Hot News Nuclear

Fuel use for Gen III+ nuclear power

In one of the entries on my series of posts on the Integral Fast Reactor, I pointed out that a next-generation nuclear-power-plus-full-fuel-recycling plant would require only 1 tonne of natural uranium fuel (or thorium, or nuclear waste, or depleted uranium) per year, for a 1,000 MWe plant. However, I recently got asked this related question:

Do you know of any sources where I can find what the fuel requirements would be for a typical 1 GW Gen 3 plant running for a year?

This is an interesting question. Two obviously modern plants to consider are the Westinghouse AP1000 (four are currently under construction in China) and the AREVA EPR (two are being built in Europe).

The AP1000 uses 4.25 % enriched fuel and achieves a burnup of 60 GWd/t (details here). The EPR uses 5% enriched fuel to get 62 GWd/t (details here). The following Excel table illustrates my calculations (blue = inputs, green = calculations, bold = results) — click on the table to download the .xlsx file and play around with it yourself.

This estimates a natural uranium metal use of 108 to 117 tonnes U per GWe per year, using an enriched fuel loading of 21 to 25 t for the two designs (1,115 and 1,650 MWe respectively, running at about 92% capacity factor). The EPR appears to be slightly more efficient than the AP1000 when levelised on a 1 GWe basis.

Note: If 0.2% U-235 tails were left over after enrichment (rather than 0% assumed above), then the value in row 9 (% nat) would become 0.51, and the corresponding U/GWe/yr for the AP1000 would be 163 t, and for EPR it would be 150 t.

My calculations, based on the performance documentation, are similar to the generalised calculations provided by the WNA, as given below:

(There are small discrepancies — if anyone can work out the sources of these, please let me know in the comments)

These Gen III+ plants are quite efficient in their fuel use (see the briefing paper The Nuclear Fuel Cycle for a detailed description of the different inputs — from which the above table was extracted). Depending of the grade of ore, this metal fuel will typically require the processing of 20,000 to 400,000 tonnes of mined ore. As explained in this excellent comment by Luke Weston (reproduced at the foot of this post), in polymetallic mines, this ore may already be extracted for other purposes. For in situ leach mining, the extraction process is different.

The above are idealized calculations for the newest, fuel efficient designs. If you want a rough figure for older reactors (which actually aren’t that much worse), the estimate of fuel use is about 24 t of enriched U per year, or up to 200 tonnes of natural uranium per GWe: see

Finally, an interesting video to look out for. Now uploaded in full on YouTube, this excellent BBC documentary (57 min) was first broadcast on 14 Sept 2011. It is hosted by Professor Jim Al-Khalili, and entitled “Is Nuclear Power Safe“.

Here is the blurb:

Six months after the explosions at the Fukushima nuclear plant and the release of radiation there, Professor Jim Al-Khalili sets out to discover whether nuclear power is safe. 

He begins in Japan, where he meets some of the tens of thousands of people who have been evacuated from the exclusion zone. He travels to an abandoned village just outside the zone to witness a nuclear clean-up operation. 

Jim draws on the latest scientific findings from Japan and from the previous explosion at Chernobyl to understand how dangerous the release of radiation is likely to be and what that means for our trust in nuclear power.


Appendix: Luke Weston on Gavin Mudd

In response to Gavin Mudd’s ridiculous article about the Olympic Dam expansion over on The Conversation, I wrote a lengthy comment in response. It’s posted over there but I will also copy it here for interested readers.

Contrary to the usual tendentious nonsense from anti-nuclear activists, Olympic Dam is not really a uranium mine. Olympic Dam is a copper mine. Following the expansion, the total copper production at Olympic Dam will be 730,000 tonnes per year, up from about 220,000 tonnes per year at the present. (Copper smelting is, incidentally, what requires most of the energy input to the Olympic Dam site, nothing to do with uranium.) After the ore is mined and milled, the copper minerals are separated and processed and we’re left with powdered mineral waste – the so-called tailings which seem to be a cause for great concern amongst environmentalists.

At Olympic Dam, however, those tailings contain a small amount of gold and uranium, and further processing of the ore (which you’ve already mined and milled anyway) to separate the gold and uranium into saleable products is economically attractive. (If those same relatively low concentrations of U and Au were present in an orebody that was not already being mined for the copper anyway, mining such a deposit would not be economically attractive.)

The gold and uranium are essentially “free” byproducts recovered from what would otherwise be tailings from the copper mine, with no additional mining – no additional hole in the ground – required to extract those resources. In this sense, it would appear that polymetallic Cu/Au/U extraction operations at Olympic Dam are actually a very environmentally friendly way to mine those metals, as opposed to the alternative of having additional, separate mines at other sites mining gold and uranium deposits. Getting the most value that you can practically get out of one single hole in the ground is an environmentally efficient, conscious approach to mining.

As Mudd points out, the Olympic Dam orebody contains rare earths, which are essential for wind turbines, LEDs, electric vehicles, fuel cells and the like, and for which demand is growing rapidly. I’m sure BHP Billiton is well aware of the chemistry of the Olympic Dam orebody, and as the price of rare earths continues to rise over the coming years, I’m sure they will pursue rare-earth extraction at the point when it becomes economically viable. As with gold and uranium, extracting these different elements from the polymetallic orebody is an environmentally friendly alternative to having multiple additional mines. (Incidentally, all other rare-earth mining prospects in Australia also seem to attract criticism from the predictable band of anti-nuclear “environmentalists”, who complain that horrible, radioactive, scary uranium and thorium will also be extracted from these polymetallic ores where it is present along with the lanthanide metals.)

