (2014), Energy and the Wealth of Nations, Dordrecht, Springer.
Hart, E. K., and M. Z. Jacobson, 2011. A Monte Carlo approach to generator portfolio planning and carbon emissions assessments of systems with large penetrations of variable renewable, Renewable Energy, 36, 2278 – 2286.
Heard, B., et al, (2017), “Burdon of proof: A comprehensive review of the feasibility of 100% renewable-electricity systems”, Renewable and Sustainable Energy Reviews, 76, 1122-1133.
Heinberg, R., and D. Fridley, (2016) Our Renewable Future, Santa Rosa, California, Post Carbon Institute.
Kitchen, C., (2021) Big Battery Bonanza Australian Energy Council. https://www.energycouncil.com.au/analysis/big-battery-bonanza/
Lenzen, M., B. McBain, T. Trainer, S. Jutte, O. Rey-Lescure, and J. Huang, (2016), Simulating low-carbon electricity supply for Australia, Applied Energy, 179, Oct., 553 – 564.
Lockyer, J., (2017), “Community, commons, and De-growth at Dancing Rabbit Eco-village”, Political Ecology, 24, 519-542.
Moriarty, P., and D. Honnery, (2010), The Rise and Fall of Carbon Civilization, Springer, Dordrecht.
The Simplicity Institute, 2017. simplicityinstitute.org/
Trainer, T., (2013), "Can Europe run on renewable energy? A negative case", Energy Policy, (63) 845 – 850.
Trainer, T., (2017a), “Some problems in storing renewable energy”. Energy Policy, (in press.) thesimplerway.info/REstorage.htm
Trainer, T., (2017b), “Can renewables meet total Australian energy demand: A “disaggregated” approach.” Energy Policy, 1`09, 539 - 544. thesimplerway.info/CanRE.htm
Trainer, T., (2017c), “Estimating the retail price for 100% Australian renewable electricity.” (In press, Energy Policy.) thesimplerway.info/REretailprice.htm
Trainer,T., (2017d), “The overlooked significance of the EROI for renewable energy supply systems.” (In press, Energy Policy.) thesimplerway.info/REsystemEROI.htm
Trainer, T., (2023). "Storage Implications of Australian Wind Data," Biophysical Economics and Resource Quality, Springer, vol. 8(4), pages 1-8, December.