How to scale nuclear power
a16z.com
a16z.com
We have plenty of energy already
The entirety of our generation capacity is aimed at delivering energy during peak use, which is 8 hours per day. That is 1/3 of the day. The other 1/3 of the day is much less energy intensive, and the final 1/3 hardly registers in comparison. Coal plants can't be powered down so if they're generating too much then the excess is sent to ground. Literally, we spend 1/3 of the day pumping energy into the ground. Ditto for today's nuclear power plants. Yet people keep saying we don't have enough energy.
What we don't have is two things:
1. A distribution system capable of handling 24/7 load. The lines feeding your neighborhood were never meant to be run 24/7 and the equipment is undersized to cool down properly. You need to upgrade so they don't overheat and cause problems. Remember, they're only meant to be run hard for 8 hours per day and they're used to not being used much at night so they can cool down in the cool night air. More nuclear generation isn't goin to do squat about that problem.
2. Smart load management. Yes, industry can't shift load as much - but residences can. Don't charge your EV while you're using your oven. Things like that. Collectively there's plenty of power to power everything you need - we just need to be a little smarter in scheduling that power. You need to manage your peak rate. We need smart plugs that shut of when you've reached your peak usage rate. Shut off the EV, shut off the electric water heater - you should be able to program what you want done and get alerts that you're going over your peak load. Load management needs to be part of our lexicon.
Continuing to do what we've been doing is not the path forward.
Putting SMRs in coal plants to leverage the existing power gen and transmission is a very appealing proposition.
It seems that storage is a key part of the solution, and being able to distribute that more so that every neighborhood can have resiliency and the ability to charge fleets of EVs w/o taxing the grid.
Batteries win on convenience and are only going to get cheaper, but it would be nice to have other options. Flywheels are cool but Beacon Power's failure leaves that landscape looking bleak. Thermal storage seems like the best/cheapest option other than batteries.
I think we have enough technical solutions today -- now it's a matter of will and economic incentives.
I think the only "exception" will be air travel, but even that can be addressed by using green fuel.
https://medium.com/@twelve.CO2/electrify-the-fuel-how-e-jet-...
Good for you -- you'll pay $2.00 per kWh for the privilege. Mad about it? Turn off the computer, insulate your house better, or wait to charge the car. YOU have to pay for it because WE all collectively pay for it when they turn those peaker plants on and CO2 is dumped into the atmosphere.
We already have such things, but the effect is greatly dampened because the utilities are so limited in what they can do. The only way to get people to change their behavior is with incentives.
I use almost double the electricity of the average household right now, but I pay zero dollars towards fossil fuel energy and I pay extra for my energy to come from solar and wind. There are four adults under my roof. By all metrics, we're the "greenest" household in the neighborhood. If we switched to propane and gasoline, we'd have one of the largest carbon footprints instead.
The economics of what you're suggesting just don't make sense. Despite our high usage, our carbon footprint is far lower. As a society we want people to use more electricity. There's no such thing as clean energy when you get it from fossil fuels. The idea that we'd incentivize people to use dirtier energy instead of spending money to improve our electric infrastructure is wild.
That's what this is about, changing behavior. Two things are necessary for that:
(1) things need to be priced to align consumer behavior (subsidies for green energy and pigouvian taxes for dirty energy) (2) consumers need to have full knowledge of prices
Right now, we are actually not bad at (1) although in my ideal world the cost differential is far more extreme (the actual cost of energy is like 4c/kWh when the sun is shining and the wind is blowing -- as the gp comment says, we make a lot of energy!). We totally fail at (2). Most consumers don't know that energy is more or less expensive at different times of the day. They see a monthly bill.
Of course, you're right if you don't close other loopholes it won't work. A gallon of gas would be $10 and a Dodge Challenger would cost $100,000 or more if externalities were priced in.
That way 0% of your energy will be green!
At some point when the prices are too much, it becomes a tax on the poor, who need to use public electricity, to pay for the rich's private batteries. It would be better to take that incentive money and spend it on public utility batteries instead.
