Nuclear is Still Cheaper and Safer Than Solar and Wind

I answered a question on Quora about why is nuclear power generating more power than solar energy ? Nuclear generates over 2500 TWh and solar is at 500 TWh and heading to 900 TWh globally in 2022.

Current bids for new nuclear power plants in China fell below $2000/kW in 2016. China’s Hualong reactor’s estimated per GW construction cost of US$2.5 billion.

China has 43 GW of reactors and will be completing 11 more reactors for about 13 GW of power. 40GW of reactors generated 285 TWh of power in 2018. China built those for about $90–110 billion. A GW of nuclear power can produce 8 TWh when operated with US level efficiency and capacity factor. China does not full operation out of some new reactors as they cannot operate a full year. Steady-state 40 GW can generate 320 TWh for full-year operation. China build is about half the cost of France’s build in the 1980s.

France built 58 nuclear reactors over 15 years and has generated over 400 TWh with them. The inflation-adjusted price was $330 billion.

France’s Nuclear Buildout Was More Cost Effective Than 2018-2022 Solar and Wind.

Germany spent $580 billion on solar and wind to get about 220 TWh. This was four times more expensive than France.

Global Solar and Wind has to increase by about five times to provide all of annual increase in electricity.

Solar and Wind need natural gas or some other power plant backup or massive battery farms that have not been built yet.

Even the 2019–2022 build of solar and wind is estimated to add 800 TWh for about $1 trillion in spending. It is more expensive than France’s build out. Double France 400 TWh to get 800 TWh and double the inflation-adjusted cost and it is $660 billion.

China can add 1GW or 7TWh for $2 billion to $2.5 billion. 800 TWh would be $220 to 270 billion. This would be four times cheaper than the projected solar and wind construction.
Nuclear is safer based upon actual deaths per terawatt hour and less polluting. Solar needs to use ten times the steel and concrete. Steel and concrete need polluting industrial processes to make. Solar uses twenty times the land.

Solar has industrial chemical waste.

Solar, wind, nuclear are all much safer than coal, natural gas and oil. The fossil fuels kill with particulates and other pollution.

Nuclear power did offset coal power usage. It is a historical truth for decades. This meant nuclear power prevented over 2 million deaths from air pollution.

Compare that to the 50 deaths at Chernobyl, zero deaths at Three Mile Island and zero nuclear deaths at Fukushima.

Nuclear power has been made even safer and with new molten salt reactors can be made even lower cost.

The Westinghouse eVinci 25 MWe reactor could get down the $2 per watt of electricity capital cost of China’s large nuclear reactors.

ThorCon molten salt reactor might be able to get to $1 billion per Gigawatt and could be mass produced at ship yards.

119 thoughts on “Nuclear is Still Cheaper and Safer Than Solar and Wind”

  1. The only fly in the ointment for SMRs is security. The nuclear regulator in most countries requires nuclear facilities to have very elaborate and expensive security. Douglas point nuclear generator was offered for sale years ago for 1 dollar but the local utility would not buy it largely because of the up scaled security system that would have been required to operate it.

  2. Nuclear reactors can manoeuvre a small amount but there associated steam turbines can manoeuvre about 40% if the unit has poison prevent capability. It does cause accelerated wear on the turbine but lately with all the installed solar and wind power many nuclear units are forced to operate this way anyway to accommodate the intermittent nature of solar and wind power.

  3. I worked at a Canadian nuclear plant for 33 years. It is by far the cleanest and safest place I have ever worked at. The Reactor I was in charge of was designed in the sixties and built in the seventies. It’s safety record is exemplary. The amount of electricity it has generated is over 80% capacity factor to this day. The site it is located in has eight reactors and could supply the electrical requirements for 4 million people 24 /7. The site occupies 600 acres and could have an additional 8 reactors. New reactors are safer and simpler and if utilities could get together and make large orders of identical models prices would come down just like anything else. Nuclear reactors certainly have a place in global warming. They make more sense than either solar or wind mainly because of there intermittent nature. New nuclear plants have a life span of 60 to 100 years. I’m sure by then there will be something better than nuclear ,solar or wind . I don’t understand how green organizations can be so closed minded and obtuse about such an obvious solution. There misinformation about Nuclear is really baffling. I’m sure nuclear scientists and engineers can come up with a design we can all live with at a price we can afford if we build enough of them.

  4. International physicians and other experts were intimately involved in the Chernobyl response. The fact that there is accurate and consistant data on health effects does not derive from trusting Soviet institutions.

  5. This article is dumb. The fact that nuclear replaced coal and ‘saved us’ from 2,000,000 deaths is a joke when used to compare this statistic with solar or wind ‘which did not do this’. Well it might have if it was older technology. Further, the argument that solar uses 20x more land for the same output is also ridiculous, it is still a tiny amount of land comparatively, and is nothing when compared to the vast, VAST amounts of land that have been destroyed, and polluted, for decades by nuclear disasters. Do not think for one second that because the media has not reported Fukushima in a while, that the damage is still not going on. It was huge.

  6. The less than 50 deaths at chernobyl are “less than 50 confirmed/documented deaths by the USSR”. The real number is much higher, but because the soviet party covered it up by attributing their deaths to other causes the real number will never be revealed.

  7. what’s the longterm cost of nuclear waste storage, what technology would you use, what storage material and where would you put , in the country using or the country mining it? what about thorium reactors?

  8. Cancer is not as deadly as you have been told. The human body is remarkably capable of repairing itself over time.

  9. Cernevoda 3 in Romania. Just went under contract with 51% funding by China. There is a worldnuclearnews article dated May 8, 2019. CGN is the Chinese company working with Nuclearelectrica.

  10. Yeah, don’t know much about that. The diagram came up automatically when I added the link in the combox. I probably should have deleted it. I only have a mechanical engineering background so the heat storage stuff is the only part that makes sense to me.

  11. The Wikipedia article does say that Canada doesn’t have any more under construction, and hasn’t been able to sell any for a number of years.

    The last one to be built was finished in 2014 in Argentina from what I could see.

