Why is nuclear fusion so hard? | Hacker News Reader

Why is nuclear fusion so hard?

blog.sigfpe.com

148 pointspanic170 comments

This is somewhat tangential to the specific issue here, but I was surprised by this:

At the center of the Sun, fusion power is estimated by models to be about 276.5 watts/m3. Despite its intense temperature, the peak power generating density of the core overall is similar to an active compost heap, and is lower than the power density produced by the metabolism of an adult human.[1]

So practical fusion power requires producing much higher power densities than are found in the Sun.

[1] https://en.wikipedia.org/wiki/Solar_core#Energy_conversion

> Despite its intense temperature, the peak power generating density of the core overall is similar to an active compost heap...

I've seen this comparison before and it also seemed quite surprising to me. But in hindsight maybe it shouldn't: The sun is already billions of years old and it hasn't burned out yet.

Another way to think of it: A larger compost pile has to radiate more heat out of the same surface area as a smaller heap because of the Square–cube law [1]. Now imagine a compost heap more than 800,000 miles in diameter.

[1] https://en.wikipedia.org/wiki/Square%E2%80%93cube_law

That also shows why fusion power always centers around exotic isotopes like deuterium and tritium rather than plain hydrogen. The reaction paths needed to fuse plain hydrogen are so improbable that they don’t happen fast enough to be useful.

Is this an another way of saying that the only way to fuse to Hydrogen is to create a structure as large and with gravitational features that of a star?

Then a star/solar energy is the only way you will ever have large scale fusion energy generation and harvesting.

I suppose the universe doesn't owe us a cheap, easy, unlimited source of energy. A lot of people seem to think that it's inevitable we will achieve amazing, magical feats of technology given sufficient time and effort that are unthinkable today. Maybe. But it's also just as possible that we already know most of the general parameters of what is possible in the universe, and the peak practical feats of technology that will ever be open to us are not all that far off from what we have now. How can we tell?

Nuclear tech is not even a century old. The idea that we can't do any better in a thousand or a million years seems so improbable to me that I can't take it seriously. There is some kind of extrange pessimism around that mindset: that we will be extinct much sooner, or that the Sun will die and thus Humanity, as if there's no obvious way to escape that apocalypsis.

I think that significant swathes of human kind throughout history have assumed that their current state of existence is the peak and final form of humanity. While it is always logically possible that this is the case, it has never yet been true.

I don't know what it is about the idea of the far future that makes people so willing to throw out fundamental physical laws. The relevant point here would be that there are fundamental limits to heat transfer, and a finite amount of work which can be done on Earth-as-we-know-it for a given unit of time, irrespective of the energy source. An "unlimited" source of energy would be in practice limited by its heat output.

Improbable is not the same as impossible. We didn't even know about atoms for a huge stretch of science, and now we're smashing them together to break them apart looking for smaller things.

Is it completely out of the question that some discovery will be made that invalidates our current understanding of physics?

Even without new physics, processes can be improved with new tech. As far as I know current computers are based in the same foundations as in the past century, but they're many orders of magnitude more powerful. Then, after some threshold, quantitative changes become qualitative ones.

Physical processes cannot be made arbitrarily more efficient. There are thermodynamic limits to computation, the same as any other form of work.

You are beating a strawman with great dedication, congrats.

Yes, it's absurd to consider that there will be any large-scale violation of thermodynamics in the indefinitely far future. Literally, nothing is less probable.

Literally nothing is less probable? Alright then. I don't know why we bother with science any more since you've apparently got this one in the bag.

Non sequitur, and rude.

> do any better in a thousand or a million years

From what I can tell, there is a much bigger chance that we'll do much worse - and much sooner.

And I was afraid someone would call me Dr. Freud for mentioning the pessimism...

Agreed, the accepted models in physics are laughably incomplete. There is still a tremendous amount we still don't know. Although it's possible we've peaked, I have doubts.

It is also a self fulfilling prophecy to declare that we have reached our technological apex, and not striving to progress further.

No, it's saying that large-scale fusion using hydrogen fuel requires a star, because hydrogen fusion reactions are very improbable. That reaction only happens when four nuclei collide at once.

Fortunately we have other fusion fuels that are much more likely to fuse, including deuterium, which makes up 1/2500 of the hydrogen atoms in the ocean. Most projects target the easiest fuel, deuterium-tritium; while tritium isn't naturally available on Earth, it can be bred from lithium.

Nope! The p-p chain only has pairwise interactions. The issue is the p+p->d step is very disfavored, thanks to the weak step. D+T is all strong interaction.

Could that be explained with less jargon? I’d like to understand it.

There is a chain of individual steps that need to happen to create the final product of hydrogen fusion, helium. One of the crucial steps is the production of deuterium from two protons. This step is incredibly slow and limits the rest of the chain. However if you skip the slow step and start by fusing deuterium and tritium you can have a much faster and efficient reactor.

Indeed. After a D nucleus forms in the sun, it reacts (with a proton, to form 3He and a photon) within a few seconds.

