I’ve joked to my friends that if there is anything that proves how important I consider the clean energy topic to be, it’s me digging into electricity pricing. I have a natural aversion to economics – I’ve demoted that aversion somewhat from the position of idealist elitism it carried back when I was a teenager (anything to do with money was about greed and not worth considering). Now I acknowledge that my prejudice toward economics is a flaw in my character which means I’m probably missing a whole lot about a fascinating and complex aspect of societal dynamics. I’ve battled that weakness a couple of times to catch a glimpse of that complexity.
So, when I used to hear people complain about nuclear energy being expensive and slow to build, I would thoughtfully nod my head, thinking: “Well, they probably have a point, it’s expensive, and it’s quite a project to build a plant. Still, it’s important because it can provide astounding amounts of reliable carbon-free energy, so we just have to stomach that slow and costly process.”
You’d think I would have learned by now about the risks of making assumptions based on hearsay?
When someone pointed me to a graph comparing the best build-rates we’ve ever had on carbon-free energy over the last half a century (first the excellent one presented by Climate Gamble, then another from Cao J et al, Science, which you see below), I had to stare at it for a while to process how wrong I had been about that “slow to build” part.
Considering one huge plant takes years (and with delays, sometimes more than a decade) to build, it seems slow. But that’s mostly because we don’t really have a useful everyday intuition about how frigging massive amounts of energy nuclear plants produce once they are online, and then keep on producing, for over half a century.
The only thing that can surpass a big nuclear plant in energy output is a hydropower dam of epic proportions – like the record-breaking Itaipu dam on the border of Brazil and Paraguay, which produces almost 20% worth of Brazil’s entire electricity needs. The biggest dams in the world (Itaipu and Three Gorges) produce only twice as much electricity each as the existing largest nuclear power plant in Canada generates alone.
I hope these visuals (above and below) help drive home the point that nuclear power plants provide a massive constant flow of energy.
So… if I was wrong about that one, what about the economics arguments?
I have said many times that an evidence-based, actual commitment to climate solutions, using all IPCC recommended technologies – including nuclear power – is the top priority issue in the world today. I’ve already dug into questions about nuclear waste, safety, and the expert views on climate mitigation power. I can’t afford not to at least try to understand the economics better, too.
What do the actual numbers tell us?
Like so many times before, when I’ve actually started looking at a topic, the dishonesty of many of the very loud and self-convinced arguments is astounding. I began to delve into this in my recent piece The Right Price For Saving the Planet Depends on the Energy Form. I highlighted some of the inconsistencies in the economic arguments against nuclear, but I didn’t actually wade into the numbers (there’s that natural aversion again). Serendipitously, however, I was pointed to an excellent blog post by Jani-Petri Martikainen, a physicist and a fellow Finnish Ecomodernist. He had done just that: perfectly illustrated the points I had made in my first piece, but with actual numbers of capital per energy produced. I was thrilled, and asked him if I could translate his table into English. He made the devilishly reasonable suggestion that I should update the 5-year-old table with newer data.
Figuring out strange things like what was the meaning of energy per time divided by time divided by time yet again, or what were the differences between four differently named energy capacities for power plants, it took me a few days and many deep breaths to do that. Finally, here is an updated version of that original table, looking at a common point of complaint in many economic arguments against nuclear power: that it ‘costs so much to build’.
That argument holds water only if we are merely interested in results in the short term. But what, exactly, is short-term about mitigating climate change and having clean air to breathe? We need to clean up fast, and we need to do it for the long-term.
Despite lack of subsidies, and excessive regulatory hurdles, or even purely miserable planning leading to unnecessary delays and growing costs in the construction phase, even when nuclear power plant becomes as ‘scandalously expensive’ as the reactor currently in construction in Finland – even then, the cost divided on the total energy output remains low in comparison to other clean energy forms. As soon as Olkiluoto 3 (which the ‘Nuclear Finland’ column in the table refers to) will come online – in about a year and a half – it will provide 10% of Finland’s electricity, and is expected to decrease the price of electricity in all the Nordic countries, as long as other nuclear plants are not shut down.