  12. The design is wrong
    the containment systems should be below ground level
    In the event of a meltdown you can use gravity to pull water in to the tanks

  13. Because (in some locations) the voters want the power to be provided by intermittent sources such as solar and wind.
    Combined with the much greater cost of having enough backup to use such sources while still being super reliable.

    We can give you 99.9% reliable for $0.15/kWh, or 99.999% reliable for $0.30/kWh.

    Which do you think people will choose? Most will go for half price option and deal with resetting their microwave clock once every couple of months.

  14. There you have it folks. I didn’t pull it out of the air. Much better explanation, DrPat.

  15. Why should anyone have to pay extra for power priority ratings…. shouldn’t power be a basic like water, etc…

  16. Actually, it wasn’t the IAEA that gave the 93,000 number and I think the IAEA just released data on the total radiation which others not conected to the USSR (which was doing its best to down play the accident.) I’m really annoyed that anybody is claiming nuclear power is safer than solar and wind power. The claim that just 50 people died because of Chernobyl is just obscene. Cancer is a really nasty way to die.

  17. Lost a few $billion reactors and few $billion in land value and have to spend a few $billion to fix the issue. And no insurance company wants to insure them so the government has to back any big loses. And the worse thing is there are other designs that are inherently safe. Why spend good money after bad money.

  18. Canada abandoned the CANDU reactors. Some issues with the pipes that carried the heavy water.

  19. If you overbuild by 20% by adding extra reactors, I doubt you get much economy of scale. The overnight cost/kW will come down the more you build, but ultimately it is what it is.

    What happens when you make bigger individual reactors? Do the overnight costs scale linearly with power, or do that flatten out at higher powers?

  20. But having decaying isotopes in the storage tank relaxes the insulation requirements! 😉

  21. This BS. Nuclear may need less materials for initial construction; however, long term it has a massive mining impact compared with wind or solar.

    Current world demand for uranium requires about 20-30 mines operating to continuously produce 400GW of power generation through LWR tech. 1/3 the fuel is replaced every 18 months. It’s 75 tons enriched fuel per GW of power. The largest uranium mine on earth produces 650 tons per year.

    Conversely, the 2 year production at the world’s largest iron and copper mines is enough to produce over 150,000 wind turbines or roughly 450-500GW of power generation. That’s only 2 mines versus 25ish.

  22. Just to illustrate the point: Over a 1-month period, 1- (30,000*24)/(5.2M*24*30) = 99.98% availability. Pretty close to four nines right there.

    Even really big blackouts (and the two US Northeast blackouts were huge), fade into the background when averaged on a long enough time scale.

    To be fair, if we’re engineering a renewable grid to be able to withstand 4-sigma events, they’re probably about as rare as the big blackouts. But 2- or 3-sigma, while engineering the storage would be a lot cheaper, would significantly reduce availability.

    Monte Carlo sims are beyond my statistical competence, and I wouldn’t know where to get the weather data sets, but figuring out how likely running through a given amount of storage would be, seems to be an awfully good thing to know before committing whole-hog to an all- or nearly-all-renewable grid.

  23. So Brian will quote a bunch of unproven estimates for air pollution deaths and exaggerates the environmental impact of solar and wind, but for nuclear he takes the lowest possible figure. I guess you have to die fast from ARS or an explosion for nuclear power to be responsible. Cancer a few years later doesn’t count.

  24. Moltex has such a proposal using a secondary or tertiary salt coolant loop. I think they are the ones who just received $10 million for development.

    GridReserve is a collection of molten salt storage tanks that stores gigawatt scale thermal energy when it’s not needed for electricity production. When demand goes up, say when renewables are off, the plant can take heat from the reactor and storage tanks to produce electricity. This is just like in a Concentrated Solar Power plant and uses the same solar salt, turning a 1GW reactor into a 3GW peaking plant. 

    Many concentrated thermal solar plants like the one in the Mojave Desert already have this tech working.

  25. It’s not a disagreement. You said, “The International Atomic Energy Agency estimated Chernobyl to be responsible for an additional 60,000 cancer deaths across Europe.” That’s factually wrong.

  26. First of all, Chernobyl is [i]not[/i] N/A. There are plenty of squirrelly reactor designs sprinkled all over the world, and some of them are still online. (In fact, some RBMKs are still online, albeit with a pretty hefty fix after Chernobyl.)

    Second, callous though this may seem, the problem with nuke accidents isn’t lives lost; it’s property damage. Chernobyl and Fukushima together contaminated about 30,000 km^2 with more than 5 Ci/km^2 of Cs-137–that’s pretty much uninhabitable for 50 years or so.

    We have roughly 17,000 reactor-years of operation, so we’re looking at roughly 2 km^2 of land taken out of service per reactor-year.

    A square kilometer of land in a swamp in Belarus is a manageable expense. On the other hand, a square kilometer in New York City is worth about $2.4B–excluding the structures and other property on the land, so it’s probably closer to $5B total.

    It’s really hard to quantify how the rate of serious accidents has declined over time, because there aren’t that many of them, but it’s fair to say that the first 30 years of nuclear power’s history have a lot more INES >4 events than the second 30 years.

    That’s due to something. Some of it is simple learning curve, but a lot of that learning curve is codified in regulation. A lot of the regulations are no doubt stupid, but far from all of them are.

    Getting regulation right is a hard problem. Dismissing it as completely worthless, however, if foolish.

  27. I believe the actual number of homes blacked out in the most recent 24 hour blackout (not 12) was about 30 000 or so. So not remotely all of Sydney.

    I didn’t really have a problem once I convinced the children to stop opening the fridge.

    And by children I mean people in their 60s, they just act like children when it comes to not being able to go a day without cold drinks.

    The more famous South Australia black out took out the entire state except for one island. But that was storm damage rather than overloading on a hot day.

  28. Isn’t he linking to a report by the International Atomic Energy Agency, the very group you said claimed up to 93 000 deaths?

  29. Just because somebody you don’t agree with cites some numbers you don’t like doesn’t make them wrong. The Union of Concerned Scientists has simlar dire numbers.

