Nuclear Scenarios

Scenarios for nuclear electricity to 2060 – Context

Back in April 2010, I wrote a post “Nuclear century outlook – crystal ball gazing by the WNA”. It looked at an interesting study, undertaken by the World Nuclear Association, that made some low/high bound projections for electricity production between 2008 and 2100, including nuclear, new renewables, fossil fuels with CCS, etc. (see figure to the left). After describing the study in some detail, I noted the following:

One underlying problem with the NCO forecast … is the lack of explicit detail about technology type/role… What of the technological mix within the nuclear domain? (For instance, what is the likely proportion of Gen II, Gen III and Gen IV technologies, and how will that mix of contributions change over time?) What would such a massive nuclear build-out mean for uranium demand? How might nuclear power growth rates be constrained (or otherwise) by the availability of fissile material? On these seemingly rather important points, the NCO is, alas, silent. But that doesn’t mean it isn’t possible to make an informed guess as to the answers…

Well, motivated by some recent discussions, I am now going to write a series of posts on BNC to try and address these questions. (I’m not quite sure how many parts I’ll need to accomplish this!) The idea is that rather than doing a single (monolithic, detailed, lengthy, behind-closed-doors, indigestible) analysis, I want to treat this scenario mapping as an iterative and evolutionary exercise, where each new post builds on the last, and takes accounts of earlier comments and suggestions.

As such, this can be thought of as an Open Science experiment, conducted in the same spirit as those for For instance:

Open Science in its most basic form requires two things: (i) the clear and complete presentation of data and methods, and (ii) for the authors to care genuinely about the correctness of their work, and to act with due diligence in response to any mistakes or problems that arise, before and after publication… To practice Open Science is to embrace the critical analysis of your work by others, whoever they may be. This allows for fault finding in the first instance, and enables deeper understanding of the conclusions in the longer term.

I’ve also created a new category for this series, called ‘Scenario Analysis‘, and will, at some point, also back-edit some other past BNC posts that also fit with this theme.

Okay, the first step will be some projections of the build out of Gen III/III+ thermal reactors (i.e., advanced water-moderated reactors: PWRs, BWRs, HWRs etc.), over a 50 year time frame (2011 to 2060). I hope to post that up tomorrow. Subsequent posts will look at IFRs, LFTRs, synergies, etc.

At each step, I will be careful to lay out all of my assumptions and constraints (and do my best to justify them — remembering that these will usually be open to disagreement), my workings/tools (e.g., direct analysis, Excel worksheets, R scripts), and also to undertake some ‘sensitivity analysis‘ — to identify input parameters that most influence, or outputs that are most influenced by, a given assumption, constraint, or parameterisation, as well as to quantify a plausible fan of uncertainty.

Although the exercise will, by its nature, be somewhat technical, I will also write a layman’s summary at the end of each post, for the benefit of those BNC readers who (mostly) just care about the bottom line. The idea is to use the BNC blog as a kind of research tool, where I present some analysis, invite criticisms, update my models/thinking and progress to the next stage. Ultimately, it might even form enough coherent material to be worth writing up for publication in a peer-reviewed energy journal. We’ll see. But it’ll be a fun experiment, regardless of where it ends up.

Initial comments/thoughts on the process are welcome, but the first ‘meat and potatoes‘ gets posted tomorrow.


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.

42 replies on “Scenarios for nuclear electricity to 2060 – Context”

I think it will be a fun exercise, rather than a labour.

TerjeP, I suspect very badly. With that in mind, I used the word ‘projection’ rather than ‘forecast’ or ‘prediction’ advisedly! A projection can be correct insofar as the model, assumptions and constraints are correct. Making an accurate prediction is a whole different ballgame, because you have to get the assumptions and constraints right. Still, we might, together, get somewhere close to what is possible and perhaps even plausible.


“If we had made such a 90 year projection in 1910 it is interesting to ponder how well we would have done?”

If we had projected that the world would be powered by fossil fuels, as we probably would have, we would have been dead right.