The 3He+3He reaction is also slow, but not as slow as that first step.

Thanks, that was interesting. Apparently the five-minute video I got my statement from led me astray.

Ah, one of my favorite astronomy-related trivia.

It also shows how a supernova can (briefly) shine as brightly as its whole host galaxy (i.e. some 10-11 orders of magnitude stronger than an average dwarf star - like you wrote, those basically "smoulder").

It continues like this though:

> The low power outputs occurring inside the fusion core of the Sun may also be surprising, considering the large power which might be predicted by a simple application of the Stefan–Boltzmann law for temperatures of 10 to 15 million kelvins. However, layers of the Sun are radiating to outer layers only slightly lower in temperature, and it is this difference in radiation powers between layers which determines net power generation and transfer in the solar core.

So if a m^3 of sun core was moved to a powerplant it would generate a lot more energy.

Not exactly. If, using your Star Trek-style transporter, you instantaneously transported a cubic meter of sun core into empty space, it would radiate a lot of power at first, but by the very fact of radiating all that power, it would quickly cool, especially if you didn't have a good way of keeping it from expanding. But even if you did have some way to contain it, it would still cool pretty fast, since it would no longer be receiving the power input of all the other core stuff around it.

> So if a m^3 of sun core was moved to a powerplant it would generate a lot more energy.

Yes, for a tiny fraction of a second as it exploded. Its pressure would far beyond what any material could withstand (and magnetic confinement does that change that, as the outward pressure has to eventually be carried by the magnet support structure).

Just to expand on that: the kinetic energy of the particles in 1 m^3 of the material at the core of the Sun is equivalent to the yield of a 5 megaton bomb. So when I say "exploded" I really mean it.

> So if a m^3 of sun core was moved to a powerplant it would generate a lot more energy.

Would it? Or would it expand into basically normal hydrogen and helium?

Provided you could keep the plasma contained in magnetic fields, it would certainly generate plenty of free neutrons to provide heat for substantial power generation.

If you let it cool, it would stop fusing. The fusion power wouldn't magically go up.

fusion halts harmlessly when conditions are unmet

Halting harmlessly is the opposite of "producing a lot more energy".

It would produce negligible quantities of neutrons.

My understanding is that neutrons are the primary component for heat generation and therefore steam and therefore energy. At low energy fusion reactions, is this not the desired outcome?

You seem to think that the outcome of an experiment is determined by desirability, not by physics. This is called wishful thinking.

Uh my counterexample is the Sun itself. Uses fusion on a daily basis to provide vast quantities of light and heat. What a fusion reactor does is it decouples the plasma superhot gravity-laden center of a sun with a magnetic field strong enough to hold tiny pockets of plasma to generate energy (once we can go straight to fusion reactions using Boron) or to generate other particles which will create heat.

QUOTE

"I took this fusion class when I was at Georgia Tech and I will never forget it. We started studying and I go, "Man, this is really hard." Charged particles don't want to get near each other. Bare nuclei are both charged, positive charged, they want to avoid each other.

And my professor had a really great way of putting it. "It's like going to the mini golf." He says, "You know how in mini golf you've got the volcano, and the volcano's got the hole at the very top, and you've got to putt your ball in a way that it goes all the way up the side of the volcano, and 'phwep!' falls in the hole." He goes, "OK. That's like fusion. The ball is like a nucleus, and the volcano is the scattering effect.

So any time you want to have a nucleus go to another nucleus, it scatters; it rolls up the mountain and it rolls down the side, it rolls over here, over there... and only when you just perfectly get it on the right angle does it go in the volcano." Now, the problem with fusion, he goes, "You can't steer the ball, you have to have enough temperature so that it can make it all the way up the side of the volcano and fall in, and then you have to have enough balls because you can't steer them there at the mini golf park", that's density, "and then because they're flying all over the place, you've got to make sure that there's a fence around the mini golf park so that they don't get away." That's confinement.

He said, "Those are your three things: density, temperature, and confinement, to make fusion happen."

I said, "Dude, that's really hard!" So, I came up with another analogy, "So, I guess fission would be like the mini golf park except now the volcano was flush, the hole was about this big around, the balls are going slow, and every time the ball goes in the hole, two more balls come out."

He goes, "Yeah, that's pretty good."

ENDQUOTE

Watch the whole thing. It's good.

https://www.youtube.com/watch?v=lG1YjDdI_c8

Re: ..."Man, this is really hard." Charged particles don't want to get near each other. Bare nuclei are both charged, positive charged, they want to avoid each other.

Are you sure you didn't accidentally quote my college dating experiences?

What point in the video is this quote?

https://www.youtube.com/watch?v=lG1YjDdI_c8&t=71m26s

> Some kinds of instability are slow enough that we can control them. For example bicycles are unstable, but many of us eventually learn to ride them.

Actually, bicycles are only unstable when they are moving slowly (or stopped), and most people never learn how to stabilize one in this unstable regime. It can be done, but it's very, very hard.