Make electricity cheaper at night. Demand will shift and it'll be a popular program rather than another eco tax that creates opposition.
Roughly 60% of US electricity comes from coal and methane. A serious response to climate change means shutting down all that fossil generation, while simultaneously electrifying all the cars, heaters, and industry.
We don't have nearly enough energy yet.
If one were to regulate hydraulic fracking and/or regulate the emissions produced by these plants then there wouldn't have been the huge build-out of them. Utilities do what they're incentivized to do.
Unregulated CO2 emissions are a huge problem for the world, and natural gas is part of that problem, regardless of what the law says.
Oversimplifying: if we're running at an average of (let's say) about half load (full load at peak, and 1/3 load at all other times, for about 4/9 ~= 45% average 24h load), wouldn't it be useful to run generating stations at or around that figure around the clock and store the off-peak surplus for on-peak use?
And as additional food for thought... How many outages can anyone remember that were caused because of lack of generation capacity?
At the time Texas was dealing with frozen power plants there were plenty other colder parts of the country that had their generation, including windmills, running just fine. Let's not pretend the lack of generation capacity caused the problem, because it didn't.
in their dreams...
PV costs are falling exponentially (-10% per year) while nuclear costs remain at best constant. If you decide today to build a reactor it will be ready in 10-20 years (don't be in the hurry). By the time it start producing it will be competing with plenty of cheaper renewables. During its lifetime (50 years) it will have plenty of time to become obsolete and so expensive.
It's like buying a Cray I in the 1980s and get delivered in 2000 just to find that a gaming console has more power than your mammoth computer but at a fraction of the cost.
Nuclear is completely doomed.
When you dig into why it takes 10-20 years to get one built, you arrive, more or less, at the feet of incompetent regulators, that order changes during construction, after plans are initially approved, and then they can/do order more changes before they'll allow the plant to operate.
Now, don't take that as "loose nuclear regulation good", that's not really a solution, what we need is much, much better regulation, so a plant can get approved and then built without interference provided the plan that was approved is followed.
And, realistically, once you have that, you can come up with a plant design that you can copy/paste in a bunch of locations, instead of building them as one-off's, and that not only reduces costs significantly, but also improves safety significantly because if one plant has an issue, you can also copy/paste the solution/mitigation/prevention elsewhere.
Similarly, the "blame the regulators" game does not work for recent reactors in Finland and France based on the latest design.
What it really comes down to is incompetence on the design side and on the production side. As well as not taking the requirements seriously, at least with some of the contractors.
The design change story was more true back in the early 80s, but even then it was only part of the story. Good projects finished quickly, in any case, back then. A bigger problem was vastly over ordering nuclear too, in a planning failure.
Fundamentally, nuclear construction has to deal with the fact that in advanced economies, labor only gets more expensive and more valuable. And the foundations of big nuclear reactors are massive scale and massive labor coordination. There's so much high skill welding, that needs to be checked so thoroughly, since it needs to last so long.
There are some, like Jigar Shah, who think a nuclear industry can be built if leaders emerge who reject the blame game and take accountability onto themselves for actually making a product that utilities want to buy, and doing it on budget and on time. I'm skeptical, after having Westinghouse and the EPCs dash my hopes, and seeing the same failures in Europe at the same time. But I hope I am wrong. However I don't see a single thing in this a16z article to make me think that they will be able to pick and fund those leaders. Let the VCs be the dumb money to find a model that works, then Shah will bring the capital that can really make it scale, IMHO.
In a database of thousands of projects over $1B in size, nuclear power projects come out third worst for cost overruns, behind (before?) Olympic Games hosting and Nuclear Waste Storage Facilities. The kicker: that's construction, after approvals.
It's not even close, either.
1. https://duckduckgo.com/?q=Bent+Flyvbjerg+How+Big+Things+Get+...
Heh. You made me realize that nuclear revival is the latest mania (catnip) for gold bugs. Just like cryptocurrency, err, I mean "digital assets".