What exactly does Mudd propose we should do with the uranium at Olympic Dam, if the status quo is not the way to go? If we are to mine the copper and gold (and perhaps rare earths too), then of course the ore that is mined contains the uranium as well. Should we simply cease the extraction of uranium from that ore? But that would simply leave all that uranium in the tailings – significantly increasing the amount of radioactivity present in that tailings waste, whilst of course it would not at all decrease the volume of that tailings waste or change its characteristics in any other way. Given the concern expressed by Mudd and other anti-nuclear activists about those scary radioactive tailings, leaving the uranium in the tailings does not seem to make sense at all.

Mudd calls the mine tailings “billions of tonnes of radioactive waste”, calling to mind nonsense fictional mental images of billions of rusty 44-gallon drums full of luminous green goo, but in fact those tailings really are just natural rock that comes out of the ground. The tailings contain natural uranium and the natural daughter-product radionuclides in the uranium series, all naturally created and naturally present in the ground. Mining the ore and extracting the uranium does not create, or add, or change the radioactivity of this natural material in any way – except for removing the uranium from it.

Removing the uranium (well, essentially all of it, not 100% of it) from the tailings removes most of the uranium daughter-product radionuclides (Ra, Rn, Po etc.) that will form in the tailings over the long term, into the future, and will therefore remove most of the radiation dose that workers or the public may be exposed to from exposure to said tailings (a long time in the future). Leaving the uranium in the tailings, as well as the uranium daughter radionuclides that the uranium will become over long timescales, will substantially increase the radioactivity and potential radiation dose from those tailings.

If you live in a part of the world where uranium (and uranium daughters) are naturally geologically abundant, then you’re exposed to natural background ionising radiation dose from that natural geology – from uranium, radium and the other uranium daughters in the soil, in dust, from gamma radiation directly from the ground, from radon in the air, and from uranium and uranium daughters in water. All these background dose pathways are completely natural – they’re a fact of life. If you’re afraid of that, move to a location where the natural geology contains minimal uranium or thorium.

Does the mining by humans of these natural rocks that contain uranium and its daughter products actually cause any real change to the background ionising radiation dose rate that people receive from that natural radioactivity, compared to the radiation doses received anyway when that radioactivity just sits in the ground naturally (and is subject to natural erosion, natural geological and hydrogeological transport) and does not get mined? Good question… perhaps Dr. Mudd could point us to some research or evidence on this subject.

It is a well-worn and predictable rhetoric sound-bite from the likes of Mudd, Ludlam, Diesendorf and Lowe that Australia’s uranium exports, in terms of revenue dollars, are less than Australia’s exports of cheese or lamb. But these people should know better than to simply think about everything in terms of the economist’s bottom line when it comes to science-based ecology and environmental best practice.

In the 2010 calendar year, Australia exported 301 million tonnes of coal, which corresponds to about 7.2 * 10^18 J of thermal energy content. The 7555 tonnes of natural uranium oxide exported in 2009-2010 contains a thermal energy content (ignoring the thermodynamic losses in a heat-engine power station, and assuming inefficient, once-through use of low-enriched uranium in LWRs) of about 3.4 * 10^18 J.

Australia’s three modest uranium mines provide a total energy output which is about 50% of all of Australia’s coal exports (a bit more than 50% of Australia’s total coal output including domestically-consumed coal). And yet this clean energy resource is supplied from three mines which have a total environmental footprint on the landscape which is far, far smaller than 50% of the environmental footprint of Australia’s numerous coal-mining holes in the ground. 15,000 tonnes of uranium oxide would give you the same energy output (in LWRs) as all that coal, and 15,000 tonnes of mineral production is a hell of a lot less environmentally intensive than 301 million tonnes.

Clearly Australia’s uranium exports are not essential for Australia’s economy. But such an enormous resource of clean energy, with such a high energy density, which is abundant in the earth and is available at such a low cost is obviously incredibly valuable and important for the global environment.

Following the expansion of mining operations at Olympic Dam, the mine will produce about 19,000 tonnes of uranium oxide per year, which will generate (in relatively inefficient once-through use in LWRs) about 800 TWh of electricity. Over the coming years, Australia’s clean energy exports (in the form of uranium) are likely to provide enough clean coal-replacement capacity to catch up with, and offset, the greenhouse gas emissions from all of Australia’s coal exports.

When natural uranium is used very efficiently, in the Integral Fast Reactor for example, one tonne of natural uranium oxide (U3O8) will yield about 8 * 10^16 J of thermal energy in the reactor. Therefore, if the 19,000 tonnes of uranium oxide from the Olympic Dam expansion was to be used in this way, the amount of energy produced would be about 1.5 * 10^21 J (thermal) – an amount of clean energy nearly 10 times greater than the energy content (about 1.7 * 10^20 J) of all the coal production on Earth. That’s basically all the energy for all the world. And for all the people that don’t yet have access to electricity. From just one mine!