  30. Did all of Sydney lose power, or just your neighborhood?

    System reliability = 1 – sum(each customer’s outage hours) / sum(each customer’s total service hours)

    The problem is that the 3- and 4-sigma weather events take down the whole renewable grid. That takes you from five nines to what feels like nine fives in a hurry.

    And note that I’m taking the high road on your “sort of a developed country” lay-up, just to prove what a woke non-MAGA citizen of the world I am. But it hurts leaving that hanging there…

  31. I know a guy working for a swimming pool startup.

    It seems that for most of suburban Australia, swimming pools (filters, pumps, chlorination machines with electrolysis and in the colder areas heating) add up to an average of 10% of the energy needs of the suburb. That’s 10% on average over a 24 hour period, maybe 30 to 50% if they all turn on at once. (I don’t know myself, I’m just going on the story he told me.)

    Their scheme (which appears to be gaining market share) is that they provide the pool systems, also provide a reselling service so the pool owners get the electricity cheaper than otherwise, and in return the company gets remote control over the usage. So the pool company has a dial that lets them turn the energy usage of the suburb up and down by some large fraction on a second by second basis.

    They then sell this level of control as a service to the power companies.

    • Want to cut power usage in this area? Give us a curve of GW versus time, to the nearest fraction of a second, and we will follow it (for a fee).
    • Want to dump a load of extra power? Same deal.
  32. Goat is talking about non-fuel salts in a secondary or tertiary loop.

    Not the fuel where I don’t think anyone wants a large external tank of decaying isotopes just sitting around.

    And as originally mentioned, there are plans to do the same thing with solar thermal.

  33. I live in Australia which is sort of a developed country.

    I’ve had the power cut out at my place this year (for about 12 hours once), and I’m in Sydney, the politically 2nd most powerful location in the nation.

    You get used to the idea that power is not THAT many 9s reliable. What are you going to do when the choice is the current government, or the opposition who are promising LESS power stations?

  34. There was an old saying of “iron men in wooden ships” referring to the old time sailors of the Napoleonic era and earlier. The ships were wood, and the men had to be unbelievably tough, as even a “good” voyage would be a national disaster by our standards.

  35. The only direct nuclear response I remember off-hand to the peaking problem, is

    1. a molten salt thermal store spliced somewhere into a conventional reactor steam island (something like a nitrate solar salt system)
    2. NACC-FIRES, which uses a resistance heated firebrick compressed air container for thermal storage, though the backend ends up looking like a conventional gas turbine peaker plant anyways…
    3. MSR with a large intermediate salt thermal store
  36. The lower melting point salt is the BeF2 at 550C. John (commentator) rightfully mentions a eutectic mix of BeF2 and LiF that melts at less than 400C. NaF, KF, LiF, all melt at like 900C. If we had the materials to handle it, straight UF4 would be the way to go, but like I said it melts at 1030C. UF6 is a gas at 133C. Put an F2 atmosphere on UF4 and it becomes UF6 – multivalent.

  37. Also from the UCS article:

    “If not handled and disposed of properly, these materials could pose serious environmental or public health threats.”

    And as some of the other links I posted show, it hasn’t always been handled or disposed of properly.

  38. You don’t consider the Union of Concerned Scientists to be a reputable source? Or the NYT or the Chinese news site?

    Can you show me any “peer reviewed journals” which conclusively state that all solar panels have no toxic materials, such as the ones listed in my sources?

  39. I’m not inclined to agree that the answer is in the middle. I think the level of regulation pre-TMI was about right. After all, large numbers of those reactors ran for a decade or so w/o polluting or having any significant release of pollution (Chernobyl is N/A). After 50 years or so of operation, the Fukushima reactors, which were of that vintage (and regulation level), had a significant release (which was due not to an accident or breakdown, but a 9.0 earthquake and a 45-ft high wall of water). And yet, Fukushima caused no deaths and any future health impacts will be too small to measure. Meanwhile, fossil generation goes on killing ~1000 people every single day, and causing global warming.

    Current nuclear regulations equate to nuclear spending many, many orders of magnitude more money per unit of public health and safety benefit (e.g., dollars per life saved) than competing sources of energy. I agree that objective risk assessments are needed, but with a level playing field between different energy sources (i.e., a similar level of allowable risk, or money required to be spent per life saved). Under such a system, and the resulting equilibrium (with, say, the same total amount of money spent on safety, but spent where it does the most good) it may be that nuclear’s “optimum” regulation level would be even lower than that which applied back in the ’70s (whereas fossil fuels’ level or effort would increase by orders of magnitude).

  40. I’m getting confused by “melting” and “dissolving” here, not knowing much about UF4. Do the solvent salts have to be above the UF4 melting point to keep it in solution? What if you use UF6 instead? (No clue how to figure the partial pressure in solution.)

  41. Storing heat in sodium nitrate is a good idea. Seems that it decomposes at 380C. LWRs are barely hot enough just to melt it, so not a fit obviously. The molten salt mixes are mostly solvent salts; MSRE salt was only 1% fuel by mass. This is because uranium tetrafluoride melts at over 1000C and we don’t really have metals that can hang with that for the type of duration we expect from nukes (uninterrupted years). You can start to smell fish with MSR when you realize that a big PWR has 87 tons of 4% initial enrichment uranium in solid form – that is the kind of inventory that gives 3.5 gigawatts for years. you can calculate how many atoms you have to fission to give 3.5 GW-y. then you have to adjust that mass for the enrichment, because what you’re going to calculate is just the fissile content. Conversion factor is 200 Mev / fission. I should really put my thoughts together into a 10-page ass kicking of the MSR. Maybe I could get some reads. Maybe I’ll start doing that on LinkedIn.

  42. IAEA press release from September, 2005:

    A total of up to four thousand people could eventually die of radiation exposure from the Chernobyl nuclear power plant (NPP) accident nearly 20 years ago, an international team of more than 100 scientists has concluded.

    As of mid-2005, however, fewer than 50 deaths had been directly attributed to radiation from the disaster, almost all being highly exposed rescue workers, many who died within months of the accident but others who died as late as 2004.