How do you maintain a nation's baseline? Coal and gas plants or massive battery storage, which raises many questions about environmental impact and resource requirements, hydro is not viable everywhere, is that the future you're hoping for?
The 'baseline' idea exists because of nuclear's high fixed costs and low marginal costs. This is also a form of generation which is complemented by dispatchable generation and storage, just the same as renewables, and does not complement renewables usefully at all. So if you wanted to build a grid based on nuclear you'd need to either overbuild it or build lots of storage, just the same as with renewables. And renewables absolutely destroy nuclear on cost.
As we electrify our cars and trucks, every house is going to end up with 2-7 days of household electricity requirements parked in their garage. Most models show that that alone should be enough to smooth over the daily variability. Using stationary, cheaper battery chemistries opens up an entirely new world of resources too.
Battery + solar is already beating out new nuclear on cost, and we have decades of novel development of the techs to bring down costs even more.
Unless nuclear can figure out a way to start dropping in cost by 10% every year, and having shorter capital lifetimes so you're not locked into the high initial cost of today for 60 years, it really will be doomed.
Once batteries are everywhere who cares when the electricity is produced.
Hey btw Vermont Utility plans to installs battery storage at customers' homes because it cheaper than updating the grid
https://www.nytimes.com/2023/10/09/business/energy-environme...
Everywhere I've looked, sufficiently inexpensive storage is either theoretical or using some tech nobody has really built out en masse before.
I know there's a few desert installs that do well, but the best places for solar tend to also be the least populated.
And yet there are plenty of grid-scale nuclear power plants, and very, very few grid-scale solar plants.
Weird, huh?
> If you decide today to build a reactor it will be ready in 10-20 years (don't be in the hurry)
Whereas one can simply wave a magic wand and a grid-scale solar plant will magically spring into existence.
"Renewables" aren't suitable for base load, and the figures for "renewables" are generally presented in a deceptive way. 1 watt of solar does NOT replace 1 watt of nuclear, because there is this thing called "night". You need 2 watts, at a minimum, even with storage. More than 2 if the weather is at all cloudy.
Mandatory nick-picking from a chauvinist French: France is “only” around 70+% because we do use nuclear for much more than just base production, and modulate the production to fit the demand variation, not because Americans operators are somehow better ;).
Also, unsurprisingly from A16Z they put most of the blame on regulation, but the main reason why nuclear is costly today is because there's barely an industry around it and there's no mass production. And unsurprisingly, custom built one-shot stuff cost much more than serial production.
Isn't that because of regulation (which was put in place in part because the population demanded it)?
And not to forget: some things are freakishly dangerous. For example, spent fuel rods provide deadly amounts of radiation in basically no time to people around them.
You know what's deadly in a nuclear plant? The industrial machinery, the heavy plates of metal, the rotating devices at high speed, the trucks crossing the plant to ship heavy stuff, the dears crossing in front of the cars of the night shift workers.
All of this kill and maim workers from time to time, like in every factory in the world. But you don't care about them. You just care about fantasies of fluorescent green stuff killing people with deadly invisible radiation, even though it barely ever happened in the real world (and it's not fluorescent green either).
And, yes, it tends to glow blue, not green.
But you are not seriously alleging that radiation from spent fuel rods isn't deadly, do you?
No, it doesn't glow at all. Water under irradiation do glow (Cherenkov effect), but the rods themselves don't glow.
> But you are not seriously alleging that radiation from spent fuel rods isn't deadly, do you?
They aren't. They would be very deadly if they were out of water with people around, but it doesn't happen under any circumstances, so the risk you're talking about doesn't exist. It is as nonsensical as saying wind turbines are deadly because “if you got hit by a rotating wing you'd be cut in half”.
Even when transported after storage in water they do emit copious quantities of radiation, hence the rather cumbersome transport "cases". You might want to familiarize yourself with the dangers during the lifeycle.
Btw., people die from wind turbine blades in accidents (when they break, for example).