Given that Australian-Obligated Nuclear Material is carefully safeguarded and watched and is not allowed to be diverted into nuclear weapons (and in fact most of this material, such as depleted uranium, highly radioactive used fuel which contains some reactor-grade plutonium in it, and reprocessed reactor-grade plutonium, is not physically useful as the fuel for nuclear weapons), would Mudd care to explain to us exactly how Australia’s uranium exports are “potentially increasing nuclear weapons risks”?

Depleted uranium, used LWR fuel and recycled reactor-grade plutonium are not a “burden” – they are enormous resources of fuels for clean energy which can be used throughout the world to provide energy, replacing the need to mine a hell of a lot of coal – and indeed, replacing the need to mine lots of new uranium which might otherwise be used.

Mudd lists “massive energy consumption” amongst the mine’s other supposedly negative effects. In fact, Olympic Dam has an enormous “energy gain” – the mine’s clean energy production (in the form of uranium) is far, far in excess of the mine’s total energy inputs. The highly efficient use of uranium, for example in the Integral Fast Reactor, as opposed to enrichment and inefficient once-through use of low-enriched uranium in light-water reactors, will increase that “energy gain” enormously.

Australia is, and will certainly continue to be, a major exporter of uranium, for civilian nuclear energy use under strict safeguards, to the world. This uranium has a very valuable role in providing clean, safe, abundant energy to replace coal and fossil fuel use throughout the world – today and into the future.


By Barry Brook

Barry Brook is an ARC Laureate Fellow and Chair of Environmental Sustainability at the University of Tasmania. He researches global change, ecology and energy.

47 replies on “Fuel use for Gen III+ nuclear power”

One difference between Barry’s estimates and the WNA figures is that Barry is ignoring losses to the enrichment plant tails. As the WNA figures assume, depleted uranium has about 0.2% U-235 left in it, because it is cheaper to get more natural uranium than to continue trying to squeeze U-235 out of depleted material at 0.2%. When U prices go up, enrichment plant sometimes take another pass at old tailings, but the effort and cost required goes up very steeply at low concentrations, so there is always some loss. For 0.2% tails, the concentration of available U-235 is only 0.51%, so the U usage goes up to ~150Te/Gwe/yr for EPR and ~163Te/Gwe/yr for AP-1000


Luke Weston said: “If you’re afraid of that [natural radioactivity], move to a location where the natural geology contains minimal uranium or thorium.”

There are plenty of places where the rocks contain negligible uranium and thorium, limestone being a good example as it has been laid down by organisms who have no need to either element. However limestone is alkaline, so accumulates uranium from groundwater travelling horizontally through the area. Moreover, someone living on the limestone area will still be breathing radon in the air which has drifted across from ubiquitous uranium-containing rocks elsewhere.

Someone living on a coral atoll deep in the Pacific will be able to escape the presence of uranium and thorium in the soil. Sure, there is plenty of potassium in the sea, however there will always be protected from it by a shield of seawater. Further, he or she will have minimal exposure to radon, as the maritime air is likely to have travelled for several lifetimes of radon since last leaving uranium-containing land. Yet again, being at sea level, they are beneath the thickest atmosphere, giving maximum protection from cosmic radiation.

If there is a clear correlation between the incidence of hard cancers and the ambient level of natural radiation, epidemiologists would be able to point out a low incidence of these diseases among the people of Polynesia. Low, that is relative to populations that to live in granitic country — in the presence of uranium and thorium, or at high altitudes exposed to cosmic radiation. Or pretty well anywhere for that matter. As far as I know, the medical statistics of oceanic peoples are as sad as our own.


Nice one Luke. Should The Conversation take me up on my request to rebut that ridiculous article (I am just enough of an academic these days with my Adelaide Uni email address) I will draw heavily on that great breakdown of yours (with attribution of course). Well done, it really needed to be said.



Thanks for putting this together in one short article. It is usefult to have a short clear post like this available to provide as a link in comments elsewhjere.

Could I suggest you add one (or two) columns to show the equivalent figures for IFR (and for coal; i.e. tonnes of coal per GWe-y in comparison with this statement: “Depending of the grade of ore, this metal fuel will typically require the processing of 20,000 to 400,000 tonnes of mined ore.“). It would be helpful to have this information in the same place to make it easy for people who have not been following BNC but are referred to this article.

Luke Weston, excellent comment in the appendix. It is helpful to have your comment promoted to a lead article where it can be easily referenced.


Barry: Please do not forget the CANDU. It requires only 157 tons of U per gigawatt-year, as compared with AP-1000 (at 0.2% tails assay) which requires 181 tons per year.



150 tonnes U per GWe-year is about 17 tonnes per TWh.

Australia uses 266 TWh.

266*17= about 4500 tonnes uranium needed for all today’s electricity.

This is similar to the uranium output of the Olympic dam mine.

The fact that all of Australia could be powered by a byproduct of a single copper mine, is testiment to nuclear power’s great energy density. Even with inefficient light water reactors the power density is amazing.


The World Nuclear Association lists the burnup for a Korean APR1400 as 65 GWd/t. This burnup is a little better than the AP1000 and the EPR. I do not know how to find the other numbers in your chart. Is the burnup the essential factor in comparing uranium usage?