    This is the number from the Chernobyl forum, which did indeed have representation from Russia, Belarus, and Ukraine on it, but also had the participation (and blessing) of the IAEA, WHO, UNSCEAR, and a variety of other international agencies.

    The number you’re citing is from the so-called TORCH report, which was commissioned by a coalition of Green parties in the EU parliament, and authored by Fairlie and Sumner. It is not produced with either the cooperation or blessing of the IAEA.

  43. The other hypothesis that fits those facts is that old cheap nuclear power wasn’t pricing in the risk, and that more recent nukes are fairly priced to defray the risk.

    The answer is almost certainly somewhere in the middle. I wish we had a better process for transparently assessing the risk, and for reforming regulations to follow the real risk.

    Messy problems like this are, fortunately, metastable. Eventually, some new technology comes along that inherently reduces the risk enough that a reasonable regulatory regime is easy and obvious. But we could get to whatever that new technology was a lot quicker if we could convince the anti-nukers that performing some experiments was a good idea.

  44. You’re essentially arguing for an all-baseload grid–which may indeed be the right answer. My biggest concern is that if all of that baseload is nuclear, it’s expensive baseload.

    Once you stray away from the basic baseload / load-following / peaking model we use today, there seem to be five strategies that are fully decarbonized:

    1) Overbuild baseload capacity to 1x peak demand, then find fun things to do with it off-peak or, worst-case, ground it out. (Good strategy if the externalized cost of baseload is very cheap.)

    2) Overbuild baseload enough to charge storage, then use storage for load-following and peaking. (Good strategy if storage is cheap.)

    3) Build baseload capacity to baseload demand, then use a combination of renewables and storage to provide load-following and peaking. (This is where we’re heading in the near term, although the baseload will be mostly carbonized. Whether this is a good long-term strategy depends on everything decarbonized and a bit cheaper than it is.)

    4) Overbuild renewables (up to about 1x of peak demand), then throw lots of storage at it to make it 3-4 nines reliable. (Good strategy for kinda cheap renewables and extremely cheap storage.)

    5) Massively overbuild renewables, up to 3-4x peak demand, then throw small amounts of storage at the system to make it 3-4 nines reliable. (Good strategy for almost literally dirt-cheap renewables and moderately expensive storage.)

    If nukes were cheap, option #1 is my fave. But they’re not.

  45. This is a puff piece. The International Atomic Energy Agency estimated Chernobyl to be responsible for an additional 60,000 cancer deaths across Europe. Other have placed the estimate as high as 93,000 deaths. Nobody operating with a good conscience can claim nuclear power is safe.

  46. If you can fill up the off-peak times and manage to not throw too much into big resistors/load follow you still need to make fuel for airplanes, rockets, non-electric vehicles etc. and that means you need more capacity, not necessarily of the same kind of nuclear.

    Hydrogen can be somewhat efficiently made from high temperature nuclear with chemical cycles and you need it for rockets and for making ammonia and nitrates. The ammonia and nitrates you need for solid fuel rockets (ammonium perchlorate), explosives (often ammonium nitrate, sometimes sensitized with NG-explosives for non-war purposes) and farming (ammonium nitrate, ammonia, potassium nitrate etc).

    Huge tanker ships you could power directly with small nuclear plants. The biggest problem is regulatory. It needs to be ridiculously safe during operation to be accepted; not the HEU kind of reactor used in subs, ice breakers etc, but maybe graphite pebble bed. They have a diesel generator -> electric -> mechanical drive train anyway.

    Trucks and cars; that’s hard. Batteries might work out, might not. Pain-in-the-ass and inefficient power-to-methane or methanol or DME we can do now; but it’s not cheap and not efficient.

  47. …cont

    Uses like electrical processing of iron (experimental) or aluminium require reliable electricity. You must never interrupt supply; frozen aluminium in the electrolyzers is awful and must not happen. Capital cost is large, so operating 20% of the time at random times of day is not acceptable. Operating at full capacity 80% of the time and reduced capacity 20% of the time; that’s not so bad (remove less heat from the aluminium during peak electricity demand, so it does not freeze). Shaves the peak a bit by having a small through in aluminium production mid-evening. Can reliably plan to do this every day. If you can get half-price electricity for doing this and have the peak-load electricity users pay for it, that probably works out great.

    Over time you want to keep trying to identify and add processes that can operate at different capacity factors without the capacity cost killing them. Solid state ammonia synthesis is a nice candidate. Car battery storage is a nice candidate (you need the battery for EVs anyway, so it’s not the same thing as a grid battery).

    On the electricity storage side you try to identify a good opportunity to make high temperature storage for peaker plant operation. That just makes the economics sweeter if you can do it; it’s not necessary. This necessarily means high temperature nuclear (e.g. molten salt or graphite pebble bed) and e.g. molten salt (not fuel) storage.

  48. It’s not that terribly hard, since you get reliable electricity. It’s not season, time of day and weather-dependent electrical noise. Even if you just dump the surplus electricity into a giant resistor you only need to overbuild capacity by ~90% above average load to match electrical use on the current grid and deal with the capacity factor which is often 80-90% today. Not very cost efficient, but not completely abyssmal.

    Some seasons require more electricity (heating/cooling). Try to do planned service in the off season. Then you casn overbuild a bit less.

    Don’t use grid electrical storage; if you happen to have reservoired hydro that’s really nice.

    There are decent uses for off-peak electricity since it is reliable. You can charge some electrical cars overnight, but don’t worry too much about transportation at first; it’s the hardest problem to electrify.

    You can play around with peak load a little to reduce it (things like big industrial freezers can run down the temperature a few degrees extra before peak and let the temperature rise back up during peak).

    A lot of energy goes into heating and cooling. If you live in a city; district heating makes a lot of sense. If you live in a suburb; heat pumps with some heat source like the ground water or air makes sense. COP is about 3-5 most depending on temperature difference. 1 m^3 can of hot water can store ~29 kWh thermal in a 25 degree C temperature swing. Per capita needed accumulator size is tiny.


  49. I asked the same question of GG up above: does the cost of a nuke scale mostly as the thermal output of the reactor, or as the balance of plant? If the former, the parasitic storage system is still really expensive. But if it’s the latter, you’d expect the extra output to scale more like a vanilla-flavored molten salt storage system, which isn’t too bad.