I’d maintain that in broad terms the energy options for the next century are in plain sight right now. Are we likely to see some advance in fundamental physics that obsoletes all the current options? Anything’s possible but it’s equally possible (probable?) that the next century will not see advances to parallel the revolutionary effects of the advent of quantum and nuclear physics on the 20th century. Revolutions in fundamental scientific paradigms are few and far between. And look how long it has taken engineering and technology to exploit those, and we’re not finished yet.

Now, if we are talking the biological sciences, then I think things are not so clear and we may see some really interesting things.


Can I jump the gun with some numbers? I suggest a sobering benchmark could come from a projected ‘universal frugal middle class’. We could assume that by 2060 the world will level out at say 9 bn people none of whom is allowed to be in abject poverty. Relative to now we will have efficiencies such as electric transport and ‘inefficiencies’ such as desalination and essential-to-survive air conditioning.

I think it is fair to assume for practical purposes that cheap coal, oil and gas will be either depleted by 2060 or carbon taxed to irrelevance. Suppose each of those people needs 5 kw continuous energy production, justified below. 9 bn X 5 kw = 45,000 GW. The projected clean energy production figure in the WNA graph (if I’ve read it right) seems to be around 8000 GW. I guess some poor folks or their descendants are going to stay poor. Alternatively some demographic segments who are now well off (Brits, Aussies) will have to accept 80% ‘lifestyle’ cuts.

The 5 kw per person average gels with both MacKay’s UK 125 kwh per day and ABARE’s 3900 PJ/yr for 22 m Australians. Thus I think that ‘equitable’ world energy demand could reach 45,000 GW by the second half of the century. I don’t see it being met by either clean or dirty sources. Fossil fuels were Mother Nature’s one time gift and we squandered it.


I’d say 45 TW is out of the question by 2060 on a build rate factor alone. I suspect 5 TW of new capacity will be tough, 10 TW is what is needed, 15 TW ideally. So yes, in 2060 the world will still be an inequitable place (no surprises). Some thoughts on the matter here — it’ll be one of the assumptions:

The energy demand equation to 2050

My conclusion:

On this basis, the world in 2050 would demand 700 EJ in thermal energy, which translates to 290 EJ of electrical energy (which I round up to 300 EJ). This would require 300/0.03 = 10,000 GWe of generating capacity. As you can see, under some pretty heroic assumptions, we are likely to need a 5-fold increase in electricity generating capacity by 2050. If we assume all existing power plants (fossil, nuclear and renewable) will be retired by 2050, then we have to build 10,000/(365*40) ~= 680 MWe every day for the next 40 years (2010 to 2050), to meet this challenge. (By the way, the scale of the problem doesn’t diminish if you favour renewables or ‘clean’ fossil fuels over nuclear — indeed, it gets substantially larger due to overbuilding required for technosolar and the efficiency losses involved in carbon-capture-and-storage [CCS]).

By the year 2100, we may want double this figure again — to 1,400 EJ of thermal power or 20 TWe of electricity generating capacity — which would give the global population of 7 billion (let’s assume we stablise our numbers due to improved standards of living and education levels, and then gradually decline), a per capita energy use of a little less than the French enjoy today. This would allow for global economic growth (in energy terms) over the next 91 years of a few percent per annum, and agrees fairly well with the World Energy Council’s scenario A for 2100.


I was about to put something along the following lines in the FaD 7 thread, but thought it might be more appropriate here.

As part of my growing education, George Stanford provided a formula which allowed one to calculate the annual growth of Gen 4 deployment possible if one were to start with no surplus stocks of fissile material (at least that’s what I think the formula allows!) It must presumably also infer that the process of uranium enrichment has ceased because breeding makes it redundant.

The formula is GR= HC x (Br-1)/FI where
GR is annual growth rate (%)
HC is number of tonnes of HM lost/GWe-year (given as rough constant of 1 ). I am assuming HM to stand for heavy metal (ie unenriched uranium +/- thorium if using thorium breeding)
BR is breeding ratio and FI is fissile inventory.

George has also explained that a BR of 1.05 for an LFTR would give the same GR as a BR of 1.5 for an IFR because the IFR needs an initial fissile load which is 10 times greater.