Perhaps you might want to work in a reactor for a bit to get more of a feel for the dangers and the many things done to contain radiation related risks (instead of just shouting how it is all a non-issue).
True, but it's a thousand time less luminous, and actually it's still not the rod that glows.
> You might want to familiarize yourself with the dangers during the lifeycle.
In fact, I studied nuclear engineering before deciding to do computer science. There I met a bunch of good friend who works a several place of the nuclear supply chain, and also my wife who's now working at the plant 8km away from were we now live. So I'm very, very familiar with nuclear stuff.
> Perhaps you might want to work in a reactor for a bit to get more of a feel for the danger
Fortunately, nobody works “in a reactor”, as it would be indeed by dangerous (300°C and 155bar isn't compatible with human life, without even talking about the radiation levels ;).
But I take your point that you never worked in the field.
Granted, the pressure and temperature are technology dependent, but it's never human suitable conditions in any case. And that's no sophism, the problem with your reasoning is that you call something “deadly” where it is in fact never in contact of a human in the first place, and could never be.
> It is also not the water that glows as such
What do you think glows? It's literally water molecules that glows from de-exitation.
> take your point that you never worked in the field.
Sure, but neither did you, and your vision of nuclear is much less based on reality than it should to have a discussion on the topic.
And, unlike you, I have worked there.
You might want to look up how Cherenkov radiation works, btw. - it's depolarization.
Anyway, it is good that you are kept away from nuclear technology given your lax and aggressive attitude towards risks there (anyone thinking accidentsare impossible has no place there). Good choice.
Yes, from afar, at CEA but I don't think anyone was working inside it.
> You might want to look up how Cherenkov radiation works, btw. - it's depolarization.
Would you mind expanding? Because Wikipedia seem to agrees with what I remembered:
The charged particle excites the molecules in the polarizable medium and on returning to their ground state, the molecules re-emit the energy given to them to achieve excitation as photons.
> Anyway, it is good that you are kept away from nuclear technology given your lax and aggressive attitude towards risks there (anyone thinking accidentsare impossible has no place there). Good choice.
Well, I see you're angry, but that's kind of a funny thing to say to someone who's started this very thread by defending safety regulation in the first place[1].
FYI There's a big difference between having a “lax attitude towards risks ” or “thinking accidents are impossible” and noting that you're completely making up risks that simply don't exist like “spent fuel rods provide deadly amounts of radiation in basically no time to people around them”. If you have people “around” spent fuel rods in the open air, you have a serious organizational problem: what are you trying to do in the first place?! (researchers in many fields are known to be completely oblivious to any kind safety measures, so maybe that's the kind of thing you did in your lab, but in an industrial setting I doubt anyone would have be doing that after at least the 70s).
The special risks with nuclear have nothing to do with the deadliness of things like hot fuel rods, and much more about broad environmental impact with moderate increase in cancer rate spread over a big populations (be them workers in contaminated areas or general population in the wide place around an accident). This is a big topic. But when it comes to immediate death risks on a nuclear plant, the radiations coming from radionuclide simply isn't one of them.
[1]: see also this comment in a sibling thread for another even more explicit example https://news.ycombinator.com/item?id=37935638
In many other countries in Europe it was because of the strong public backlash against nuclear after Chernobyl , and because of the oil glut in the 80s which made fossil fuel affordable again for decades.
Capacity factor of 100% is not feasible because you need maintenance of the plants (probable true for every single form of electricity production).
Regarding whether or not stricly 100% of nuclear in the grid is feasible, I don't now enough about grids to say whether only using load-following nuclear power plants (without hydro/gas turbines) could do the job. We have hydro in France, might as well use it rather than have only NPPs, and dams are great for load-following.
It is feasible, but the problem would be the peak in winter: that would mean having more than twice as many nuclear reactors to be able to deal with a few winter nights with about 100GW of consumption. That'd be very wasteful.
> We have hydro in France, might as well use it rather than have only NPPs, and dams are great for load-following.