Now, what would be great is if that analogy could be stretched further. It would pretty wonderful if someone found a way to do fusion where, like a bike, it's stable once it gets going.

Well, that is actually possible. That's what happens in stars and hydrogen bombs. The problem, of course, is that "getting going" for fusion means getting very big. The reason fusion energy is hard is precisely because to be practical and safe we have to "ride slowly".

These are called self-stable plasmas, and are already been explored. For example, TAE's field-reversed configuration.

If you understand German, trust me and listen to this podcast about Wendelstein 7-X O:) : https://alternativlos.org/36/

This is a really excellent podcast. Very informative.

Almost trotted this one out myself. So good!

Here is a thought experiment( no cats harmed or wetted). Show a cat how the toilet flushes, then try to flush the cat. You have two hands, the cat has 4 paws, with sharp claws as well as teeth. This will give you an idea....

Stars have gravity and mass- we do not, so we must try and compress the plasma, heat it AND stops the paws from grabbing the exterior. In truth magnetic confinement is a poor way to constrain a plasma that wants to leave at high speed = quenches the reaction. There have been many avenues tried and they are making incremental advances, year by year. They say we are 5-10 yearsaway? The ITER Tokamak is well advanced. https://www.iter.org/mach/Tokamak New approaches are being tried that might leap past the Tokamak.

Controlled fusion has always been 5 to 10 years away, just as we've always been at war with Eurasia.

I remember distinctly reading a popular science magazine article when I was young (it had that great plasma picture with all the plasma arcs fizzing over the floor of a giant lab) that stated that we could have fusion power plants in 25 to 50 years. That was 30 years ago.

It seems I was indeed wrong about the deadline but not about the "forever" part, see [1] and [2].

[1] https://www.newstatesman.com/sci-tech/2014/11/forever-20-yea...

[2] http://blogs.discovermagazine.com/crux/2016/03/23/nuclear-fu...

That is partly true,but we ave advanced to the point where the next generation of machines take 5 years or more to build. They have reached the break even and gone substantially past it, but they have not been ablt to reduce it to the turn-key appliance stage.

No fusion reactor has reached breakeven.

All deadlines have an implied "with equivalently aggressive funding" addendum

Try making a compressed, spherical ball of water using your two hands.

“Confining a plasma is like trying to keep water in an inverted cup by blowing on the water with fans.”

That is fantastic analogizing right there.

Plasma instability is just awful. See [1]. I thought he was going to go on about some approach to active stabilization of plasmas. But no. There's been work done on that, but it's been years. Anyone following that?

Then there's the first wall problem.[2] Fusion generates heat and neutrons. Lots of neutrons, which break atoms apart. Finding something which will stand up to that in an experimental machine has been tough. Finding something which will stand up to that in a long term production environment is really tough. In a fission reactor, you can use water to slow and stop the neutrons, and you just get some tritium as a byproduct. A fusion reactor's first wall faces a vacuum, so that's out.

[1] https://en.wikipedia.org/wiki/Plasma_stability

[2] https://en.wikipedia.org/wiki/Plasma-facing_material

Yep, studied Nuclear Engineering at NC State, didn't stick around for the grad program. My friend did, for plasma physics, and is now working at ITER. Plasma confinement is still improving.

This a totally ignorant question - can you add “something” to the plasma to achieve the “jello like” analogy to water?

That's actually a very reasonable question. The answer is: no, but explaining why is non-trivial. The short version is that at 10 million degrees, matter behaves so differently than it does at room temperature that none of your everyday intuitions apply.

Warm Dense Matter can be thought of as ionized jello.

Would zero gravity be helpful in order to make it easier to get fusion working?

The particles in a fusion reactor are moving at 10 km/s. They don't particularly care about gravity.

For components of the reactor, gravity tends to be helpful. It's much harder to cool superconductors or generate steam if your liquids aren't flowing along the bottom. Turbines don't work nearly as well if their inlets are sputtering like a coffeemaker.

They are moving a lot faster than 10 km/s.

Good question. Under the influence of a magnetic field, charged particles do all sorts of counterintuitive things due to the Lorentz force, which acts at right angles to the particle velocity and the field.

One example is a sideways precession due to the interaction between gravitational acceleration g and the Lorentz force. Fortunately this effect is basically negligible, because gravity is so much weaker than the electromagnetic forces operating on the particles.

The difficulties tend to come from the geometry of the magnetic field itself. For example, in a torus shape, there is an unavoidable drift in the vertical direction due to the combination of centrifugal force in the big circumference, and the force implied by the gradient of the magnetic field.

My understanding is that stellarators are designed in such a way that this sideways drift in one part of the device is canceled by opposite drift in another part due to the twisting of the field.

Space also comes with a natural vacuum, not just zero-g. That might be more interesting for a reactor since you would only have to care about shielding the magnetic coils instead of building and cooling a vacuum vessel at the same time.

Most fusion is powered by gravity, so at this point, low gravity has been nothing but an impediment.