South Korea on the other hand is an example of a country where the construction costs of nuclear have stayed consistently low. They produce 30% of their electricity with nuclear power, and are in the process of building two more plants.
Even more complicated analyses of economics
There are many other things to take into consideration in the question of energy economics, of course. There are very long and in-depth discussions on that for nuclear power for those who would like to dive in to that. I would like to give an idea about some of the major factors at play, and once more lift up some of Jani-Petri’s excellent points.
Jani-Petri spends quite some time delving into the use of what is called a discount rate: a tool for appraising construction project investments vs their returns in the future. He notes some strange tendencies in the way the authors of chapter seven of the Intergovernmental Panel on Climate Change (IPCC) Working Group III -report try to hide proper comparison of the costs of energy forms in their supplementary materials, perceivably for no other reason than because of the discomfort of letting their readers know that nuclear power is almost always the lowest cost energy solution, and always the lowest cost zero carbon energy solution.
For more detail, you can find an in-depth account of the numbers from the annex III of the WGIII report in his article on Discounting and costs (Part 2) IPCC WGIII report on mitigation. Note, these are comparisons of Levelised Costs Of Energy (LCOE), not simply construction costs – more on LCOE’s below. Shortly:
nuclear power is the lowest cost zero carbon source of electricity no matter what discount rate was used.
This is his table based on IPCC’s numbers. Nuclear is set as the baseline, the figures in green=more expensive than nuclear, the red=cheaper.
But there are other kind of aspects to the question of economics of electricity apart from investment costs. What if we turn the question around, not how much it costs to produce x amount of energy, but what value does the electricity have for the consumer, whether it be a family or a hospital.
Availability of energy has value
How much would you be prepared to pay for electricity that was always available in the morning, when you were fixing breakfast to your family, then again in the evening, when it was time for dinner, and stayed on until you went to bed? What about the same amount of electricity, but available only during the daylight ours? Or perhaps only in the evening, and not in the morning, or vice versa – and you could only have an approximate idea about which way it would be, a day or a two in advance?
The ability to put electricity generation potential to use is called capacity utilization. Variability in an energy source inherently lowers the degree of that utilization. In another blog post, Jani-Petri draws interesting parallels from a point of history when coal power won out against hydro, not because of price, but particularly because of its high availability. On the prospects of moving from that constancy to a more variable power scenario, he asks:
Who bears the cost of lower utilization? Labour? Lower salaries and/or more irregular working hours anyone? Vacations in the winter since solar power produces mainly in the summer?
The value of electricity to private as well as corporate consumers of electricity has a lot to do with the reliability of it being there when they need it. Considering that the capacity of storing electricity is low (in the manner of minutes rather than hours on a national scale), there are natural complications to electricity production that will vary with weather and time of year. Where it can flexibly be applied, it is very welcome, but that type of energy reaching a majority share of the market is challenging, to say the least. All this also plays a role to the actual value of electricity to the consumer. Some services, say things like hospitals, schools, trains, and cold storage facilities, really need a reliable, constant flow of power, a ‘base load’ kind of electricity. The cost of outages tends to be very greatly larger than the price of the electricity itself that was lost.
It is clear that being constantly available increases the value of electricity. Constant power enables things we simply could not achieve with intermittent supply. It is not a simple task to assign numbers to this, but it is also not something we can ignore.
What about all the auxiliary costs?
Of course there are also a number of different specific costs we could discuss: operational and maintenance costs (for nuclear, often comparable or smaller than coal), costs for expanding the electrical grid (considerable cost for renewables, not included in their price), and costs for energy storage (another big cost and largest limiting factor for renewables).
Take the fresh news about the world’s largest battery, installed in Australia to help buffer wind farms: the quick back of the envelope calculations by Robert Hargraves (Ass. Prof. in Math and founder of a thorium molten salt-nuclear energy company) puts the battery’s estimated added costs for electricity at almost 40 cents/kWh. I am happy to update the piece with a better estimate, if I find one – but for context, even a tenth of that price would still double the price for energy per capital investment (which, as seen in the table before, for European wind is around 4 cents/kWh, vs 1 cent/kWh or less for nuclear).