He has also stated that a typical ALWR will need 150-200 tonnes of uranium (HM)/GWe- year. Consumption of uranium at this rate is clearly unsustainable if one takes affordable uranium stocks as 20 million tonnes and wishes to produce 10000GWe -years/annum.

However, if HC is an approximate constant of 1, one can forget sustainability with breeders for a few thousand years. The same goes for break even reactors except that they wouldn’t allow for expansion once existing fissile stocks (start loads) had been used, though, by definition, they could continue ad infinitum at their existing levels.

Key questions, therefore, seem to be how much fissile start charges are available now, who owns them, will said owners make them available for power generation and, if so, to whom.

If the current total of existing stocks of fissiles are inadequate to get where we want by 2060, we will need to create more. It seems that there are 3 ways of doing this:
1) By breeding more while concurrently generating power from the breeding reactors (Gen4).
2) By use of spallation.
3) By continuing to enrich uranium as we are doing now in the plants we are using now for as long as it takes to produce enough fissile material to get where we need to be.
The optimum method chosen would depend upon cost and, possibly for method 3, upon uranium availability.

I have attempted to repeat George’s lesson in my own way, though it is, no doubt, already well understood by most other readers. The reason is that I might have the wrong end of the stick and, if so, some kind person can correct me.

George is fairly obviously of the view that, sooner or later, preferably sooner, we’ll need breeder reactors. If the only choice is between LWRs and breeders, this is, without doubt, the case. However, depending upon how much nuclear electricity we need to generate and upon the current size of the fissile inventory plus that which will be added during the lifetimes of the present fleet of Gen 2/3 reactors, iso-breeding (breakeven) might serve as well as true breeding. One could also argue that reactors with a conversion rate of <1 would still give very good mid-term sustainability (a few hundred versus a few thousand years), but shouldn't be considered as other than a transitional means to the desired end of break even or full breeding.

The conclusions one draws would seem, to a great extent, to depend upon the amount of start charges that will be available for new generation reactors and I am hoping that Barry might come up with relevant information on this. Even without this information, it seems clear that we need a new generation of reactors as soon as possible, that they should produce power more cheaply the LWRs (inherent safety, modular construction etc) and address the waste and proliferation concerns of the anti nuclear lobby.


A minor point, but MacKay’s 125 kwh per person per day is the present UK primary energy figure. His 2050 guestimate is 50 kwh / person / day electric, (=125 kwh /p/d primary at 40% efficient) plus 7 kwh /p/d assorted biofuels. He assumes modest economic growth (times are going to be tough) and a lot of efficiency gains from electrifying things. Even so, it still ends up at 2 kw(e)/person, continuous, or about 20,000 Gw(e) for the world, which is Barry’s estimate for 2100. For 2060, even 1 billion at this level + 3 billion at half of it + 5 billion at a quarter, total 7,500 Gw(e), would be very optimistic.


Douglas, I’ll go along with most of what you say, but I have a few comments.

– First, a minor point. In that formula, HM (heavy metal) is thorium on up. BTW, the formula is only an approximation, but probably good to within 10% or so.

– HC, the heavy metal consumed per GWe-yr, depends almost entirely on the thermal efficiency of the fuel cycle. The approximation of 1 tonne per year assumes (I think) a thermal efficiency of ~33%.

– I don’t understand your remark that “If HC is an approximate constant of 1, one can forget sustainability with breeders for a few thousand years.” With a doubling time of 15 years, and given the fissile material that has and will come from LWRs\ (under a reasonable growth-and-phaseout assumption), LFRs could be providing a half-decent standard of living to a global population of
~10 billion by 2100. But not by 2060.

– With their smaller initial fissile requirements, break-even LFTRs could come on line more rapidly at first. But (again with a reasonable energy-growth scenario) their growth will cease before the end of the century because they will have exhausted the fissile supply — leaving none for the startup of breeders.

– As you say, the needed fissile could also perhaps come from elsewhere, such as spallation or fusion reactors. But the fissile need would be large — something like 1,000 tonnes or more per year for continued expansion of the LFTR fleet — maybe 100 or more times the current production rate in LWRs.