  50. Makes sense, assuming that you have MSR deployed.

    The big question is how much of a nuke’s cost scales as the thermal power output, vs. the balance-of-plant stuff. If most of the cost is in BOP, then higher thermal output and dumb molten salt storage should flatten the $/kW scaling quite a bit for the part of the output dedicated to storage.

    scaryjello, do you know the answer here? (I know you’re not a huge fan of MSR, but what do you think would happen to the scaling in this case?)

  51. Could… but you also could over-spec the nuclear to account for the nominal high-demand conditions, and implement a full power-fabric-wide smart cut-out system, as outlined (grossly) in my answer to DoctorPat.

    Much more effective. 

    Just saying,
    GoatGuy ✓

  52. I think the idea is that you can melt thousands of tons of an innocuous high-heat-of-fusion salt to molten state in the 3rd loop.  

    Well outside of any possibility of becoming radioactive. The molten salt is just stored in VERY well insulated hillside tanks, as you see at refineries. It is amazing how well 2 feet of rock-wool will insulate a huge tank. Ought to stay molten for months.  

    This is an idea that is uniquely suited to MSRs, from the point of view that they naturally ‘run’ at far, far higher temperatures than either boiling or pressurized water reactors. 600°C to 900°C, typically. Some are specified at even more extreme (but dangerous) temperatures. Most I’ve reviewed seem to hit a decent compromise at the “dull red glow” heat level. 


    Is hot enough. And pretty darn safe. 


  53. But, but, but… Doc!

    The impartial hand of Socialism must certainly be the solution, right? REQUIRE…

    • № 1: All houses install high-amp appliance “smart cutouts”
    • № 2: All businesses install high-watt interior lighting smart cutouts.
    • № 3: All high-demand industrial production defers to peak-social demand
    • № 4: Electric street-trolley service has mandatory slow-downs
    • № 5: All domestic, commercial HVACe cuts back in stages…

    … 22° C (72° F), 24→75, 26→79, 28→82, 30→86, 32→90.
    … seriously, we EVOLVED to handle this for 100,000+ years.

    • № 6: All e-car recharging is halted during hi-demand levels
    • № 7: All e-stove/range is limited to 1 burner + oven. (surprisingly efficient)

    • № 8: Master cut-outs for houses exceeding 25% of switchbox plate ampacity

    … that gets people’s attention, real fast.  
    … having “red, yellow, green, blue” breaker lights works wonders.


    It would work (especially № 8). If new social justice and fair-use socialism is enacted, one need only look at one’s abode (or business) e-panel, to SEE which of the many breakers has high-loads. Which ones will be automatically cut-out in succession when tight-production threatens the whole grid. Won’t be pretty for high-watt server farms. Well … its about time to solve that, too.  

    Musing away here in the Socialist State of Kalifornia.

    Just saying,
    GoatGuy ✓

  54. Do like in France: they (originally, tho’ not kept up with it) planned for 133% of PEAK as their continuous nuclear power production (along with her scant hydroelectric resources). And then they built like crazy. Economies of scale, of streamlined rubber-stamped reactor production. Not one-offs. Not each-is-designed-from-scratch.  

    THEN, with all that extra non-PEAK capacity, they built a bunch of aluminum-from-ore smelters, and electric-arc steel-and-alloy production facilities. Which work almost exclusively at night and during longer low-power consumption periods. Using up a lot of the otherwise fallow power.  

    This is not to say that she planned to use ALL of the non-consumption power, but only some of it. The nuclear reactors however are also designed to ramp up, and down, and up, and down too. Not as much as “just following consumer load” would account for, hence the off-peak production of electrical-energy intense industrial products.  

    It is the smart thing to do.  
    As well-before-nuclear, the absolutely gargantuan Niagara Falls hydro was.
    Alcoa Aluminum (or its predecessor) put America’s first Al smelter there. 
    Because of all the OFF-PEAK and over-capacity-even-in-the-day of it. 

    Just saying,
    GoatGuy ✓

  55. It’s the capacity factor on the nuclear package that’s driving the cost; In principle you could equip a nuclear reactor with a molten salt heat storage system, in much the same way as you could solar, and just have excess capacities on the turbines. (In a molten salt reactor, this would be essentially built in.)

    I think the grid can absorb a little solar without any added peaking requirements, due to the power being naturally produced around peak consumption times. A few percent, maybe.

    And more time of day pricing would make sense, to shift at least industrial loads towards the normal lulls in consumption.

  56. The first paragraph of your first link says “without toxic pollution”. Did you even read the places you linked to?

  57. The Verge and Forbes contributors (i.e. not staff) are not exactly peer-reviewed journal. Have you got any reputable sources, rather than fossil industry shills?

  58. All these people here talking about the wonderfulness of nuclear———-the problems with coal. I haven’t read any negatives of WIND other than land space and cost. Any other concerns? Health issues, annoyance (distress), audible and inaudible sound, shadow flicker, decommissioning, conflicted officials, etc.) Can anyone posting here who lives in a rural community comment about their experiences of living with a 600 foot wind turbine located 1000 ft from their home? Or do all of you live in the city far far away from wind turbines and only experience them from the highway as you fly past at 75 mph?

  59. You and many others see the data in that graph as being damning for nuclear, but my takeaway is the exact opposite. The chart shows that nuclear is not inherently expensive, as we built it at a very low cost before. It is clear that expensive nuclear power was a *choice* (i.e., a choice to apply absurd levels of regulations, standards of perfection, and other burdens).

  60. You could just use batteries to handle the demand peaks in an all-nuclear system. The amount of battery storage capacity for a 100% nuke system would be far lower than that required for a 100% (intermittent) renewable system.

  61. Scientific consensus, from formal bodies like the UN and the World Health Organization, is that there were only 50-60 clear deaths. Statistically significant increases in cancers (other than highly treatable thyroid cancer) were not observed. That is, any effects were (and will be) too small to measure. Pessimistic theoretical estimates of (too small to measure) total eventual deaths top out at ~10,000.