The APR1400 has the ability to turn a rod end-for-end during refueling so the end part of the fuel can be equally used on both ends. I’m not sure if this minor interesting fact is the essential burnup difference.

Do you consider the APR1400 a generation III+ design? Korea has two units under construction and several planned. The UAE has plans for four units.


Burnup is not the essential factor on its own. 100 GWd/t is a high burnup but if you need to use bomb grade uranium to get it, the reactor is a poor resource performer.

The other extreme is the natural uranium fuelled CANDU: this reactor uses fuel shuffeling and a lower neutron loss heavy water moderator to allow the use of non-enriched uranium. It’s burnup is poor, around 8 GWd/t, but due to the use of natural uranium the resource efficiency is improved over the light water reactor (LWR). Lower burnup also reduces the fuel fabrication costs because there is less fuel damage and corrosion.

Slight enrichment makes sense for a CANDU. With 0.1 percent tails (efficient centrifuge enrichment) and a 1.2 percent U235 enrichment level, it is possible to get down to about 100 tonnes per GWe-year.

For reference, this is excellent:

Click to access te_699_web.pdf

In particular check out the “extended burnup in CANDUs” part, and its graphs. Shows the same results as Barry’s, and adds CANDUs with the different effects of enrichment and tails assay.


To my knowlegde, the two GenIII designs that are pushing thermal efficiencies towards 40% are the EPR and Mitsubishi’s APWR while US-designs stay around the the typical 33%. The rationale is less about saving uranium at the front end -everyone pays more or less the same price after all- but about reducing the volume of spent fuel. At the same time that American utilities are willfully paying their famous tenth-of-a-cent-per-kWh and wait for the government to take the spent fuel, many European countries require NPP operators to organize disposal themselves.

E. g. Swiss NPPs share funding a cooperative dealing with disposal of any kind of radioactive waste according to their rated thermal power. Thus they are intrinsically motivated to get the most electricity possible from the heat produced. They care less about how much natural uranium is needed for fabricating their fuel pellets as long as it is cheap. So they would not have much interest in building a CANDU; they may in fact end up being punished for that if the disposal volume for SEU spent fuel turns out to be higher than for LEU spent fuel. Total disposal cost (incl. lower level waste and funds for decommissioning; are around 1 Rappen per kWh electric, which is within an order of magnitude of 1 US-cent.

Another interesting effect of a comparatively expensive back-end of the fuel cycle: Swiss NPP are among the world leaders in increasing burnup: At the front-end, any savings for manufacturing fewer fuel assemblies are consumed by increased enrichment needs, the benefits of such a policy are reaped when it comes to disposing of a smaller volume of fuel.


Great video:

The biggest health hazard from the Chernobyl and Fukushima nuclear plant accidents is to one’s mental health.


The fourfold increase in Olympic Dam production won’t happen without 700 Mwe additional power supply. So far there is no coherent explanation where that power will come from.


GeorgeS, on 27 October 2011 at 5:19 AM said:

Great video:

The biggest health hazard from the Chernobyl and Fukushima nuclear plant accidents is to one’s mental health.

I forgot to add that you can fix this problem with education.


The reasoning behind higher efficiency plants is not lower waste, that is just incidental (but good for public relations). Storing the spent fuel in concrete canisters costs almost nothing, roughly 1% of the total levelized cost is for the canister (dry cask) storage.

The real reasoning is simple economics. It costs very little more to have a better steam turbine and a more efficient steam generator, economizers etc. if you’re building a whole new plant. But you get a big improvement in the electricity output.

Compared to 33%, the EPR’s 37% gets 12% more electricity per unit reactor power. That is a big boost. You make 12% more money from the same reactor and it costs only a little more. The business model improves.


For example, suppose that using a more efficient turbine generator and axial economizer to improve steam quality in the steam generator, adds 12% to the total cost of the plant. With 12% more output every year you get the investment back in 1 year! And then you get 12% more electricity, very year, for 59 years.

We can see that improvements in efficiency can be very rewarding. The 12% reduction in waste is just gravy.


Well, I agree that dry storage for whatever you like how long you like is not an interesting cost driver. However, where the waste bill is proportional to thermal power and constitutes an important portion of the fuel cycle cost, its contribution to total economics will be a strong incentive to increase thermal efficiencies.

If on the other hand, you’re a private company with comparatively high lending cost but your fuel disposal is charged modestly and by electric output, you may prefer a lower thermal efficiency as long as the plant is cheaper and quicker to build, so there is less investment to finance.


Of course, a reactor core that is already as big and power-dense as physics and control aspects allow it to get can only be used to deliver more electrical power via increased efficiency in the turbine island. Thus higher per-plant-output is surely no lesser motivation for increasing thermal efficiencies.


Re. the tailings and enrichment: A lot of the 700,000+ tons of DU in the USA have U-235 fractions of 0.3% or even more. Given that we’ve got these “leftovers” from about three decades of fuel production, a laser isotope separation system that could economically get that down to 0.1% could plow through the tailings and produce 15 years worth of fuel for all the reactors in the States.

Cyril writes (apparently referring to once-through LWRs): The fact that all of Australia could be powered by a byproduct of a single copper mine, is testament to nuclear power’s great energy density.