This is why I have little confidence that Tokamaks and similar will have much success.

Instead of trying to prevent the non-linear behaviour of plasma it should be utilized.

Look at (grossly underfinanced) attempts like focus fusion, where there is no attempt to confine the plasma, instead self-interacting nature of the plasma is used to focus high temperature regions to a small region where higher temperature fusion processes can occur.

Long time plasma confinement is a dead end.

I'm a fan of IEC, what happened to the Polywell?

https://en.m.wikipedia.org/wiki/Inertial_electrostatic_confi...

This is probably another ignorant question, but why can't we "aim" and shoot protons at each other to make them smack together without requiring all that heat and pressure?

Say you shoot the deuterium and tritium ions at each other in beams. Making two ions fuse isn't quite the same as hitting a fixed bullseye. Typically, physicists consider a 'cross section' for a reaction between two particles as a function of their relative velocity: a notional area that allows one to compute the probability of a reaction occurring, which may be thousands of times larger or smaller than what we typically think of as the physical size of the particles.

The cross sections for fusion reactions are significantly less than 1000th those of elastically bouncing off, even at the optimal velocities. Fusion reactions release about 1000 times[2] the typical energy needed by the reactants in order to get close enough for there to be any significant probability of reaction. So a one-pass beam system is always going to lose out to elastic scattering.

[1] https://en.wikipedia.org/wiki/Nuclear_cross_section [2] D + T releases 17.6 MeV of energy; typical fuel temperatures are 15 keV.

Like hydrogen-boron fusion?

https://newsroom.unsw.edu.au/news/science-tech/laser-boron-f...

https://revolution-green.com/breakthroughs-make-commercial-l...

Other: Google for "LENR". Some people claiming (for years) they are near to success.

I've wondered, aren't there any fusible elements that can be solids in vacuum? Then shoot pellets at each other at a sizable fraction of c. Way better density than a particle beam.

For instance, lithium-6 is a stable solid metal with the right nucleon composition and binding energy to fuse into carbon-12. I have no idea if it actually would to a useful degree.

Solid in a vacuum at what temperature?

Whatever's practical for your giant pea-shooter and reaction chamber. But lithium is a soft metal at room temperature, so this part doesn't seem too constraining.

If you want to sustain a fusion reaction for generating power through heat, you're not going to do it at room temperature, and there're not many materials that will remain solid at fusion temperatures.

I said "shoot pellets at each other at a sizable fraction of c". The point is to hold together before the splat, so that by high density enough nuclei actually hit each other. This is a bang-bang kind of thing, not a continuous reaction.

We don't have any trouble creating fusion explosions. The whole trouble of fusion is sustaining a reaction. Fusion bombs already exist, and what you're describing isn't that different from an Ulam-Teller device.

Existing fusion bombs need a fission bomb for ignition. This lower-bounds both the bang and the unpleasant byproducts. A smaller bang can be used much more practically. For example, 20th-century cars were also powered by explosions. Somehow they outcompeted the previous dominant technology of continuous reactors, steam engines.

To dismiss my half-assed idea by pointing to another, I'd bring up something more comparable, like inertial-confinement fusion.

https://en.wikipedia.org/wiki/CERN

Maybe we just haven't tried hard enough.

While researching a comment for another thread I discovered an interesting fact: the US has now spent approximately as much money on fusion research ever (since 1955) as it did on the Manhattan project, alone. Perhaps we shouldn't be so surprised that progress is so slow.

Does fusion still have much of a role now that renewables are increasingly competetive? We can pretty much already harness infinite energy for all our needs from an already working fusion generator our planet happens to orbit around.

> now that renewables are increasingly competetive...

... with legacy energy sources.

When talking about fusion / nuclear vs renewables, think in terms of max energy produced per site, not cost per unit.

Similar to total thrust vs specific impulse.

The largest respective power generation stations by source -- Three Gorges Dam / Itaipu Dam (Hydro, 22,500 MW capacity, ~100 TWh/yr), Kashiwazaki-Kariwa (Nuclear, 7,965 MW capacity, 60 TWh/yr, currently suspended for earthquake-proofing), Tengger Desert Solar Park (Solar, 1,547 MW capacity, ? TWh/yr), Alta Wind Energy Center (Onshore Wind, 1,547 MW capacity, 2.68 TWh/yr, Gansu not included due to utilization issues), Walney Wind Farm (Offshore Wind, ~1,000 MW capacity, >1.3 TWh/yr?).

And then realize that hydro & wind are both location-limited. And solar has a large footprint: Tengger is 43km^2.

Nuclear (and eventually fusion) scales footprint much more slowly with capacity. Kashiwazaki-Kariwa is 4.2km^2.

For most of my life fusion has been teased but not delivered. fission has been pimped for decades yet cannot get private funding, or build reactors in a decade, or provide solutions for the waste. Yet still the industries persist with their propaganda while renewables beat them in cost and time to operation. If I managed the funds I'd cut nuclear and redirect to wind and solar immediately.