All this is very scenario-dependent, of course. I look forward to seeing Barry’s calculations.

— George


George, regarding this point:

HC, the heavy metal consumed per GWe-yr, depends almost entirely on the thermal efficiency of the fuel cycle. The approximation of 1 tonne per year assumes (I think) a thermal efficiency of ~33%.

I explained this in another comment a while back. To reiterate, for those who might be interested:

1 fission of a 239-Pu nucleus (bred from fertile 238-U) yields about 190 MeV of useable (non-neutrino) energy.

A mole yields 6.023E23 (Avagadro’s constant) x 190 x 1.602E-13 (joules/MeV) = 18.3 TJ of energy.

Thus completely fissioning 1 kg of 239-Pu gives (1000/239)*18.3 = 77 TJ = 7.7E13 joules.

Now, 1 GWh of energy is 3.6E12 joules.

1 GWyr (the output of a 1 GWe power station, run continuously over a course of a year) = 8760 x 3.6E12 = 3.154E16 joules.

So we require 3.154E16/7.7E13 = 411 kg of 238-U ‘feedstock’ (bred to 239-Pu and other TRU fissile isotopes) to deliver 1 GWyr.

Assume the IFR plant runs on a Rankine cycle at 35% efficiency operates at 90% capacity factor (in reality the efficiency and CF might both be higher), we would need 411*0.9/.35 = 1057 kg, or roughly 1 tonne of uranium.

The actual amount of uranium used will depend on the purity of the waste stream (residual actinides left with the fission products after pyroprocessing)

More here:


Thanks, George and Barry. George, I think we might have been at cross purposes over the term, sustainability. In this context i was using it, probablty confusingly, to imply that we wouldn’t be running out of uranium anytime soon so long as we moved to full or iso-breeding. I was not referring to the sustainability of our species which you may be correct to think may be predicated upon the availability of adequate fissile materials.


George, If the LFTR is designed for 1 to 1 conversion, there is not problem. If it is designed to achieve a 1 to 1.05 breeding ratio you get a U-233 doubling every 20 years. LFTRs can be started with U-235, and the excess stock of weapons gade U-235 in the American inventory can be used to start something over 300 GWs worth of LFTRs. This will mean that the entire stock of RGP can be devoted to IFR breeding, plus the 100 tons or so of WGP, in the United States’ inventory. If thorium is added to the IFR core, according to Indian researchers, the fissile inventory of IFRs could be lowered to 4 tons per GW. That will give you 100 GWs of IFRs to play with. In addition, there is still a considerable amount of U-235 in spent fuel, If the uranium in spent fuel is re-enriched to 19.75% in can be used in DMSRs at a ton per GW. That will give us an extra 300 GWs of thorium breeding DMSRs, plus more U-233 coming out of fast reactors for still more LFTRs.

IFRs cannot breed at ratio higher than 1 for ever. And there is no point to maintain a large fissile inventory if a small fissile inventory will do the job more efficiently, and at a lower cost.


Remember that the first cubic mile of oil that uranium replaced, it replaced by burning in Magnox reactors. Maybe it was only a cubic km.

“Start charges”? A Magnox reactor’s start charge is, you take some uranium, and you take some carbon …

I say for really big reactors, that can cover the fuel demands of, say, ten million cars each, the prescription can have helium substituted for carbon. Not so abundant, but still, no isotopic adjustment required.

(How fire can be domesticated)


I can’t see things working out as depicted in the WNA graph. I suspect world energy production will have plateaued by 2060 helped by draconian rationing and population control. CCS will never happen beyond demo stage and any case affordable fossil fuels will be largely depleted . Since new renewables will either be capital intensive or depend on permanent subsidies and gas backup I don’t see how they can increase out to 2100. By 2060 nuclear will have to dominate the energy mix or the world economy will fragment.

As to the output and technology type it’s hard to say. It’s hard to imagine future adults who are now middle class children graciously accepting massive reductions in average energy use eg from 5kw to 2kw. It’s also hard to imagine a fear ridden world with over 10,000 NPPs. Even if the problem is technically solvable the politics may not be.