As Barry points out, used in IFRs that byproduct can power the entire planet, not just Australia. I think this sort of fact has to be trumpeted from the rooftops. Once people grasp the energy density concept at these extremes, it gets far easier to make them see the difference between nuclear and wind/solar, and the inadequacy of the latter.


Today’s centrifuges can also economically use the enrichment tails of older diffusion enrichment operations. Laser enrichment is promising but it is only getting started in commercialization right now.

Let’s do a sanity check on what Tom Blees said If the Olympic Dam byproduct can get all of Australia’s 266 TWh of electricity consumption, then the IFR could produce about 150x that or roughly 40,000 TWh which is twice today’s global electricity consumption.

Imagine that, the byproduct uranium of a single copper mine can produce the world’s electric demand twice over!

How is this possible? Simple, splitting an atom of uranium gets us 50 million times more heat than burning an atom of carbon.

It is that factor of 50,000,000 that makes nuclear attractive.


I agree that it was a good video, but as you can imagine I thought it was quite an oversight to talk about transmutation with accelerators as a solution to the waste problem and completely omit even the mention of fast reactors or, for that matter, of the prospect of using LFTRs to eliminate waste. Still, as a response to FUD about radioactivity in general, it was a welcome contribution.


The various rates of conversion of uranium by this or that type of reactor was surely a little concern when you consider the tiny quantities involved. More or less as Barry says in the first sentence of the post, 1 GW of electricity is going to convert one tonne per annum of uranium. At 1 kW per person, that amounts to one gram per annum per person. Regardless of the reactor type.

What should matter to an environmentally concerned critic is how much waste our power generation is costing the environment. When anyone accuses, “what about the waste!” we should have a ready reply, and not about one gram of fission products. That is exactly where we should be loudly and proudly proclaiming how much CO2 is being released per person by our consumption of nuclear generated electricity. “It is small, and decreasing”

In recent threads the calculations seem to be based on the cost of the resources consumed. But that is an old paradigm, where it is burnt carbon that costs money and dumped CO2 is a right. Applying it to uranium misses the different evaluation of the impact on the environment.

Once we have scored the point about superior CO2 emissions, we should listen to our critics about the amount of waste (yes, waste) accumulated by slow-neutron reactors, Gen III or not. There is a lot more mass in once-used fuel than the fission products in it; the unfissioned is somewhere near twenty times the fissioned. What are we going to do with the stuff?

Hard-nosed operators may not want to recycle it, not even for fast-neutron reactors. Considering that uranium enrichment is getting cheaper and cheaper, the probability that once-used fuel gets recycled is diminishing. Heck, not even bomb makers are bothering to reprocess fuel any more!

It is the environment that needs Gen IV reactors, not the perception of vanishing resources.


Actually, Tom, I thought the same thing…no mention of Fast Reactors. Strange. And I’m saying this as part of the LFTR ‘crowd’.

But I think this reflects the very immature level we are at yet with breaking out of the very small Gen IV Reactor audience. Even with the recent endorsements of IFR and LFTR by well known groups or individuals, it’s still not yet talked about as part of the general energy solutions. It’s not page 1, above the fold, so to speak yet. In fact it’s still page 25 and often as a footnote. So we all have a long ways to go. Takes time. Once we get closer to both proof of concept and deployable demonstrations, as the shift from Gen III to Gen IV looks realistic “to most policy makers and game players” then you will see more about both.


There was also another episode of BBC’s Horizon series broadcast in 2006, entitled Nuclear Nightmares, and it’s well worth watching that one as well if you can find access to a copy of it.

With regards to the kinds of calculations Barry is discussing in his post, I think the WISE nuclear fuel cycle calculator looks kind of interesting and valuable, but you do have to treat it with great skepticism because it is an anti-nuclear group that has created it, after all.

I think all the fundamental arithmetic programmed into that webpage is sound, from playing with it a little, but I think that some of the input model parameters (which you can change and adjust) are quite pessimistic and biased a little, by default, to make nuclear energy look worse.

Still, if used appropriately and treated skeptically, it looks like a nice little software tool.

Thanks for re-posting my comment by the way, Barry, I’m glad you found it valuable :)


An excellent new article has appeared in The Independent:

New life for old idea that could dissolve our nuclear waste… A nuclear programme that was abandoned two decades ago has emerged as a possible 11th-hour solution to Britain’s plutonium-waste headache, which the Government has to decide on within weeks.

Read the full article here:


Luke Weston,

I think that some of the input model parameters (which you can change and adjust) are quite pessimistic and biased a little, by default, to make nuclear energy look worse.

Could you list the values you would suggest should be used in the calculator.


David Walters @
(and all the other IFR and LTFR proponents)

But I think this reflects the very immature level we are at yet with breaking out of the very small Gen IV Reactor audience.

I am one of them.

I have two main concerns:

1. cost

2. time until they are commercially viable – cheaper than Gen III+

My gut feel is that IFR and LFTR will not be cheaper than Gen III+ for many decades.

Is there anything persuasive – e.g. by competent, impartial estimators – as to what the LCOE of the proposed system is likely to be?


I would urge anyone assessing reactors on the basis of cost-per-energy, such as LCOE , might keep in mind that such orthodox engineering calculations fail to include considerations arising from climate change. You might just be following a red herring.