    To put this in perspective, pollution from fossil power generation causes ~1000 deaths PER DAY, worldwide. That’s on the order of 10 million deaths over the ~60-year period that nuclear has been around. And, of course, Chernobyl is completely non-applicable to the merits of nuclear power, as practiced anywhere else.

  62. I have both solar panels and an electric car. It is true that my residual power bill is very low. So, I may be “electricity independent” but definitely not energy independent overall (heating costs, energy used to make products I buy, air travel, etc..).

    But it’s not like the solar array was free. It will take almost 10 years to pay for itself, even though it’s competing against very high CA utility costs (~25 cents/kW-hr), and despite enormous subsidies, such as the tax credit for 30% of the system costs and net metering.

  63. Nuclear is a bit worse, as it can not attain as high a temperature. Nuclear’s thermal efficiency is ~33%, compared to ~40% for coal and 50% – 60% for efficient natural gas plants.

    But who cares….. Not a climate change issue. Can sometimes be a local water issue (although that can be designed around). I suppose you use a bit more fuel, but nuclear fuel is extremely cheap compared to coal or gas. Thus, nuclear’s fuel costs are lower than other sources, despite the lower thermal efficiency.

  64. There is complete consensus among climate scientists that discharge of heat is a negligible contributor to global warming. The effect of greenhouse gases (CO2, methane, etc..) is many orders of magnitude higher. All we need to do is get off fossil fuels (although other things like agricultural practices play a role too).

    While there is definitely a safety difference between the Chernobyl design and modern light water reactors (LWRs), there is little more to be gained with advanced reactors vs. “old” LWRs, as LWRs have never caused any significant loss of life, over 50 years of operating hundreds of reactors.

  65. Solar generation does not coincide with peak demand at any locations. Peak demand occurs in the late afternoon/evening. There is a several hour delay between peak solar generation and peak demand (which occurs after solar generation has falling off significantly or even gone to zero).

    This is actually causing a real problem in the Southwest, where solar is significant, but still only ~10% of overall annual power generation. Arizona is actually being paid to take CA solar power during the midday hours, because there is more solar generation than they can use. A lot of solar in the region is actually shut off during those hours. It is hoped that several-hour electricity storage will soon be affordable. It’s only use will be to store solar electricity for a few hours, from the generation peak to the demand peak.

  66. The notion that reducing system pressure will reduce costs by a factor of 8 is absurd. Most of the cost of a nuclear power plant is for structures and components whose design is not affected by pressure.

    The main cost driver is excessive regulations, requirements and fab QA requirements. Basically, double standards where nuclear alone is held to a standard of absolute perfection. No new reactor design will change any of that (make those double standards of perfection disappear). Nuclear’s future will require a political battle, not a technical one.

  67. I haven’t heard of that, and storing a bunch of radioactive molten salt sounds like a terrible idea. You’d need hundreds of extra tons to turn a baseload reactor into a viable peaker replacement at anything like an economical rate. The NRC would soil themselves. So would all the non-proliferation people.

    There are lots of proposals for using MSRs for process heat, but that’s different.

  68. Several issues with various things you have here:

    1) For OECD countries, nobody’s going to put up with anything less than three nines grid reliability. If you can use a smart grid to shift some load around so that nobody cares and you can do with a bit less storage, cool. But the moment that somebody throws the switch and nothing happens, any pretense of global high-mindedness goes out the window. Same thing with doubling or trebling the price of electricity.

    2) For developing countries, they only care about cost and time to electrification. If an off-the-grid village can get modest electrification out of a solar micro-grid, they’ll weigh the time factor against the cost of going it alone. But if you tell them that they have to wait for more-expensive electrification because the rich people say that they’re now responsible for saving the planet, they’ll quite rightly tell them to do sexually acrobatic things with the proverbial rolling rubber donut.

    3) To get enough gas-powered gas turbine reserve capacity to make a difference, there’d have to be a guarantee of some kind of ROI on a system with a capacity factor that’s <10%. That guarantee goes right to onto customers’ bills. (See #1 and #2.)

    4) Per my #1, I don’t think that there’s any possibility of your #3 happening. Either there’s a technological fix that finesses away the storage requirement, or we won’t come close to de-carbonizing any time soon.

    Despite this, I’m optimistic. There’s a lot of tech coming.

  69. Given that we’ve been operating hundreds of PWRs for over 50 years, and they have yet to cause any deaths of measurable public health impacts, it’s hard to see how they are not safe enough. (Fukushima was an old BWR, but even that event will have no measurable public health impacts.)

    Storage is already a big issue, even though solar is still less than 2% of US demand. It’s more like 10% in the areas where the problems are being seen. One problem is that the peak in solar does NOT line up with the peak in demand, which occurs several hours later. We’re hearing about how 4-hour storage is starting to become somewhat affordable. But the thing that 4-hour storage may be used for is to merely store solar power for a few hours, from the daily generation peak to the daily demand peak. (nothing like storing solar power overnight, so that we don’t need traditional baseload power, etc..). That will be valuable, since it will then allow solar to act as a demand peak shaver, leaving a remaining steady baseload profile that can be covered by sources like nuclear.

    To illustrate the magnitude of the current problem. California actually pays neighboring states to take their solar power during the midday hours, because solar generates more power than they can use at those times, but the utilities have a mandate to produce all that solar. The other states are actually turning off their own solar to take the excess CA solar power.

  70. The numbers are based on official statistics and scientific consensus (the UN, WHO, etc..). Credible estimates for Chernobyl’s total eventual death toll range from ~60 to ~10,000. And Chernobyl is not even applicable to the merits of nuclear power, as practiced anywhere else. Aside from Chernobyl, there have been few if any deaths from nuclear power. Estimates of total eventual deaths from Fukushima, the only significant release of pollution in non-Soviet nuclear’s entire 50_ year history, range from zero to ~100. Meanwhile, fossil power generation causes ~1000 deaths PER DAY, and is a primary contributor to global warming.

    The use of nuclear power (in place of fossil fuels) has SAVED millions of lives, and would have saved millions more if not for political opposition. Global warming would have also been less severe.