I’ve been looking at your link to the UK energy calculator plus having a quick browse through the Pathways Analysis Report.

The Report contains a great deal of useful and interesting information. It has David MacKay’s fingerprints all over it . It continues to invite the public to offer its own recommendations, but assiduously avoids any mention of the cost implications of said recommendations although, in the annexes at the end, it lists fuel and capital costs of various energy generating technologies. Nuclear is recorded as having significantly higher costs than anything else, including offshore wind and CCS coal. Furthermore, these are overnight costs which omit finance charges.

Four levels of nuclear rollout include total phase out, new build of 33GW (of which 24GW will be replacing retiring nuclear plants and is the option stated to be the most likely eventuality), 99GW, stated to be highly ambitious and 150GW, stated to be just about within the bounds of possibility but highly unlikely. All plants built up to 2050 are assumed to be PWRs since it is assessed that new technologies won’t be ready till after then.

It is also stated that the government will enable nuclear to the extent of streamlining planning and simplifying regulation/licensing, but will do little else and certainly not contemplate subsidies.

I found all this profoundly depressing. I then started to play with the calculator, but it didn’t appear to respond to the pathways I chose to opt for by informing me of the implications. I’m probably too thick and innumerate to use it properly.

The DECC is inviting a public response, but the response seems to be dictated by the framework setting out its own uncosted assumptions. In other words, it seems to be designed to get the answers it wants while providing a cloak of democratic accountability.

What do you think? Are you going to respond? We have an energy minister who is avowedly anti-nuclear, but, as his party is junior in the coalition in which the senior party is pronuclear, he claims that he is honour bound to give nuclear a fair crack of the whip. However, his actions seem to belie his words.


I’m not so depressed at the cost assumptions, because I think they are per kw peak, not average. On that basis, onshore wind is comparable to PWRs, but unreliable and there aren’t anything like enough sites, and offshore wind is uneconomic by a factor of at least 2, not counting backup requirements. CCS loses because the capital costs are only slightly better than PWRs and coal is £80/Te to import..

I may have the rose-tinted specs on, but the report looks to be a softening-up exercise for an ‘all options open’ approach, including major nuclear build. It seems to have worked on our formerly antinuclear energy secretary already.

What implications are you expecting the calculator to tell you? It is a highly simplified model, the outputs are a CO2 emissions rate, a ‘difficulty’ score for what you are trying to do, and a note of your success in keeping us from freezing in the dark during 5 consecutive windless days in midwinter (like we got in 2009/10)



I’d like to be able to find somewhere to buy a pair of rose tinted spex like yours.

You suggest that the report and planner are a part of a softening up exercise. I wondered if that was the case when I first read MacKay’s book, but when he moved from his so-called Economic plan to one he named his Consensus plan (having joined the Ministry) which sharply downgraded nuclear’s contribution, I began to worry. I interpret the current activities as being aimed at shoring up this Consensus plan.

I was probably expecting too much of the calculator. I changed all the levels for each category simultaneously (supply and demand) and hoped the overlying graphs would alter accordingly to reflect the outcome of my personal accumulative plan. It was my intention, for example, to up nuclear and reduce energy imports, offshore wind and CCS coal but the calculator allowed the upping of nucler but not the other reductions – leaving me with too much supply for the demand I intended.


We should probably take this conversation to email. Ignoring the underscores, L_E_Collie_at_BTinternet_dot_com

The calculator will do what I think you want. Open it, then click on
The string of numbers in the url gives the model settings. This is the ‘easiest’ path to hit the target I’ve found so far. Lots of nukes, heat pumps and PHEVs, not much else. If they believe their own model, it’s telling them to go nuclear.


Charles Barton wrote:

IFRs cannot breed at ratio higher than 1 forever.

. Well, of course not, since forever would require an infinite supply of uranium (and other resources). So I suppose you mean there would come a time when they could stop breeding once we reach that Utopian day when the population has stabilized and everyone has enough energy, which is indeed correct. From then on, break-even operation — in whatever mix of reactor types is most economical — is all that would be needed.