To be fair, any economic analysis can only include so many factors, and then must assume “everything else being equal”. However, at least on this website, any vision of the future should include estimates of the savageries that climate change will wreak on the world’s population and its infrastructure.

Contingencies in orthodox planning include predictions of the likelihood of a “once in 100 year flood” for example. However such planning implies a commitment to the belief that the next hundred years climate is going to be similar to the last hundred years climate. Please check if that really is your committed position, because you will be denying climate change.

If the future will really will be business-as-usual, of course gigantic, well proven, slow-neutron reactors would come up as the cheapest way to supply power across an ever expanding grid.

On the other hand, if we are selecting blueprints for an energy supply to be rolled out in an emergency response, then the considerations will be about whether it can be rolled out at all, how fast, and across which grids.

There should be other considerations too. Will the components be simple enough to be made in factories across the world with standard equipment and standard workers? Will components be small enough to travel on standard railways, standard roads and lifted into place with standard cranes? Can they be assembled by standard construction methods? Can the reactors be operated autonomously, remotely, or by occasional visits of trained staff? Will they need access to copious water supplies?

These are the sort of issues that a blog like this should explore when it comes to choosing power supplies for survival in a world of global climate catastrophes.


@ Peter Lang: Well, to just start by giving one example, the energy intensity for uranium enrichment, which is 2300 kWh/SWU by default, is a very pessimistic worst-case figure, where 100% of your enrichment is done by the worst possible, oldest, most inefficient gaseous-diffusion plant you could find, of which only one or two remain operational in the world today. Uranium enrichment by gas centrifuge has an energy intensity which is much lower, about 50 kWh/SWU.


Roger Clifton, good point. All metrics have their limits. The limits of the levelized cost are that they don’t include external costs, and also omit systemic costs. You’ve pointed out the external cost weakness already, but I’d like to point out the importance of the systemic cost. The systemic cost is particularly important for the direct financial analysis. Even if the levelized cost of solar gets down to a few cents per kWh we will still not power countries with it due to the intermittency cost – batteries cost more than today’s solar so it is quite mad.

Are we happy with 2 cents per kWh solar if it means a 20% solar 80% fossil grid? A glass that is 80% empty seems quite empty to me. But Karl Friedrich Lenz will now step in to tell us it is better than nothing.


Luke Weston, a good example of this is the Georges Besse project. It will save 98% on the Georges Besse industrial facility energy bill.

Pretty darned good Negawatts. And it’s a nuclear project – Georges Besse is an enrichment facility in France. It saves 3 GWe, four reactors worth of electricity freed up for the French grid.

For perspective: 3 GWe is enough to power all housholds in the Netherlands.

At 3 billion euros, these ‘new’ nuclear plants (new as in will be grid connected) cost only 1 euro per Watt.

Maybe this will even convince some die hard energy efficiency fanatics of the merits of modern nuclear power technology.


Luke Weston,

Thank you for the one example. You may have interpreted my question as rhetorical. It wasn’t intended to be. What I was seeking was a list of your suggest inputs for all the entries in the calculator so I (and perhaps others) could use it based on more realistic default values. I wouldn’t know where to start looking to get a consistent of input values to enter, nor which to use for the various technologies. If you know them, it would be useful to have them posted. Even better than putting them in a comment field would be tadd an addendum to the lead articel wiwith a link to the calculator, your suggested input values, references for them, and some explanation of how to use them.


Cyril R and Roger Clifton,

The limits of the levelized cost are that they don’t include external costs, and also omit systemic costs. [Roger Clifton has] pointed out the external cost weakness already, but I’d like to point out the importance of the systemic cost.

This is not a “limitation” in the LCOE methodology. The LCOE method can handle external costs and all the “systemic” costs you mention. In fact it is commonly used for analysis the effects of externalities and . Examples are:

• Treasury modelling of the effects of the CO2e Tax and ETS
• ExternE – effect of externalities on LCOE
• EPRI studies on the effect of CO2e price
• OECD/NEA/IEA studies of the effect of CO2e pricing
• Martin Nicholson’s recent paper:
• NREL modelling showing the average cost of transmission enhancements to support wind generation adds about $15/MWh, which can be added to the LCOE for wind
• Similarly with solar PV. The LCOE can be calculated based on what ever inputs you use, such as before after subsidies and with or without batteries.
• You can also use LCOE for a mix of technologies such as wind and gas, or a mix such as BZE did with their “Zero Carbon Emissions by 2020” fantasy.

The LCOE methodology is not the problem. It is how you use it.

I’d argue there is an important use for the LCOE method to determine the LCOE of technologies without carbon tax, renewable energy certificates, government subsidies and tax incentives, feed in tariffs, etc. But calculating LCOE without these distortions, we can then evaluate how much incentive or penalty is required to make one technology competitive with another. For example, it seems to me the capital cost of nuclear needs to be roughly halved to make it a viable proposition for Australia, or a morass of more government interventions can be imposed to achieve the policy objectives of this government at this point in time.


Peter Lang, I’m going to have to fully disagree with you on this issue.

First, regarding external costs. The levelized cost methodology relies on discounting. This is an important financial advantage because this is how real investors invest. Money now is more important than money in the future.