I don't think you're fully aware of the facts (not "propaganda") of nuclear. Old fission reactors are the silent workhorses of our civilization and without them we'd be a much much deeper trouble than we already are. Had previous administrations invested more in fission, climate change would be a much smaller and more manageable problem today.

Nuclear has other advantages over solar as well, such as not requiring the vast amounts of land solar does, sometime which will probably cause more and more problems as the industry grows.

Yeah this is something people keep forgetting. The reason fossil fuels have such an impact is because the industry is happening at such a scale.

Solar and wind are working at miniscule size ... and ... "no" impact (well, wind clearly has impact already, but ... let's forget that).

Well actually, solar does have an impact. In Spain there's a couple large wind farms I visited and ... not that it should surprise anyone but there's nothing growing near or below them. Below the panels there's dried out, sad grass. Now I guess this is what you'd expect, but still. It kinda sucks.

What happens when we have regions in every country the size of a decent state/province essentially covered in solar panels ? Habitat destruction, of course (even in deserts).

But for now, when we're talking a decent number of football fields, we can all happily pretend no impact.

Truth is, for the amount of power it produces, nuclear fission is not just amazingly, but almost absurdly low impact (meaning you'd get laughed out of the room if you predicted it in 1950). Far LOWER, not higher than solar, in impact, in danger (in lives lost / KwH, even if you include nuclear bombs), in place occupied, in nature poisoned/displaced ("environmental impact"), ...

Fusion can do better. That makes it worth pursuing.

Your facts appear incorrect.

Renewables consistently produce more than 60% of Portugal's energy needs, and have topped 100% on occasion. The figure for Spain is more than 35%, and more than 35% of electricity consumption in Germany.

I'm not sure how you get from 60% to "no impact". It's really quite a puzzling statement.

As is the suggestion that windfarms cause some kind of ecological blight to farmland. Not only is there no scientific basis for that claim, but if it were true, anti-wind propagandists would surely have jumped on it and made sure they mentioned it at every possible opportunity.

Perhaps you have some independent evidence to back you beyond the "couple large wind farms in Spain" you visited?

I'm very pro-renewable and new nuclear, but when you say "60% of Portugal's energy needs, and have topped 100% on occasion" do you mean the entire country's energy requirement, or just 'electricity'?. Because as rosy as that sounds, I doubt it takes into account oil-based travel infrastructure, which if the last pie-chart I recall seeing for an average country has any bearing, is a BIG slice.

The GP's post had "below the panels" in the description of the "wind farm", which was preceded by "actually, solar does have an impact". Surely the charitable explanation is that "wind" was just a mistake?

I think Fission would be great if we had somewhere to put the waste, if it didn't cost $10 billion dollars to set up a plant and people weren't phobic of it. Overall Nuclear makes more sense than coal and arguably more sense than wind and solar. But then there's the risk of the meltdown. No politician, no utility CEO, no corporation commission is going to take the risk of a meltdown when the political cost of pumping just a little more CO2 into the air is very small.

Your point about political risk is very valid, but your points about the waste and cost need to be reconsidered. A big part of the cost is artificial, not inherent: regulation. A person may or may not think the regulations are a good thing, but it's not an inherent cost. The waste is not much problem. The amount of waste to power the lives of a whole family for their entire lives would easily fit in a shoebox and be easily contained in water. And most of that radioactivity dissipates very quickly.

Don't forget to mention that old waste can even be consumed in modern reactors.

Perhaps one day we'll put the waste in cars, a la "Back To The Future ".

I can believe that human beings can run one nuclear fission plant successfully for one hundred years without a serious accident. I cannot believe that we can run thousands of plants for hundreds of years without a serious accident. There is too much complexity to prevent a black swan event, and in the case of nuclear fission, a black swan renders the surrounding area uninhabitable for an immensely long time. Nuclear fission is a fine bridge technology, but we cannot build systems that rely on it and expect them to be sustainable.

This thread on reddit had an excellent discussion on nuclear power.

There was some especially interesting discussion around thorium based nuclear power which has the potential to be much less dangerous.

https://www.reddit.com/r/Physics/comments/a1c7kt/why_do_peop...

Regarding solar placement, we can place panels on housing and commercial buildings and parking spaces. Check out canopy style parking covers(sun protection)in the southern US.

I once did my own calculation to answer the question: "How much land do you have to cover with solar panels to meet Toronto's electricity needs". The answer: about 3.5x the size of the city. Covering every square inch of the city is not sufficient, you need a lot more land.

You can double-check my calculations: take the output of largest solar farm installations, note their land area, adjust for solar radiation at your chosen location/latitude, then use the target city's electricity consumption to see how many of these solar installations will be required. Note, I'm talking solely about land-use requirements here, not cost/battery requirements etc.

So covering cities in solar would work great if the overall electricity mix was 50% nuclear for base load, 50% solar/wind/hydro for peak times, but relying heavily on solar will bring its own issues.

