Of the 50-odd nuclear plants currently under construction, around 1 in 3 are Russian VVER designs, being built by Rosatom. Sanctions on the supply of all kinds of electronics mean that few of these will be completed on time, if ever. in promoting sales, Russia has relied heavily on concessional financing through Sberbank, which is also sanctioned. That’s going to make future sales just about impossible, and create big difficulties in fulfilling existing commitments.
With the exception of the EPR money-pit, the only remaining large reactor design still in the market is China’s Hualong One. Given the experience with Russia, buyers outside China may well be cautious about this option.
So, if there is any chance for new nuclear, it rests with Small Modular Reactors, none of which actually exist (there are small reactors, but they aren’t modular, that is, mass-produced).
{ 22 comments… read them below or add one }
Chetan Murthy 05.21.22 at 4:54 am
“it rests with Small Modular Reactors, none of which actually exist”
Indeed.
God, it’s such a farce. The idea that people who pretend to be engineers can just prattle on about these designs that nobody’s actually built, much less built numerous instances of, as if they’re already mass-production-ready, is just ridiculous. Or thorium: yeah yeah somebody tried it half a bloody century ago, and maybe somebody’s trying it right now. But saying that’s real, is like saying that AI was real in 1980, my dude.
Such a farce.
Brett 05.21.22 at 5:14 am
The cost per kilowatt-hour on those is pretty unfavorable compared to large-scale nuclear plants, never mind solar and wind. Maybe that could come down if they could actually mass-manufacture them, but that’s not really feasible with most regulatory set-ups in potential buyers.
It’s a pity, because it seems like small-scale nuclear could be really useful. You don’t need as many safety precautions, because the thing just can’t cause that much damage even if it’s blown wide open. You could swap them in for industrial heat and electricity currently generated on existing industrial sites by burning natural gas, or even figure out how to use them in ships for power.
Jesse 05.21.22 at 1:13 pm
That is my big worry. If we don’t have Nuclear to give reliability to a clean energy supply (beyond the limited amount hydro can do) we will stall out once we get to 60-80% VRE electrical production. And with the lower reliability, we will also stall out on overall electricification of the rest of energy. This will be due to a combination of resistance on land use for more VRE and transmission, the uneconomic nature of the storage needed to long term stabilize etc.
Even hydrogen production from ‘extra’ VRE power will never be economic even if the electricity is ‘free’ due to the low utilization of the electrolysers (it will be MUCH lower than the CF of the energy supply as the priority will go for immediate use).
All that being said, we in the west don’t seem to be able to execute on single site 10+B$ projects anyway. Where $1B range is chemical plant size, which we do reasonably well on, with many per year being built currently.
I’m optimistic for the larger SMRs, I think the ideal point for economics will be as big as possible while still relying on passive safety, (and passive circulation is a bonus), complete vessels being shippable, and ‘off the shelf ‘ balance of plant equipment. I think ~300MW looks promising there.
Molten salt with integrated thermal storage will also help for grid integration, especially with high VRE penetraion.
Gary payes 05.21.22 at 1:31 pm
Any thought given to utilizing the small IFR concept? Several nuclear cycles exist which would be capable if using existing spent nuclear fuel to run, with the remaining (significantly smaller quantity of) transuranics having significantly shorter half life compared to existing fuel cycle wastes.
One possibility would be to convert select existing power plants to become nuclear waste processing plants using these new reactors to generate power while processing existing reactor waste, since the power infrastructure is already on site
A J Paxton 05.21.22 at 5:54 pm
‘EPR money pit’ – explain, please!
John Quiggin 05.21.22 at 11:05 pm
The EPR is a French nuclear reactor design, notorious for projects running way over time and over budget. The archetypal case is Flamanville, originally projected to be completed in 2012 at a cost of 3 billion years, still unfinished today with a cost of nearly 20 billion. The UK is building a couple more, again at massive cost and behind schedule.
John Quiggin 05.21.22 at 11:07 pm
@4 People have been talking about IFR and other “4th generation” concepts for at least 20 years, but none has progressed even to prototype stage.
Ronald 05.22.22 at 2:16 am
Jesse, 15% of Norway’s road vehicles are electric and represents around 34 gigawatt-hours of storage on wheels. That’s equal to a Tesla Powerwall per family. There’s no need to worry about variable generating capacity topping out at 60-80% of generation.
Jesse 05.22.22 at 12:59 pm
Hi Ronald,
EVs will be helpful for short term variations, first by flexible charging and maybe a bit by V2G. This will cover minutes to hours variability. With electrified heating a powerwall equivalent capacity is still only a few hours. EVs will not help at all with seasonal storage or low VRE periods of days to weeks. People will still need to drive and if the available power is less than demand over days, the car fleet will not be able to be recharged. Would you let your car put power back to the grid when the forecast shows you won’t be able to charge for days? Also the storage requirements for 100% VRE will be in TWh even for a smallish country (hundreds of TWh in total). To avoid this much storage MASSIVE overbuild would be required, which will run into even more land use and economic constraint, like >50% power curtailed levels of overbuild.