Even this largest battery in the world only adds about 20 seconds worth of buffering capacity for the total electricity consumption of Australia (calculated from the 2014 total) – or about a few hours worth for 30,000 households. It enables a supply during the short transition time it takes to power up the back-up fossil fuel sources – it’s not a solution for reliable long term electricity supply.
There are also costs for handling waste. Renewable companies are largely not obliged to take care of their hazardous waste, but costs of waste collection and handling are included in the responsibilities of the nuclear industry. Even so, looking at the total cost from mining, processing, and handling of waste, the Word Nuclear Association (WNA) estimates nuclear power’s total costs from fuel to be considerably smaller than that of coal and gas’ plants, thanks to the vastly concentrated energy contained in uranium (easier transport and storage with such small volumes):
the total fuel costs of a nuclear power plant in the OECD are typically about one-third to one-half of those for a coal-fired plant and between one-quarter and one-fifth of those for a gas combined-cycle plant.
The total costs for decommissioning a nuclear plant are also only a fraction of its price per energy produced – estimated at about 0.1-0.2 cents/kWh. No doubling or ten-times factors to the prices there.
The WNA present OECD estimates on the levelised costs of energy (LCOE) as well as their system costs for four different countries (see below). Nuclear is the cheapest option in all but one: the US – where its ‘only’ the third cheapest, and still cheaper than offshore wind and Solar PV. This trend isn’t surprising, considering that nuclear costs in the US have risen to a class of their own through a few decades of increasing regulatory burdens. Despite all that, it’s still almost as cheap as the cheapest low-carbon energy form (onshore wind). You can more about that in my earlier piece, or in the paper from 2016, analysed by its authors here.
The true costs of energy forms
The big complication through all the price comparisons, is that the myriad types of environmental and health costs of different energy forms are not consistently included in their overt price, but externalised to the society at large. Ultimately, projects like the EU ExternE are (while not above criticism, still) the ones we should look to in order to form a more comprehensive view on the best targets of investment for long-term energy solutions. As I wrote in my piece on Nuclear Waste: Ideas vs Reality,
If we are serious about protecting human health as well as that of the environment, we need to step above our gut reactions, our ‘we just know’ estimates on the problems of different energy forms, and we need start comparing all the external costs (all risks and health impacts) from each type of energy – estimates that take into account the impacts from mining, operating, and waste.
Hydro power has its own drawbacks, like a few spectacular accidents (largest being the 1975 Banqiao dam failure that cost 170,000 lives), and flooding large areas, nevertheless, it is a great clean source of energy – but we don’t have enough waterways left to harness in order to replace fossil fuels. Wind is intermittent, but it can be a great help up to a point – IPCC does not see a realistic scenario for more than 30% of the world’s energy share for all renewables taken together, by 2050.
Nuclear power has the advantage of being a long-term solution that can be expanded fast. It provides us with massive amounts of clean energy for a low cost to the consumer, the environment, and the society at large.
For more of my articles on climate and energy, look here. Even better idea, however, is to read the short, evidence-dense book Climate Gamble or browse the graphs in their blog. If you would like to have a discussion in the comments below, please take note of my Commenting policy. In a nutshell:
- Be respectful.
- Back up your claims with evidence.
Appendix: the sources and calculations of construction costs
I have used newer numbers on the costs for all but Finnish nuclear (data unchanged), and updated the capacity factor for offshore wind to 40% from the 35% Jani-Petri had used. There were some scientific publications I found about lower life-time of wind power plants as well as lower capacity factors, but after some consideration I kept the more charitable values – rather be more optimistic than pessimistic to make the fairest possible estimates.
My way of calculating the cents/kWh was far from the most efficient (my mathematician spouse simply used the numbers in the table in one equation), but I am reasonably confident in them after confirmation by said mathematician and since I reached very similar results as Jani-Petri’s.