Unfortunately this makes the levelized cost method highly unsuitable to calculate long term external costs such as climate change. It discounts these to negligible levels. Climate change damages 100 years from now barely show up in the levelized cost method. But if people in 1911 made choices that damage us, we’d still be stuck with the full charge and consequences of the damages.

Second, also regarding external costs, the levelized cost method assumes the option of replacement and choices. Your power plant has a limited lifetime and you can replace it with something else after that. The worst thing that can happen to your powerplant is that it breaks down beyond repair. We know what it will cost to buy a new one.

This does not fly if enlarged to the level of our planet. Wreck your powerplant, buy a new one, with lots of options. Wreck our planet, where do we buy a new planet? What are our options?

Due to these limitations the levelized cost method is not suitable to calculate future damages of climate change into the cost per kWh.

Third, regarding systemic costs. It is certainly possible to consider the cost of solar or wind for 90% of our electric supply. This is among the things that are needed to solve climate change. Theoretically this can be done but rarely ever is done in practice. In practice people look at individual projects cost. If systemic costs are considered it is usually to a low level that is not relevant to solving climate change. The climate scientists tell us to reduce emissions 90% below today’s level. That is such a deep reduction, David Mackay has suggested the best way to think of it is “no more fossil fuels”. So we must consider the total cost of supplying our societies with energy without fossil fuels.

The studies that have been done suggest that it won’t be possible to power most of our society with nonfossil sources. In stead the studies all point out that we’ll just settle on fossil backup for 60-90% of our power. We’ve talked about the DeCarolis and Keith study that showed this, and they were being optimistic in not considering the systemic fossil backup burning inefficiency penalty arising from so much intermittent wind power.

This is where the external cost and systemic cost issues come together: the risk of fossil fuel lock in.


Cyril R, @ 31 October 2011 at 12:24 AM

Yes we fully disagree on this. I see your argument and Roger Clifton’s as in the same basket as “ignore the figures, just go with my beliefs and my gut feel”. That is the approach that got us 50 years of anti-nuclear activism and 30+ years of pro-renewable energy activism (and many other examples I could give). If you want to ignore discount rates, we can’t even begin a conversation. I have no interest in trying to discuss that. It is too basic. Furthermore, it would lead into subjects that are banned from discussion from BNC. So, we’ll have to leave this topic as “we agree to disagree”.


A policy to roll out gigantic Gen III reactors would assume that we can make century-long investment decisions with reasonable confidence that external events can be neglected. On the contrary, we have massive external events ahead — the message of climate change is of change itself. Much of the change will be predictable, but there are persistent predictions of a higher frequency of catastrophic events. These are massive externalities, making costings meaningless on this timescale.

Flexibility and adaptability to the impacts of climate change require power sources that can be rolled out in emergency response to global crises. Crucial considerations now are strategic, rather than financial. (That is not to ignore the figures, rather that the cheapest option may be a bad gamble.)

In the 1930s, building a giant passenger liner such as the Queen Mary Queen Mary , would have seemed like a good strategic investment. It must have costed out as the cheapest passenger-kilometre, (or soldier-mile to the strategists). However within a few years cataclysmic events blew away the future-planning, and replaced it instead with the massive rollout of the small and simple Liberty Ships , which economies of scale made even cheaper still.

Massive investments in liners in the late 1930s might well have had good costings analyses, but were simply bad strategy. To be fair to the capitalists of those days, they didn’t know for sure that WWII was going to disrupt their future-planning. On the other hand, we know darn well that climate change is going to disrupt every long-term plan.

We have been warned, so we have no excuse for planning on the basis of a non-existent future..


Cyril R. said: “Burnup is not the essential factor on its own. 100 GWd/t is a high burnup but if you need to use bomb grade uranium to get it, the reactor is a poor resource performer.”

Sure, a reactor burning at 100 GWd per tonne of enriched fuel, and with enrichment of high fuel-grade, 20%, leaves a byproduct of depleted (say, 0% U235) uranium. However each gigawatt-year of production leaves only 3.65 tonne of high-level waste in the form of once-used fuel rods, and 100 tonnes of relatively innoffensive doorstops, paperweights etc from the DU. If we are to account primarily for high-level wastes rather than resources used, such a reactor would be rated highly.

A CANDU reactor running natural uranium at 8 GWd/t would leave (365/8=) 46 tonnes of high-level waste for every gigawatt-year. In comparison, it would be a faint virtue to be leaving no depleted uranium.

IRIS (Gen III) is planned for 10% enrichment and 80 GWd/t, so 1 GWa would be leaving 1.25 tonnes of high-level waste and 18 t of DU, by my reckoning. So Gen III does look like an improvement.

Using the rule of thumb of 1 GWa/t in the post, a recycling fast neutron reactor would extract 365 GWd for 1 tonne of high-level waste, and leave no tonnes at all of depleted uranium. However that does assume a) a fast neutron reactor, and b) full recycling.

Small fast neutron reactors lack the geometry required for a high breeding ratio, but still achieve a high burnup. Toshiba’s 4S is planned for 34 GWd/t at 20% enrichment. That’s still a rate of only 11 tonnes of once-used fuel per GWa, so even the smallest fast neutron reactors look good.