Depends on what kind of hydro though: mountain barrages are, but not really Run-of-the-river hydro.
Fabulous how you get that from my comment. Talk about confirmation bias.
It's literally the opposite, as illustrated by France doing most of its daily load variations from nuclear.
Said another way: traditional nuclear power generation cannot jump onto the cost learning curve (Wright's Law).
Time will tell if small modular reactors can or will make that leap. Decades from now.
H2 is closer. (Because we've pointed a fire hose of money towards H2 and not nuclear.) So the window for a nuclear revival, in any form, is small and getting smaller.
At least with smaller reactors it's more feasible to get operating data and be on a path to economies of scale. The US now has a second chance to get it right, and the US can always be counted on to do the right thing, after exhausting every other alternative.
[1]: My wife work in nuclear engineering, and I did work for the rail industry for some time. It's not exactly the same, but it's also not that different compared to other regulated industries, like healthcare for instance, where it's full YOLO compare to both of the above.
"...a nuclear power plant can generate over 6,507 kW/acre compared to solar at 23 kW/acre, or a natural gas plant at 114 kW/acre."
So in effect it has a negative land requirement in those places.
1: incremental construction costs: (NB, numbers speculative, think order of magnitude). Roughly speaking, 100e6 dollars gets you a decent solar/wind utility scale installation (heck, you can do it retail for a few 10e3's on SFH rooftops), but only (maybe) a feasibility study for nuclear. Definitely not an operating plant. Spending 100e6 dollars a year gets operating solar/wind generating capacity on an ongoing basis, same on nuclear gets you a working plant in how many years?
2: backend costs: I don't think nuclear has a business plan about the cost of decommissioning a plant. Or maybe it ends up like orphan wells: https://en.wikipedia.org/wiki/Orphan_wells_in_Alberta,_Canad....
Reason 1: means that the business risk associated with solar/wind is far less than that of nuclear. Given that nuclear and solar/wind are very similar as being base load [1] and not well suited to tracking fluctuating demand (so, both need the same sort of backup / storage tricks, albeit for different production profiles), solar/wind win the "cheap 'n' cheerful" race hands down. Also, manageable by private entities, so no WPPSS scandals ("whoops!" https://en.wikipedia.org/wiki/WNP-3_and_WNP-5).
Reason 2: decommissioning/recycling costs are far lower. This reduces taxpayer risk (ie., externalizing costs). There's nothing exotic: aluminum frames can join the "beer can to beer can in six weeks" loop and the other materials (can't be bothered to find the references just now) aren't as hard to reprocess as detractors claim. Also, again, it's smaller increments, so no massive 6 reactor plant that will take twenty years to dismantle. It's far easier to develop a 'panels to panels' industry with a steady stream of little projects than with a "my career was about remediating Splitters Bay" model.
[1] nuclear has to 'throw it away' by paying someone to take the power, they can't ramp down. Solar/wind has to 'throw it away' by not taking what's on offer. No way to make back six hours of curtailed sunshine.
About powerplants in France shutting down because of cooling during the summer... I find it hard to believe this is a widespread issue, if an issue at all. Many of the existing nuclear reactors are reaching an age where massive maintenance schedules need to be executed. If the problem is extending the thermal range, that's relatively easy to fix.
Source: I used to build powerplants.
https://www.reuters.com/business/energy/high-river-temperatu...
This is why fusion projects are looking for ways to directly generate electricity. If they run steam turbines they've lost before they even start.
> the success of nuclear power is much more about project management, financing, and policy than it is cutting-edge engineering or safety.
But suppose you optimize the hell out of that, then what do you get? Maybe, just maybe back to 1960s costs, at the very best.
That's not good enough. We need more advanced tech that delivers energy more cheaply. Defining "efficiency" as MWh/acre is just plain silly, and greatly reduces the credibility of the article. What a useless metric, unrelated to economics, that shows a gaping hole in the analysis. Its just boosterism because except for the useless metric, there's nothing appealing about nuclear as a power generation tech, in its current forms.