“In recent years the capacity factor for South Korean power reactors has averaged up to 96.5% – some of the highest figures in the world.” http://www.world-nuclear.org/information-library/country-profiles/countries-o-s/south-korea.aspx
Nuclear costs, Korea (and others): http://www.world-nuclear.org/information-library/economic-aspects/economics-of-nuclear-power.aspx
Korean reactor type: https://en.wikipedia.org/wiki/APR-1400
“The APR-1400 is an evolutionary Advanced Light Water Reactor which is based on the previous OPR-1000 design. Under Korean conditions, the reactor produced 1455 MW gross electrical power with a thermal power capacity of 3983 MW (4000 MW nominal)”
Commercial operation 2015: http://www.world-nuclear-news.org/NN-First-Korean-APR-1400-enters-commercial-operation-2012164.html
“overnight costs … $2021/kWe in South Korea” = 1723 euro/kWe (1.7 mrd/GWe)
1723 euro * (1400 MWe = 1400000 kWe) = 2.4 mrd euro
1400 MWe, 90 % cap. factor, 60 years. 8760 hours per year.
(1400 MW * 0.9) * (60 a *8760 h) = 662256000 MWh
2.4 mrd euro / 662256000 MWh = 3.6 euro/MWh = 0.0036 euro/kWh, 0.4 cents/kWh
*UPDATE: Better estimate of the cost would probably be 2330 dollars/kWe, which lands us at 1.97mrd euro/GWe. That would result in 0.42 cents/kWh, so no practical change to the table value (0.4). I’ve updated the figure while also updating with official cost statement on Finnish nuclear.
From what I could find, the 8.5 mrd euro total cost of Olkiluoto3 was still as in Jani-Petri’s original sources, and no change in the other numbers, so this one I didn’t re-calculate.
UPDATE: 2018 TVO came out with an official calculation of price, down 3mrd euros from the earlier worst-case estimate by Areva. Final price is 5.5mrd euro. I have updated the table accordingly, price per GWe down to 3.4 from 5.3, per cent price to 0.7 cents from 1.1 euro.
“Finland’s four existing reactors (about 2700 MWe net total) are among the world’s most efficient, with an average lifetime capacity factor of over 85% and average capacity factor over the last ten years of 95%” http://www.world-nuclear.org/information-library/country-profiles/countries-a-f/finland.aspx
Lifetime of Finnish ERP expected 60 http://www.new.areva.com/EN/operations-1707/epr-reactor-economic-and-competitive.html
Many nuclear plants have outlived their expected long life-times: https://www.scientificamerican.com/article/nuclear-power-plant-aging-reactor-replacement-/
Capacity factors of nuclear: https://www.iaea.org/PRIS/WorldStatistics/ThreeYrsEnergyAvailabilityFactor.aspx
European on-land wind capacity factor less than 21%
“For two decades now, the capacity factor of wind power measuring the average energy delivered has been assumed in the 30–35% range of the name plate capacity. Yet, the mean realized value for Europe over the last five years is below 21%; accordingly private cost is two-third higher and the reduction of carbon emissions is 40% less than previously expected. We document this discrepancy and offer rationalizations that emphasize the long term variations of wind speeds, the behavior of the wind power industry, political interference and the mode of finance. We conclude with the consequences of the capacity factor miscalculation and some policy recommendations.” http://www.sciencedirect.com/science/article/pii/S030142150900144X
1900 usd/kW European average 2016 World Energy Council https://www.worldenergy.org/wp-content/uploads/2017/03/WEResources_Wind_2016.pdf
1900 USD/kW to Mrd euro/GWe
1600 euro/kW * 1000000 = 1600000000 euro/GWe = 1.6 mrd euro/GWe
One wind turbine about 4.2MW (onshore or off) https://www.vestas.com/en/products/turbines/v136%20_4_2_mw