Norway does not need nuclear though, they are nearly uniquely advantaged with huge amounts of reservoir hydro relative to demand. The could add roughly as much wind as they have hydro power as reservoir extenders and probably have the capacity to fully electrify their energy with those syses only. Other than the fossil fuel they sell to pay for it.
Christopher Landee 05.22.22 at 3:06 pm
What is the “EPR money pit?” Are you referring to the international Fusion project in the South of France?
Ronald 05.22.22 at 10:57 pm
Jesse, a 50% overbuild of 2 cent solar is 3 cents. That’s not exactly onerous.
Chetan Murthy 05.23.22 at 3:08 am
Jesse @ 9: There’s two kinds of shortfalls: those that last a few hours (less than a day) and those that are several days long. For the former, as you agree, EVs and home powerwalls can do the trick. For the latter, that’s what wide-area grids are for: so an area with terrible weather and hence reduced solar/wind generation can get power from other areas with better weather.
If you really just mean “nukes need to be part of the solution” then …. well, just say so, eh? B/c your argument isn’t one of covering for short-term gaps, right?
lurker 05.23.22 at 8:49 am
@Roland, 11
Norway is well north of the lower 48 states, you’d need ridiculous amounts of overbuild for solar production to be relevant year round. Wind power is the way to go there: it varies but periods of low productivity do not go on for months.
Jesse 05.23.22 at 10:35 am
Ronald, 50% curtailed is 100% overbuult for average. It would still need days of storage just not weeks in many climates. I should also note quoted LCOE is typically on a plant basis, not a grid basis. Even if 0.02 from a plant, that is only when it is producing, and does not include everything neccisary to make it reliable. There is a curve for each case of overbuild vs storage and the storage comes down frustratingly slowely. I should add that solar needs more like 5x+ overbuilt in norther winters. In Ontario it can be less than 5% CF for a week or more.
Chetan, yes my argument is not for overnight level gaps, it is for the longer low output periods.
I like to think from system design cases, what are the most challenging conditions to cover, and think about the alternatives for covering it, and then given they elements used to cover it think on how should we use those elements the rest of the time, and what would make sense to add to complete decarbonisation.
My design case here (Ontario where i can get some data easily and follow closest, moderately northern, currently cooling peak but will be heating peak when electrified) is for a cold winter morning when wind has been <15-20% CF for a week and <20-25% for a month. Solar does basically nothing, and hasn’t dpne anything at all for 15 hours right when peak demand hits as businesses start up. To avoid using NG backup assuming all VRE I need about 125 GW of VRE and 100hrs storage as the lowest price mix to cover the 20GW of current demand in that case. I can get hours down but need a a LOT more VRE to do it. To electrify everything we will probably need to bring that peak up to 50-60GW, with heatong being the main driver (pumps dropping in efficiency during the design case) and even with shifted EV charging it can’t be shifted long enough to drop weekly averages. That would bring us up to near 100x as much VRE as is already installed, and there is already politically effective backlash, I do not think anywhere (other than maybe offshore wind) will get to this kind of build density before the backlash shuts down expansion. Note also that a high penetration VRE grid has the VRE eat its own value as the CF is highly correlated within type, and when curtailing value goes to 0.
Of we can’t build enough VRE how do we cover the design case? NG with CCS as backup has a brutal cost structure, and as we get to NG as a backup only (hydrogen would be high storage but maybe can help here, but that puts even more of the yearly variability on NG) but the year to year demand fluctuations for NG are going to be enormous, and it is hard to believe we will pay to maintain NG supply infrastructure (Storage and extraction) that will be in the money once a decade or so. Things that don’t get used, don’t get maintained, especially if they are an entirely parralel system to the main one. This is where the stall out scenario comes from. We start building this case, and as incramental capability becomes more and more uneconomic and unpopular, we just stop and use NG for 20-40%.
Hydrogen off VRE that has the first cut going to power has brutal economics as the Electrolyser CF is going to be very low for any reasonable quantity. VRE is very spike, so hard to match off take capacity with how much will be available.
Instead a system that is nuclear dominated, with enough capacity to cover daily peaks would be able to also power electrolysers sized for the difference of daily peak and average output with ~90% CF. This keeps the nuclear also above 90% output while doing something productive. If something unexpected happens, curtail the hydrogen in the short term to cover the issue.
Jesse 05.23.22 at 12:55 pm
I should add, while continental scale grids will help, there isnstill a fair bit of correlation in VRE availability at that scale. It also adds instantaneous geopolitical risk, where with Uranium in the case of a well built out nuclear system will have many suppliers and lead times are a year+, so much more time to sort out geopolitical changes in time to keep the fuel supplied.
Transmission is even more at risk of buildout stalling as its built through many places, any one of which can block it. See recent failed lines in the NE US for example. There is also the increased transmission losses from really long distances as well.