  71. Don’t many MSR proposals use the molten salt to store energy much like the concentrated thermal solar plants? In the case of the MSRs it would enable the constant production of thermal heat of nuclear to match the daily variation in demand and could even go a long way to provide the variation needed to accommodate significant solar voltaic.

  72. Now the deciding question is: what happens when you run out of stored energy?

    This can go a number of ways:

    1. You’re running a smart grid so hospitals and elevators keep going while HVAC systems, factories, swimming pool heaters/filters, and eventually households shut down, in an order determined by risk, political power, and whether they paid for priority ratings in their contract.
    2. You have some backup gas turbines, because having gas that you only use for a few dozen hours per year is negligible in terms of environmentalism.
    3. You have not been able to have any backup plan at all because admitting such a possibility would result in political backlash such that anyone suggesting it would not be in power long enough to implement it. Hence trains will grind to a halt between stations, hospitals will black out, and aluminium smelting pots will freeze solid and be destroyed.

    I suggest that the choice of options depends on the local political culture. We all have our favourite targets that would go with number 3.

  73. You’re right. 50 is the wrong number. That overrates the real death toll which was actually much lower.

  74. The jet engine was an Air Force farce that nobody believed would ever work. The MSR that operated did so without incident until Nixon pushed breeder reactors to gin up jobs in CA. There are audio recordings of him saying this almost verbatim. I couldn’t parse your last sentence.

  75. I have solar and an electric car. There are weeks at a time when solar isn’t up to the task, and that is in late Spring, ie last week. I’m in Arizona. When the sky is clear I’m in fat city, cloudy not so much. My system is new so I don’t have records for a winter yet. But I can see that it would take an enormous battery backup to not be tired to the grid.

  76. Depends what you mean by “the vast majority of locations”

    Majority of locations in the USA? Maybe. I’m not quite familiar enough with the ratio of sunny hot states vs cold places there.
    Majority of locations in the currently industrialised world? All those cities in Northern America and Northern Europe and Northern China and Japan and Korea? Mostly cold where the big energy use is heating and light, especially in winter. I’ve spent enough time in northern Asia to be clear on this.
    Majority of locations in the world as a whole to take into account the ongoing industrialisation of South and South East Asia, Africa and South America? Ok, that should push things towards daylight use again.

    And remember we are in the takeoff stages of the world’s automotive fleet changing over to recharging at night, rather than running on fuel.

  77. That’s a lot of solar to fully charge a hundred kilowatt hour battery. I work during the day so my car just sits in a lot, not in my driveway plugged into a 15kW cluster of panels.

  78. Oops. Should be watt-hours per watt. But hey, what’s a little factor of 3600 between friends? (Fortunately, I’d been computing stuff as if it were Wh through the whole model.) Thanks for the catch.

    100 MW of capacity–>600 MWh of storage. In other words, you’ve got enough storage to deliver 6 hours at full capacity.

    This is a hard number to derive without going full Monte Carlo on about 10 years of weather data in each ISO/RTO. 6 hours seems about right as an average for most places that have both solar and wind spun up at some reasonable capacity, but it’s a guess.

    And the average is of course not what you want. The big question is what amount of storage you need to (ahem) weather 3- or 4-sigma bad luck. That’s a lot easier to do when 50% of your grid is reliable baseload capacity than if you’re trying to build a grid out of 100% tree-hugger-approved renewables.

  79. 6 joules per watt? As in, 6 seconds of storage for all of installed solar and wind rating?

  80. Nukes are no worse about waste heat than any other thermal power technology. Seen one Rankine cycle power plant, you’ve seen ’em all. The only difference is what you’re using to boil the water. The objective is always to make the the steam as hot as possible on the hot end and as cold as possible on the cold end.

  81. Something I’ve never been able to wrap my head around: If you want to completely de-carbonize, and you decided to go all-in on nukes, what do you use for peaking?

    Theoretically, you can use nukes for peaking, but the economics get ugly really fast if you don’t have the capacity factors well above 90%. Nukes work best as base load capacity. But solar and wind aren’t dispatchable, and hydro is intensely geography-limited.

    My guess is that you could make solar/wind work with a considerably more modest amount of storage than would be needed to run an all solar/wind grid, but that’s a guess that needs a lot of modeling to back it up.

    I played around with the LCOE for all solar/wind + storage, it it’s very sensitive to how many joules of storage you need per watt of solar/wind capacity. Below about 6 J/W, it’s pretty competitive with the $90-ish/MWh LCOE for nukes. Above 6, things go south pretty quickly. However, if all you’re using the solar/wind/storage for is peaking, the consequences of a really bad day (i.e. cloudy and calm) are mitigated a lot, and you can get the necessary grid reliability with a pretty small amount of storage.

  82. Is that a negative learning curve, or a positive freakout curve? My money’s on the latter.

  83. Depends on who’s editorial you read on “the history”. US Govn’t built two experiments – one was a wacky jet engine that shouldn’t have needed experimental evidence of its impracticality – the other was a mixed “success” that they are still considering entombing at a cost of $150M. The authors you read chose to emphasize the “success” where labcoats in the age of iron men and wooden reactors managed not to spill the MSRE pot.

  84. >solar is at 500 TWh and heading to 900 TWh globally in 2022.

    IEA says solar is at 720 TWh this year, and will reach 1130 TWh in 2022. Source:

    >Solar and Wind need natural gas or some other power plant backup or massive battery farms that have not been built yet.

    In the US at least, 30% of our power comes from nuclear, hydro, biomass, and misc other renewables besides solar and wind. Only 9% comes from solar and wind. So we have plenty of backup, and don’t need battery farms at large scale yet. They are being added in a few places where they make sense. By the time we need “massive battery farms”, their costs should be low enough to be acceptable.

    >Solar has industrial chemical waste.

    This is an unsupported lie. Solar panels are made of aluminum, glass, plastic, silicon, and copper. They are no worse than other uses of those metals. Note that most silicon in the world is used as an alloy in steel.