1600 euro/kW * (4.2 MWe = 4200 kWe) = 6720000 euro
1.6 mrd/Gwe, 20 years, 20% cap fact
4.2 MWe, 20 % cap. factor, 15 years. 8760 hours per year.
(4.2 MW * 0.2) * (20 a *8760 h) = 147168 MWh
6720000 euro / 147168 MWh = 45.6 euro/MWh = 0.0456 euro/kWh, about 4.6 cents/kWh
If the age would be shorter as claimed by some:
1.6 mrd/Gwe, 15 years, 20% cap fact
4.2 MWe, 20 % cap. factor, 15 years. 8760 hours per year.
(4.2 MW * 0.2) * (15 a *8760 h) = 110376 MWh
6720000 euro / 110376 MWh = 60.9 euro/MWh = 0.0609 euro/kWh, about 6.1 cents/kWh
If the cap factor was 25%:
(4.2 MW * 0.25) * (20 a *8760 h) = 183960 MWh
6720000 euro / 183960 MWh = 36.5 euro/MWh = 0.0365 euro/kWh, ab out 3.7 cents/kWh
Off-shore Danish wind 41% http://energynumbers.info/capacity-factors-at-danish-offshore-wind-farms
Capacity factor US wind, 2016 32% for 2016, 34% for 2016-2017 august https://www.eia.gov/electricity/monthly/epm_table_grapher.php?t=epmt_6_07_b
UK and Danish wind farms shorter lived than expected. “The analysis of almost 3,000 onshore wind turbines — the biggest study of its kind —warns that they will continue to generate electricity effectively for just 12 to 15 years.” …
“The decline in the output of offshore wind farms, based on a study of Danish wind farms, appears even more dramatic. The load factor for turbines built on platforms in the sea is reduced from 39 per cent to 15 per cent after 10 years.” http://www.telegraph.co.uk/news/earth/energy/windpower/9770837/Wind-farm-turbines-wear-sooner-than-expected-says-study.html Study: http://www.ref.org.uk/attachments/article/280/ref.hughes.19.12.12.pdf
One wind turbine nowadays about 5 MW
“Currently, more than 92% (10,936 MW) of all offshore wind installations are in European waters“
“Onshore investment cost assumed at USD2005 1,750/kW, and offshore at USD2005 3,900/kW.” -> Offshore 3.3 mrd euor/GWe
3300 euro/kW * (5 MWe = 5000 kWe) = 16500000 euro
5 MWe, 20 % cap. factor, 20 years. 8760 hours per year.
(5 MW * 0.4) * (20 a *8760 h) = 350400 MWh
16500000 euro / 350400 MWh = 47.1 euro/MWh = 0.0471 euro/kWh, about 4.7 cents/kWh
If the age would be shorter as claimed by some:
5 MWe, 20 % cap. factor, 15 years. 8760 hours per year.
(5 MW * 0.4) * (15 a *8760 h) = 262800 MWh
16500000 euro / 262800 MWh = 62.8 euro/MWh = 0.0628 euro/kWh, about 6.3 cents/kWh
German solar investment, page 41, ca 1350 e per kWp -> 1.4 mrd/GWe https://www.ise.fraunhofer.de/content/dam/ise/de/documents/publications/studies/Photovoltaics-Report.pdf
German solar farms about 100 MW
Capacity factors apprx 10-11%
Cheaper than US solar http://www.motherjones.com/politics/2012/11/cost-going-solar-continues-drop/
Solar farm 100MW
1350 euro/kW * (100 MWe = 100 000 kWe) = 135 000 000 euro
100 MWe, 10 % cap. factor, 25 years. 8760 hours per year.
(100 MW * 0.1) * (25 a *8760 h) = 2 190 000 MWh
135000000 euro / 2 190 000 MWh = 61.6 euro/MWh = 0.0616 euro/kWh, about 6.2 cents/kWh
The Tesla battery calculation:
to get a sense of how much battery capacity this new record battery adds to the Australian grid. The capacity is given as 129 MWh. I take that to mean it can hold and return that much energy. Googling gives Australian energy usage at 10,077.84 kWh (2014) per capita, so about 10 MWh. Times 23.46 million people gives total consumption at about 235 million MWh.
So how much is 130MWh of Australia’s total consumption?
A day would be some 600 000 MWh. An hour is 27 000 MWh. A minute is still more than 400 MWh. I get 130MWh to be somewhere around 20 seconds of Australia’s total electricity consumption.