The big question on overbuild is how far below average capacity can the output be for a week or month time scale across the well connected region? If a typical grid has peak = 1.5x average demand, wind needs 2x peak overbuld and solar about 3x peak overbuild with infinite storage, to cover the average use. If peak lines up with low availability times. Say meerly 1/3 less than average as in my design case, now we need 4x overbuild for wind and ~8x for solar (assuming 33 and 20% average CF respectively). That gets to 50% of overall output curtailedas average output is 2x average demand. Still need a fair bit (days) of storage as availability can drop well below that sort of level for days to week time frame.
Scott P. 05.26.22 at 4:39 pm
Things that don’t get used, don’t get maintained, especially if they are an entirely parralel system to the main one.
Not necessarily true — the Thames Water Barrier has multiple parallel systems ready to go in case power gets cut, power has never been lost in over 4 decades of operation, and yet those parallel systems still work perfectly.
Dragon-King Wangchuck 05.27.22 at 12:02 am
I used to have hope for SMRs, but they’ve wasted too much time. SMRs are done. There’s no path forward remaining for them.
NuScale’s current timeline has the FOAK plant generating in 2029 and fully commissioned in 2030. The promise of SMRs is to lower the cost through mass production, but until they have mass production, costs will be even higher than what we’re seeing for current fission plants. The target LCOE for that plant is $65/MWh – and that’s before the current supply chain woes and rampant inflation.
$65/MWh is not actually cost competitive today. PJM’s 2021 cost of power was $63.97. And that’s up almost 50% from 2019, so it’s an outlier high in a high cost market. If power markets can normalize to the post-pandemic load shapes, this is going to drop. And that’s the price now – what is this going to look like in 2030? Solar gets cheaper every year – $65 is already more expensive than current utility scale solar and only slightly cheaper than current utility scale solar with 4 hours of storage (as per the NREL 2021 Q1 benchmark. By 2030, solar plus storage will beat the pants off SMR.
And there’s the nail in the coffin for SMRs. For SMR to get their price down, they need orders so they can mass produce and get the economies of scale. And who is going to order a bunch of SMRs when solar plus storage is much cheaper? And I mean cheaper even than the projections made by NuScale – IOW, the cost assuming this FOAK nuclear plant is built without going overbudget – a fantasy scenario with new nuclear builds. In 2030, solar plus storage will be mature and highly trusted, it won’t come with catastrophic disaster risk, is way more flexible and can provide grid firming services, and is going to be a lot more socially acceptable than what will essentially be experimental nuclear reactors.
They wasted too much time. If their FOAK was close to built and scheduled for energizing in the next couple years, I can see them getting some orders. But by 2030, they will be just a sideshow at best.
Ronald 05.27.22 at 3:20 am
Lurker, it may seem odd, but solar may be more cost-effective than wind in Norway, despite solar in Oslo only producing 60% as much energy annually as in Sydney. This is because solar can go on roofs and save households and businesses the retail cost of electricity. Solar output in December is basically zero but this is not a problem thanks to Norway’s massive hydroelectric capacity. In flat Finland, this is more of a problem, but they can upgrade connections with Norway.
Of course, if Norway continues to improve its grid connection to the rest of Europe, they may want a lot of extra wind power to help meet winter demand.
derrida derider 05.31.22 at 4:50 am
Count me as another one who hopes SMRs come good – which they may. I think though it’s a little say “they’re too expensive because no one’s mass producing them” as an argument against mass producing them – you would, after all, have said exactly the same thing about solar panels 15 years ago. If they work and you need them just bear with the higher cost until you are mass producing them.
And I think some parts of the world might really need them for that last 20% of renewables. Sure, Norway has the hydro resources to get it – and perhaps some adjacent countries – past a dunkelflaute (google it). But many countries have a climate where long dunkelflauten are likely and with little opportunity to access the massive storage required to overcome these in a fully renewable grid.
I hasten to add that Australia – John and I’s home – is not one of these. We can fairly cheaply go 100% VRE if only because of our size and latitudes – the sun always shines somewhere on the dry continent. We can afford to stay passionately anti-nuclear but others may not be able to.
Thomas Jørgensen 06.17.22 at 10:37 pm
https://app.electricitymap.org/zone/AUS-QLD
Australia is currently running 14 gigawatts of coal power. Yhea, totally would have no use for a reactor or seven. That solar transition is really solving the issue in a timely fashion.
That was sarcasm, but holding up Aus – Which has.. well, the dirtiest grid in the first world, as an exemplar of the viability of non-fission approaches to cleaning up electricity production is… well, it does not really prove the things you think it does.
John Quiggin 06.19.22 at 6:14 am
The problem in Oz is the rejection of carbon pricing, usually by the same people who want to push nuclear https://johnquigginblog.substack.com/p/if-the-opposition-wants-a-mature
reason 06.19.22 at 8:58 am
I’ll ask my usual question – why are we not putting solar panels on supermarket car parks basically everywhere. It seems to me a very obvious and relatively cheap win/win. It provides not only a source of power (and supermarket car parks are the obvious place to charge EVs for people without dedicated car parking spaces) but shade for the cars. And construction is relatively easy – no obstructions.