At Canadian solar intensity levels? I'd be surprised if it's only 3.5x. You can't count on average intensity for the year either, you really need to look at the lowest intensity and go with that number. You can use battery storage, but you are not going to be storing electricity from the summer for the winter.

Anyway even at 10x, Canada has a LOT of land, that's totally viable and wouldn't even make a dent in land-use compared to say agriculture.

Yeah, it's probably a conservative number. But the Toronto area (~20% of Canada's population) would not be much different from Chicago, New York, Boston, Philadelphia, etc.

And yeah, Canada has a lot of land, but once we ask Canadians "How would you like to chop down 5 Torontos worth of forest to cover them in silicon panels and make it into a lifeless desert?", then maybe some people will reconsider the relative merits of nuclear.

5 Torontos worth of forest is nothing. In BC for just the last two years we lost forest half the size of Switzerland to fires. If you think how much land is cleared and maintained clear for agriculture in Canada, 5 Torontos is a rounding error.

Land is cheap and abundant it's a non-issue for the kind of area we'd need to even power the whole world 100% on solar.

Habit losses would be more than made up for by not continuing to accelerate climate change.

Also, you can put solar panels on the roof of buildings in actual Toronto and not have to lose any additional footprint.

No one has ever suggested that renewables = solar.

Renewables are always a mix of solar, wind, hydro, and - eventually - wave.

Hydro is limited by your geography and largely tapped out. In Ontario it's about 20%. Trying to replace the remaining 80% of supply (plus future growth for EVs) with wind and solar (there's not much waves in Ontario) is an incredibly daunting proposition.

Again, the mount of electricity consumption is public info. The amount of output a wind turbine or solar farm per km^2 is available. Put these together and ask yourself how easy would be to convince people to do this.

Our old nuclear plants are our golden-egg-laying goose. Not only should we not kill it, other places should get their own.

> If I managed the funds I'd cut nuclear and redirect to wind and solar immediately.

To make those a reasonable base you need to add enough storage to keep everything working even when your current weather does not meet the energy demand. I think setting up that storage is currently going to be the biggest cost factor.

Base load power is not a useful feature. To follow the demand curve you generally what to turn massive amounts of production off at specific points in the day, base load is simply what you keep running during those periods.

Nuclear power plants for example where known to be hard to throttle which was a significant design issue. Historically, the tradeoff is worth it as the cost per kWh was less, but again not a useful feature and you would expect renuables to kill off other forms of base load power based on price.

I used to be a reactor operator and never saw any evidence they were hard to throttle. That's why they are used for warship propulsion: loads in battle are not predictable by their nature, what with changes in course, catapult launches, etc.

Let’s be more specific.

Compare to hydro which can often do 0-100% or 100%-0 in minutes and has little opportunity cost for doing so. It’s relatively cheap to add max capacity even if you can’t increase average capacity much etc. On the other hand increasing nuclear max capacity is as expensive as increasing maximum sustainable capacity.

Nuclear also has issues going from sub 5% and ramping to 100% with little notice. Thermal stress is an issue ramping up and down frequently over the life of the system. 2x cycles per day is possible but would adds noticeable lifetime costs, 30x per day would cause huge issues. Which means you need to look at ramping as an expense each time you do so which causes incentives to avoid doing so.

PS: Civilian and military reactors are also significantly different designs with different goals. Military reactors are more expensive per GW of capacity, making them a poor though relevant comparison.

You're both right- technically, nuclear reactors are on of the fastest-ramping sources available. A reactor in SCRAM can drop down to <10% of full output in minutes, and many reactors are capable of running at continuously at reduced power.

However, zero reactors are capable of doing so economically. The cost of a nuclear plant is in its construction, upkeep, and staff. Those costs are the same at 110% and 30% output, but revenue isnt. In order to turn a profit a nuclear plant needs to be running at 80%+, 24/7.

Ultimately that means nuclear needs batteries almost as much as solar and wind, since only ~30% of grid energy is constant 24/7. This should be no surprise- all three technologies have nearly no reliance on fuel, and are funded pretty much upfront. If you have no marginal costs then your only rational option is 100% output.

There is a flaw in your reasoning: By definition if a plant needs to be throttled then it's surplus power generation has negative (or zero) utility/value and therefore does not contribute to it's profitability whether it's running or not. Throttling the plant doesn't reduce it's profitability, since the power it would product at that point has no value - that's why you need to throttle it.

I'm not making an argument from semantics- whether or not you approve of the above explanations, they are the reason nuclear power is commonly said to be incapable of load-following. If you want a detailed and reputable investigation, the OECD has one[1].

The inability to load-follow is not a design flaw; any reactor built in the last ~40 years can ramp an order of magnitude faster than coal or gas-steam, and will run very happily at anywhere from 50-100% power, ramping up and down multiple times per day. Many can ramp at hundreds of megawatts per minutes. In fact all of them can ramp from 100%-10% output in about a minute, since that's legally required for reactor SCRAM.