Given a bet between Web3 and nuclear as foundational future techs, would pick Web3 in a heart beat, and let me tell you just how little I believed that silly pitch when a16z was making it... much less now!
[citation needed]
I know it's almost a meme now, but not really true. Basically anything renewable doesn't work like that, including hydropower, which makes up a large portion in certain regions.
https://www.eia.gov/energyexplained/electricity/electricity-...
Hydropower is basically the same thing, the steam has just condensed already.
> Hydropower is basically the same thing, the steam has just condensed already.
I hope you're kidding.
86% is nearly all to be fair(edit as of 2017)... not 99% nearly all but it's well more than half.
Is 80% "nearly all"? Dunno, that's semantic. But it's a pretty high percentage.
Also, interesting to see them go all in on a big future with retro-tech.
The initial research was done in rather cold seawater, but they did demonstrate that the process was faster with warm water. Last I heard, they were working on testing the viability of post-filtering hot brine from desalination plants, which should be much more efficent.
Steam generation for electricity is mostly a dead end, IMHO. Being attached to it is like saying that typewriters have been the backbone of business communication for a century, so we had better invest big in them for the future.
In 1950, if you were a Great Power and wanted to remain one, you had to have nuclear technology tout de suite. Electricity generation was a sop to pacify the masses. There were all sorts of overt and covert subsidies.
Nuclear power only continues to exist for those military applications.
Back in 2008 or so China had an ambitious policy to grow its nuclear fleet about twentyfold in the next two five-year plans. Last I looked, they had only nearly doubled their fleet and the rate of growth is utterly anemic compared to growth of PV and wind.
If China can't do it, with its ability to ride roughshod over everybody and anybody, why would things be better here?
or does it involve people investing in currently non-existent "modular" reactors?
The time-scale is so long that there are no geological formations that last that long
For the entire period the waste must be monitored to ensure that it does not come into contact with humans, whom it will make very sick.
The waste does not look, taste or smell poisonous so some sort of warning sign has to be placed on it.
That sign must be readable over a time scale many times longer than we have had writing.
What selfishness. Making the next thousand generations pay for our current consumption.
The most expensive energy possible to imagine.
But because it is not our generation, but the next thousand generations (plus) that pay the bill....
Golly.
Geologist here, highly skeptical of this statement.
Most of the land area of each continent has had essentially no seismic/tectonic activity for tens to hundreds of millions of years. With sufficient burial depth (which doesn't have to be super deep, less than a kilometer certainly) there is little reason to suspect that in most areas that haven't been disturbed in 200 million years, they will be disturbed in another 200 million years. The surface of the earth has had cycles of erosion and deposition associated with climate variations (i.e. ice ages) but the effects of these aren't always that deep.
This isn't to say that the other issues aren't important--communication of dangers to future humans (or other intelligent creatures) that may not share our language, especially. However it's a social/political problem, not a geological one.
I would rather burden future generations with this rather than continuing to produce CO2 at the scale we are doing now.
https://en.wikipedia.org/wiki/Onkalo_spent_nuclear_fuel_repo...
It feels more solvable than trying to figure out how to reduce CO2 without nuclear (to me)
Thank you for making my point
We can stop producing co2.
There is a price we have to pay
But our greed means "I would rather burden future generations"
We are all human and have flaws and it's best to work around them rather than idealize about a world that doesn't exist.
Imagine, the lead in our pipes (for those of us that live in Roman settlements or small American towns) is sooo radioactive Lead-205 has a half-life of around 1.73×10^7 years.
That's nearly 20 million years!!
Wait, is that a good thing or a bad thing, having a long half life?
Is that related to the toxicity of lead, or is that something else?
Are the really toxic isotopes long lived or short lived?
Still, at least we can all agree that just having a nuclear fission reactor out in the open is absolutely going to poison the planet forever.
Maybe: https://en.wikipedia.org/wiki/Natural_nuclear_fission_reacto...
By definition.