And there is no point to maintain a large fissile inventory if a small fissile inventory will do the job more efficiently, and at a lower cost.

. Can’t say I understand this comment, Charles. It seems to be an argument against LFTRs — their small fissile requirements means that the existing fissile inventory would take 8-10 times longer to sequester in reactor plants than would be the case with IFRs of equal total capacity.


“Can’t say I understand this comment, Charles. It seems to be an argument against LFTRs — their small fissile requirements means that the existing fissile inventory would take 8-10 times longer to sequester in reactor plants than would be the case with IFRs of equal total capacity.”

I can’t say I understand this counter argument. The goal is to produce energy, not to sequester plutonium and other minor actinides in a reactor. The TRUs can harmlessly sit and wait in a concrete cask and be protected by the gamma radiation of the fission products they are intimately mixed with until they are ready to be used.

If the deployement of reactors is not limited by the total amount of fissile material available to start up new reactors you can deploy whatever combination of LFTRs and IFRs is faster, cheaper or better suited to its particular purpose(which may not in all cases be to produce electricity).


. You wrote:

The goal is to produce energy, not to sequester plutonium and other minor actinides in a reactor. The TRUs can harmlessly sit and wait in a concrete cask and be protected by the gamma radiation of the fission products they are intimately mixed with until they are ready to be used.

. Of course it is, and they can. Charles was the one expressing concern about the fissile inventory, not I.
. I agree fully with what you say. Nevertheless, more than a few people *are* worried about the fissile and TRU inventories, so Charles is not being unrealistic in that regard.
— George


George, If we look at global requirements for starting FI, they turn out to be quite large. The Indians, as you are probably are aware have evolved a three stage approach to the creation of a large enough FI. The Indian approach included the production of RGP by light and heavy water reactors, the use of RGP to start FBRs, the production of both Plutonium and U-233 in their fast breeders, and the use of the U-233 in thermal heavy water thorium breeders. Such a long range and complex program is required because the Indians lack a large FI to begin with and so have to build one up. On the other hand the FI in the United States, counting weapons stockpiles, is about twice what would be required to generate 100 % of our electricity with LFTRs. Assuming that we will electrify all surface transportation, construction, mining and agriculture, the present FI is probably large enough to handle all of that provided that LFTRs are used.

A IFR fleet would require a huge FI, 10 times what a LFTR fleet would require. The start up charge for an IFR, based on current plutonium costs, would probably run to at least one billion dollars per GWe. Charles is not being unrealistic, he is being frugal.

By the way, LFTRs are effective actinide burners. Since we are not aiming for the highest possible breeding rate, some of the extra neutrons that could go into breeding could go into nuclear waste disposal.


Charles, I’ll advocate for LFTRs whenever possible, but we need to be careful with the facts. This early paper from the French MSR group

does the neutron balance calculations for a MSBR-like LFTR started on spent fuel plutonium. The age of the spent fuel matters, because the excellent fissile isotope Pu-241 decays with 11-year half-life to useless Am-241. For 50 year old Pu, 7.8Te/Gw(e) are required for startup. Granted, they are using an old and conservative design, we can do better – but not seven times better. LFTRs can sequester and then destroy all the spent fuel actinides, but the power output available from this source alone is insufficient.

I still don’t think this matters, because 10,000 Gw(e) of LFTRs will need 10,000 / (0.5%) = 2 MTe of U-Nat if we can manage 1 Te U-235/Gw(e) and 6 MTe of U-Nat at a pessimistic 3Te U-235/Gw(e), and there is more than enough U-Nat. At $1000 /kg for U-Nat, it comes to at worst $600 million for the fissile, which is bad but not impossible, and needs to be compared with the cost of IFR-bred fissile.



. Interesting point about the Pu-241 (whose half-life is 14.4 years, not 11 — not that that makes much difference). I’ve tried to find the isotopic composition of used LWR fuel of some given age, but without success. Can anyone help?