That's as fast as an average gas turbine[2] and at the high end it's faster than the fastest gas turbine. Nuclear plants are the most agile nonrenewable plants on the grid but the fact is that a nuclear plant operating <80% output does not generate a return.

> Throttling the plant doesn't reduce it's profitability, since the power it would product at that point has no value - that's why you need to throttle it.

I get what you're saying, but it's wrong. The grid is not a free market and excess generation is not legal. Electricity markets are based around contracts and producing too much is just as heavily penalized as producing too little. If you generate so much excess that you bring grid frequency or voltage out of spec, people start going to jail.

To rephrase: If a nuclear plant throttles, its costs are constant and its revenue is lower. If it doesn't throttle, its costs are higher and its revenue is lower. Nuclear sees rising revenue/kWh by throttling, but not enough. Because of all that nuclear is not profitable in short term contracts. Generally they aren't used for ancillary services either, but that's more due to conventions- for instance they are used for frequency control in France.

[1]: https://www.oecd-nea.org/ndd/reports/2011/load-following-npp...

[2]: https://www.wartsila.com/energy/learning-center/technical-co...

Nuclear plants have high capital costs and low marginal costs for each bit of energy they make. That doesn't mean they can't produce all the power: France is at 80% from nuclear.

France reduces their penalty from load following by being the largest power exporter in Europe. Most grids drop down to ~30% of peak demand at night, while France only drops to around 75%[1]- almost exactly how much nuclear generation it has, because those nuclear plants are running at 100% as much as possible and exporting power. In 2015 nuclear made up 72% of their generation but only 40% of consumption- 40% of their consumption results in such an oversupply of nuclear that they have to export that much.

If anyone else in Europe tried this, or if the US tried it, it would be prohibitively expensive. Even in France, it's still very expensive- the EDF is 36 billion euro in debt[2]. French nuclear gets 9 cents per kWh in operating funds, slightly lower than the US- but the US is MUCH larger, and distribution costs are included in that. And again, the EDF is heavily in debt.

Nuclear doesn't readily beat natural gas on price, and the areas that it's better than renewables are rapidly shrinking. End of the day we'll need ~40% nuclear in the US, but that's not much growth compared to how much renewables need to grow.

[1]: http://clients.rte-france.com/lang/an/visiteurs/vie/courbes....

[2]: https://www.edf.fr/en/the-edf-group/dedicated-sections/finan...

> The cost of a nuclear plant is in its construction, upkeep, and staff. Those costs are the same at 110% and 30% output, but revenue isnt. In order to turn a profit a nuclear plant needs to be running at 80%+, 24/7.

The same can be said about wind and solar power.

...and I literally did, in the next paragraph.

I know nothing about the subject, but my understanding was that iodine pits could be a problem for thermal spectrum reactors if they had to suddenly raise power output after dropping it.

OTOH when you already have a steam turbine ready to go thermal energy storage looks like a mighty tempting option.

Can anyone in the industry chime in on thermal energy storage in fission plants? Do the economics work out?

When you factor in storage costs, how do solar and wind compare to alternatives?

Currently, solar that has 3-4 hours of output at 50% of peak primary solar outbut, is about $35-$40/MWh in Colorado, Nevada, and I think even Indiana had bids in that ballpark. This is between 1/5 to 1/2 the cost of nuclear, depending on who you believe for nuclear's costs. Storing a MWh probably costs around $80-$150 currently.

Why so little storage? It's not really needed at the scale of projects that are being planned, and storage gets cheaper every year.

So let's say we could start building a nuclear plant and get it done in 10 years (which is highly speculative, but let's give nuclear the benefit of the doubt). What are the costs going to be 10 years from now? Or more importantly, what are costs going to be 30 years from now? Because the storage lasts 10-15 years, so you get to replace it with cheaper tech at that point, whereas with nuclear you've locked in your costs for the next 60 years.

This is why no sane entity wants to build nuclear, unless they are ideologically motivated. There are some nuclear startups that may make nuclear cheaper, and less risky to build, but they are still startups without functioning prototypes or manufacturing facilities. By the time they start production, the utility will be ready to replace any batteries that get deployed on the grid today, so one may as well start deploying wind/solar + storage. And add more time of use rate plans so that energy is used efficiently.

We don't know, because we've never built grid-scale energy storage.

Still generally cheaper.

People regularly go off grid solar for less than the cost of grid electricity over the lifetime of their system. They save on the grid distribution system and overhead, but lose out on vast economy of scale. Peaking power plants also tend to be more cost effective than large scale grid storage. Having excess capacity also vastly reduces the need to storage.

Existing investments complicate the issue, but falling prices for both solar and battery systems regularly change the equation and this shift is only going in one direction.

Wow this is an insane view point, I actually thought you were going to suggest the opposite...

Frankly the potential capability of nuclear far outweighs anything else.

If the gen public were more on board with it, nuclear could be rolled out and be the main source of power in the world in 30-40 years, removing the need for fossil fuels for power generation.

Even building them in remote locations and taking the transmission losses would be better than diving head first into wind and solar, do you have any idea how destructive the mining is to get the elements required for some of the solar and wind solutions?