GNEP planned a worldwide rollout of small fast reactors, with international traffic in recycled fuel. GNEP hasn’t been cancelled, but it is stuck on the backburner of the Obama administration.


Correction: that should be 4.56 t out of IRIS. That is, 1 GW from IRIS would leave 4.56 t/a of fuel rods to be disposed of as “high level waste”, in the absence of reprocessing.

Sure the hot waste is a problem, but it is political rather than technical. And it is trivial compared to what the equivalent CO2 would have done.


Roger Clifton, I feel that this is somewhat false advertising. The high level toxicity of the spent fuel is in the fission products and the transuraniums, not in the uranium. All reactors make more or less (+-30% depending on electric conversion efficiency) of the fission products that are the biggest radiotoxicity hazard.This is what makes the heat that needs to be cooled and what makes it so toxic. Enriching fuel does’t matter much. In fact with higher enrichment there is less fertile which hurts breeding which makes you need more fuel. This, combined with the fact that faster neutrons will leak out of the core more than slow neutrons, is why fast reactors need fertile blankets to breed well. IFRs and LFTRs can use the transuraniums that we have today as good startup fuel and they consume or don’t make any (respectively) in operation, so that puts the short term radiotoxicity and heat about the same but with a faster drop (fission products decay rapidly).

The space taken up by the fuel rods is a non-issue. Journalists often ask questions in articles they produce, like, do we have enough space to store the waste? This is a total joke as the waste is several orders of magnitude smaller than any other energy system, including coal, solar, wind. A golf ball of waste for a lifetime of energy or a soccer ball, both are tiny for a lifetime supply of energy.

If IRIS gets 80 GWd/ton with 10% enrichment that is very poor. LWRs get 60 GWd/ton with less than 5% enrichment and would be pushing over 100 GWd/ton with a 10% enrichment.

For cycles that reprocess sustainably (ie not MOx LWRs with low conversion ratio), the burnup becomes unimportant in resource use. IFRs and LFTRs can use low burnup and still use very little uranium or thorium resources as they recycle the fuel rather than throwing it in a spent fuel storage cask just when it starts to breed some nice amounts of fissile. But it is mostly an aesthetic issue. LWRs are so much better than coal, and they make good startup fuel for Gen IV reactors as a byproduct so it is a win win to build as many of these now as we can.


Peter Lang, you’ve misunderstood. I agree with discounting on a discrete private investment and that the LCOE is very useful. This does not mean that we can apply the same discount rate to planetary decisions such as climate change over a multi century timescale and pretend that this will make sure we still have a habitable planet centuries from now. Indeed it can be shown by an appeal to numbers, as you asked for, that the world is worth nothing a 100 years from now in the economic mindset. For example with a quite modest real discount rate of 5%, the world a 100 years from now is worth (0.95^100=) 0.006 of today’s value, or about half a percent. That’s essentially nothing. If your car is worth half a percent of its new value, you would certainly consider to buy a new car and send the old one to the scrap yard.

I was merely pointing out this dillema of long term valuation, that it exists and we don’t have an answer to it. The world in 2100 is fully depreciated in the economic mindset and this implies we should just get a new planet, as we would buy a new car when its depreciated. I’m not trying to judge anything, its merely an observation of a very real problem.

There is a way out of this problem. The dillemma can be avoided at low cost using nuclear power and electrification of society, with many environmental, geopolitical stability, and health benefits as well. Costs of nuclear will go down with a massive buildout and focus on a few standardized reactors as is proven in France for example. We just have to allow it since in most countries it is prohibited by direct law to build nuclear plants.

If we frame the problem and benefits of the solution in a broader way than climate change we have a more robust case. In my opinion, it is a mistake to frame the energy fossil fuel use problem as a purely climate change problem. It helps to show the incredible energy density of nuclear power compared to coal and biomass, to relate as an example to this thread. Rather than hammering the problems of climate change and scaring people of, it may be more productive to focus on the benefits of nuclear power compared to its alternatives. That, after all, is how people make most of their choices in day to day life.


Cyril R derides the IRIS figures. On checking, I find that I have been sloppy — my figures in the previous comment should have all carried the qualifier, “thermal”, except that the fast reactor figures are based on its “electric” production.

However, IRIS’ burnup of 80 GWd/t (th) at 10% enrichment seems quite reasonable. At 200 MeV per fission, 1 tonne of U235 would give 936 GWd, according to Wikipedia. So 10% of that is quite close. Moreover, IRIS is a small reactor at 335 MW, so we must expect more neutron escape and somewhat less efficiency at converting U235.

The burnup rate reflects not so much on immediate costs, as fuel is cheap, but on the amount of once-used fuel that has to be stored and apologised for as “waste”. (Of course, fully recycling fast neutron reactors have 100% burnup and only fission products to bury.) While we continue to build slow neutron reactors, we continue to pile up once-used fuel. And okay, some only-twice-used fuel, which about as far as MOX can go.

However, burnup is increased in proportion to the enrichment. Further, enrichment allows reactors to be designed smaller – and faster. For example, 4S has a lifetime of between 10 and 30 years, depending on how fast it has been pushed. This is more like a battery than a power station.

Now, with a design life of less than a human generation, we are back in the ballpark of planning on a timescale we can predict.


Leave a Reply (Markdown is enabled)

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s