. I have a problem with the 10,000-GW number, and here’s why.
. The US is currently using energy at the annual rate of ~0.33 quads per million people per year.
. The global population is now ~6.8 billion. Some projections have it leveling off at ~9 billion around 2060 (seems to me a trifle optimistic}, or maybe 11 billion by 2110.
. Say we want to see a world where the energy globally averages 0.2 quads per million people (60% of current US per-capita capacity). Then with 9 billion people, the energy requirement would be 1800 quads per year, or 60,000 GW. If provided by reactors, that capacity would mostly be electrical, so say the requirement is ~50,000 GWe.

. In any event, 10,000 GWe is just scratching the surface of the future need.

. As Barry has observed, there’s no way that we’ll have an equitable world by 2060.

— George


I am trying to understand George Stanford’s numbers. He suggests 50,000 GWe for 9 billion people in 2110. My calculator says that works out at 5,500 Watts/person.

I suspect that a very significant fraction of the human race will not have electricity one hundred years from now. My home did not get electricity until 1948 even though we were located in the “first world”.


I hope you’re not right on that call. Meeting all human needs is the only way to create the worldwide demographic transition the planet needs to stop population growth.

You’re talking about ‘natural’ effects that stop population growth… Easter Island events. Not fun.


Charles Barton,
If there is a rapid build up in NPP capacity over the next 50 years the price of fissiles has to sky rocket as you suggest. That would give a cost advantage to any technology with a low start up charge.

Absent new technological breakthroughs the LFTR seems to have the lowest fissile start up requirement. Now some genius needs to come up with a prediction of future fissile prices as a function of new reactors coming on line.

Twenty years ago, Babcock and Wilcox created an expert system that I used to optimize the design of power transformers according to the spot prices for copper wire and silicon steel. Eventually, we were able to design a medium transformer weighing 40 tonnes in 24 hours, something that used to take 13 weeks.

The same Design Automation software would be capable of optimizing NPP designs according to fuel prices. This might be important once the price of fissiles increases by an order of magnitude.


Back in 1948 we were dirt poor but did not realize it as nobody in our village had running water or electricity. I can remember throwing all those beautiful oil lamps out with a sigh of relief as it was my job to clean them!

Most of Africa is no better than we were bW


I seemed to have messed up my reply by pressing the submit button prematurely. Of course I agree with you. My sympathies are with those many people in the third world who will be forced to wait for electricity.


I vaguely remember having seen a paper from some french researchers(in english of course) discuss a releastic plan for how to scale nuclear power up most rapidly given a finite amount of natural U(I think they investigated the 5.5 million tonnes known resource we had a few years ago as the lower bound and the 27 million tonnes reserve base as the upper bound).

I seem to remember the conclusion as such: use all spent LWR fuel available from existing and future LWRs to build fast reactors. Stop building LWRs when molten salt reactors come online; start the molten salt reactors up on U-235, the fissile inventory required should be lower than that of the initial fuel load for an LWR. Use surplus fissile material produced in the fast breeders to start up more MSRs.

I might be misremembering some aspects and filling in shit on my own; but I distinctly remember thinking it was similar to the indian plan, except with molten salt reactors rather than the solid-fueled, heavy water moderated iso breeders(or close at least, depending on what neutron economy they achieve) the indians appear to be striving for.


Sounds like the ‘large scale deployment’ paper Barry mentioned in an earlier post. I expect he will be discussing it later in this series.

Thanks for the 1/2-life correction, shouldn’t post from memory, especially to an expert audience!

The Pu isotope mix is given on page 15 of the above linked paper. It obviously varies a bit with the initial enrichment and burn-up of the spent fuel. In some of their other work they assume a slightly different mix, together with minor actinides. Per tonne of mix, at five years after discharge:-

Np 63 kg
Pu-238 27 kg
Pu-239 459 kg
Pu-240 215 kg
Pu-241 107 kg
Pu-242 67 kg
Am 53 kg
Cm 9 kg

Which corresponds to a Pu composition of (3.1%, 52.5%, 24.6%, 12.2%, 7.7%) for Pu-238…242. They don’t give a breakdown for the minor actinides.