  85. 50 deaths at chernobyl???..if you assume your readers are that stupid..who’s the one being stupid…don’t even mention it will be contaminated for thousands of years…drink a glass of water from Fukushima…I dare you.

  86. I believe I’ve read that MSRs could theoretically be constructed for 1/8 the cost of PWR due to the difference in system operating pressure.

    The history of how PWR beat out MSR is quite interesting and unfortunate.

  87. If you can use U238 to breed fuel, much easier with MSRs, uranium already mined will last for millennia, and is effectively a renewable resource. Then there is thorium, which is at this point, an expensive to dispose of waste product of rare earth production.

  88. CANDU does allow the use of natural uranium. I’m pretty sure there is no other available reactor that does. Deuterium is likely the only moderator effective enough to allow natural uranium usage.

  89. I would stipulate that the CANDUs are the best pressurized water reactor designs available, because they use heavy water, and avoid the huge pressurized reactor vessel required for most PWRs. The superiority of deuterium as a moderator over any other candidate can not be exaggerated.
    Simple hydrogen actually slows neutrons better, since the nearly equal mass of the neutron, and proton maximizes energy transfer in collisions, but simple hydrogen absorbs too many neutrons.
    The CANDU can actually operate on naturally occurring uranium. There is no U235 enrichment required.
    The reactor of the CANDU is less expensive to build because it avoids a huge pressurevessel. From Wikipedia:
    “In CANDU the fuel bundles are instead contained in much smaller metal tubes about 10 cm diameter. The tubes are then contained in a larger vessel containing additional heavy water acting purely as a moderator. This vessel, known as a calandria, is not pressurized and remains at much lower temperatures, making it much easier to fabricate. In order to prevent the heat from the pressure tubes leaking into the surrounding moderator, each pressure tube is enclosed in a calandria tube. Carbon dioxide gas in the gap between the two tubes acts as an insulator. The moderator tank also acts as a large heat sink that provides an additional safety feature.”

  90. That’s likely an artifact of the absurdly high increase in the cost of employees imposed by the French government.

  91. Ah, but PV power coincides with peak demand in the vast majority of locations, unlike most of nuclear generation, which is around the clock. There’s lots of talk about the duck curve, but that’s only seen in places with high cost per kWh, and generally in the spring, when the sun is high, but air conditioning loads are not yet high.
    If you don’t consider the power demand at the time a kWh is delivered, you are comparing apples, to oranges. Then you have to consider the financial advantage of using your own power. At least for individuals in the US, you don’t pay federal or state income, social security, medicare, sales, or self employment tax on the dollars you save, as opposed to the ones you must earn to buy electricity.

  92. thermal heating of oceans?? seriously. If the entire world ran on nuclear power and all plants used wet cooling (ie discharge the 32c or so water temp into the ocean) versus, say, dry cooling or cooling towers, the volume of warm water discharge would be about equivalent to 150 cubic kilometers, annually. The U.S. nuclear industry discharges about 500,000 olympic swimming pools per year.

    The oceans have about 1.3 BILLION cubic km of volume. IE, 10 million times more water than the discharge. Literally a drop in the ocean. Compare this to the entire Arabian Gulf which is warmer than the nuke discharge water and 57 times more volume, and I don’t see that melting any ice.

    MSRs and other 4th gen nukes are far more coolant efficient and is even less than an issue.

  93. Surely deployment at sea introduces too many variables especially during a time of unpredictable weather patterns and a potential for more once in a century or feasibly worse storms? We couldn’t use the surface the units would need to be located on the sea bed at a suitable depth to avoid turbulence and tides?

  94. The differences between “old” uranium nuclear, and “new” thorium or at least passive design nuclear is so great that one really has to treat these as different technologies, which only strengthens the case for nuclear. Chernobyl, 3-mile Island, and Fukushima were all old nuclear, with Chernobyl being “old old.”
    One thing that is not addressed about nuclear, however, is thermal heating of the oceans and rivers. How much does this contribute to melting icecaps? Small nuclear input and huge ocean, I know, but still it’s continuous and if nuclear is built out too, does it matter and how much? Would newer designs minimize that?

  95. Yes it does. It mentions that the actual deaths per TWh is lower for nuclear than from either wind or solar. What else do you infer from the word “safer”?

  96. You know ThinkProgress is the work of political thug, George Soros, who axiomatically fights against uranium as an energy source. I don’t trust their figures. I do suspect that Perovskite solar cells will win the day, based on the work UK universities conducted during 2017-19 on greatly improving persovskites durability, while yielding 20% efficiencies. This will be the way forward, and not the Hedge Fund tyrant’s.

  97. The figure of 50 deaths in the Ukraine seems very dubious. The old USSR was expert at covering up anything that was not flattering. Deaths in 10s or 100s of thousands in all probability.

  98. If you build dozens of nuclear power plants with the same design then you can build them a lot cheaper. But PWR shouldn’t be what we such be building. LFR are safer, produce much less waste and can be cheaper.

    As for the cost difference between solar, wind and nuclear. Solar and wind will get a lot cheaper in the future. And storage is not an issue if solar is less than 50% of peak demand. Also peak energy is a lot more expensive than base load energy and peak energy is what solar is competing against.

  99. The title is misleading… The article doesn’t state why nuclear is “safer” than wind or solar.

    • The last 7 candu’s were built on time in 4 years and less and on budget at under $2700/KW in $2019 or 3 cents a kWh for public power like TVA
  100. The mass production of Small Modular Reactors (SMR) should dramatically reduce the cost of nuclear energy production especially if they’re deployed out to sea for the production of renewable synthetic fuels (hydrogen, methanol, gasoline, jet fuel, etc.) and industrial chemicals (ammonia, urea, chlorine, formaldehyde, etc.).

    Uranium extracted from seawater is also a renewable energy resource!


  101. From link “Under the best scenario, the cost of French nuclear power over the last four decades is 59 €/MWh (at 2010 prices) while in the worst case it is 83 €/MWh. “

    A TWh is 1 million MWh, so 1TWh is 59-83 million a year.

    Multiplied by 400 TWh. Multiplied by 40 years.

    944 billion euros to 1328 billion in 2010 euros

Comments are closed.