Projecting total demand 50 years out is a guessing game. Amongst many other factors is the effective exchange rate between future use of fission primary energy and present use of fossil fuel primary energy. This can be as bad as 2.5:1, for fission –> electricity –> heat replacing combustion –> heat in industrial processing, or as good as 0.6 :1 for fission –> electricity –> heat pump –> heat for domestic heating. Heat pumps with average COP>4 are available now, and they should get cheaper. The 10,000 Gw(e) figure is 25,000 Gw(th), so we only differ by a factor of two, not five, as your 50,000 figure is scaled from US primary energy. I’m expecting some efficiency improvements (more GDP per kwh), and the world is not going to remotely equitable in 50 years, though that shouldn’t stop us trying to do better.


Galloping Camel wrote:

I am trying to understand George Stanford’s numbers. He suggests 50,000 GWe for 9 billion people in 2110. My calculator says that works out at 5,500 Watts/person.

Luke wrote:

The 10,000 Gw(e) figure is 25,000 Gw(th), so we only differ by a factor of two, not five, as your 50,000 figure is scaled from US primary energy.


. Actually, my suggestion was 60,000 GW total energy, or about 6,700 watts per person. (Current U.S. consumption is about 11,000 watts per person.)

. I’m assuming that (ideally) reactors would displace a very large fraction of the global consumption of fossil fuel. The thing is, most of the fossil fuel is consumed to produce heat for non-electrical applications, whereas much of nuclear’s useful output is electrical (although direct use of nuclear heat, as for desalination, is of course important, which is why I scaled down from 60,000 total to an arbitrary 50,000 electrical ). Thus much of the space heating, process heat, and transportation would have to be provided by using GWe, using uranium’s energy with a conversion efficiency of ~30%.

. The 50,000-GWe estimate might in fact be too low, since providing transportation energy might be even less efficient than assumed. See the paper at

. And Luke, thanks much for the spent-fuel breakdown,

— George



That paper you linked in your last post was really interesting. I had absolutely no idea about the huge scale of energy available from the Uranium cycle. Somehow I got the impression that Thorium would turn out to be more important in the long run owing to its abundance in the Earth’s crust being about 4 times greater than Uranium.


George is being a bit mean with the thorium reserves. The World Nuclear Association,
gives a figure of 2.6 MTe, for available reserves at $80/kg, and 4.4 MTe for U on the same basis – but no-one has bothered looking for thorium much, as there is no market. As with U use in breeders, we don’t care much if it’s $8000/kg instead of $80. At that price, reserves will expand enormously, as with U, but no-one knows by exactly how much. Enough for millenia, perhaps forever, as with U. However, there is no thorium in the ocean, it does have to be mined. At the moment, it is recovered as a byproduct from rare earth mining, and as the miners have to pay to dispose of it, the effective price at he mine gate, in crude form, is $0


Luke, you wrote:

George is being a bit mean with the thorium reserves

. I don’t recall saying anything about thorium reserves. Like uranium, there also seems to be plenty of thorium to last us for as long as we exist.
. We have fertile material coming out of our ears. What has the potential to limit the rate of nuclear expansion is the availability of fissile material. If LFTRs can be shown to be better breeders (shorter doubling time) than IFRs, they might indeed turn out in the long term to be just what the doctor ordered.
— George


Sorry for any offence. In the paper you linked, of which you are co-author, there is a table of fertile reserves, which includes 20 MTe of phosphate rock U and 1.2 MTe of thorium. gallopingcamel’s comment suggests that he is interpreting the table on the basis that those figures are comparable as they stand, but they aren’t.

There’s enough U, and Th, in average granite that they could be extracted profitably to fuel breeders, if we had to, and vast quantities of richer ores to go at first. Fertile availability is a non-issue, whichever cycle we run.


Douglas Wise wrote:

The DECC is inviting a public response, but the response seems to be dictated by the framework setting out its own uncosted assumptions. In other words, it seems to be designed to get the answers it wants while providing a cloak of democratic accountability.

It is good to see that this aspect of the new UK energy calculator hasn’t been overlooked.


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