My big problem with nuclear (fission) is that while on paper it looks like a great solution, in practice our civilizational apparatus does not appear to be up to the task of long-term responsible implementation. Fission power looks like a long game of hot potato, where everyone hopes they won't be the ones on the hook when the eventual equipment failure, natural disaster, or time to decommission eventually arrives. (And, as an observer outside the industry, building a plant while taking all those into account in advance seems to be something we are either unable or unwilling to do.)

I posted further up thread, and I am not an expert, that thorium based nuclear power may be safer with less dangerous waste and potential for weaponization.

https://www.theguardian.com/environment/2011/jun/23/thorium-...

"Don't believe the spin on thorium being a greener nuclear option"

and

https://thebulletin.org/2018/08/thorium-power-has-a-protacti...

"Protactinium separations provide a pathway for obtaining highly attractive weapons-grade uranium 233 from thorium fuel cycles. The difficulties of safeguarding commercial spent fuel reprocessing are significant for any type of fuel cycle, and thorium is no exception."

What I got from the first article you linked is that Thorium reactors produce less radioactive waste but are unproven technologies.

Don't stop at the first sentence. Read the whole article. It answers the propaganda claim of "less radioactive waste":

"Less" is only if need thorium would magically materialize out of nothing, that is, if you ignore the whole process. In reality, thorium inevitably has to be prepared with uranium rectors:

"Thorium cannot in itself power a reactor; unlike natural uranium, it does not contain enough fissile material to initiate a nuclear chain reaction. As a result it must first be bombarded with neutrons to produce the highly radioactive isotope uranium-233 – 'so these are really U-233 reactors,' says Karamoskos.

This isotope is more hazardous than the U-235 used in conventional reactors, he adds, because it produces U-232 as a side effect (half life: 160,000 years), on top of familiar fission by-products such as technetium-99 (half life: up to 300,000 years) and iodine-129 (half life: 15.7 million years). Add in actinides such as protactinium-231 (half life: 33,000 years) and it soon becomes apparent that thorium's superficial cleanliness will still depend on digging some pretty deep holes to bury the highly radioactive waste."

Not to mention that these hypothetical reactors simply don't work: if they would be viable means to produce energy, nobody would wait for the taxpayer subsidies, there's enough money which couldn't wait to make huge profits, if they were possible:

"'Without exception, [thorium reactors] have never been commercially viable, nor do any of the intended new designs even remotely seem to be viable. Like all nuclear power production they rely on extensive taxpayer subsidies; the only difference is that with thorium and other breeder reactors these are of an order of magnitude greater, which is why no government has ever continued their funding.'"

Remember, the longer the half life, the less dangerous the waste as it dumps its energy more slowly. Consider technetium-99, for instance (from Wiki):

"The weak beta emission is stopped by the walls of laboratory glassware. Soft X-rays are emitted when the beta particles are stopped, but as long as the body is kept more than 30 cm away these should pose no problem. The primary hazard when working with technetium is inhalation of dust; such radioactive contamination in the lungs can pose a significant cancer risk."[1]

Laboratory glassware, safe distance 30cm. Hardly a "pretty deep hole"!

[1] https://en.wikipedia.org/wiki/Technetium-99

That's completely irrelevant to the issues mentioned, once again:

The economics aren't there. Solar and wind are well below the costs of nuclear and still dropping fast. Costs of energy storage are dropping fast as well. I think the window has closed on nuclear for the time being. Thinking as an investor it's too risky and too expensive to produce the same fungible commodity as a solar cell + battery.

We're fast getting to the point where wind and solar is going to wipe out fossil fuel generation - for new construction - in some places it already has. Never mind nuclear, nuclear just isn't in the game anymore.

> If I managed the funds I'd cut nuclear and redirect to wind and solar immediately.

Why? Research has shown that nuclear fission could be safer per kWh than wind and solar. Just because it's not as good as fusion doesn't make it worthless.

> nuclear fission could be safer per kWh than wind and solar.

Sounds completely improbable. Care to provide some reference to back that statement up?

Some examples of research that has reached this conclusion:

https://www.iaea.org/sites/default/files/publications/magazi...

https://www.nextbigfuture.com/2011/03/deaths-per-twh-by-ener...

It might sound improbable if you are comparing from a purely technological point of view. But the reality is that the energy output of nuclear is so much higher than other technologies, it easily makes up for any increase in risk by requiring much fewer plants. That's what makes nuclear so attractive in the first place.

You fail to mention that Herbert Inhaber's article was published in the International Atomic Energy Agency (IAEA) Bulletin in 1979 [1] (before Three Man Island), drawing from his study published the same year which was heavily criticized even at the time [2] and does not take the problem of nuclear waste into account at all, and otherwise dis-ingeniously cherry-picked number of deaths associated with each energy source up to that point.

The second article you quote is just a blog, and the conclusions are just as dubious, for example trying to weigh falls during roof construction into the risks of solar etc.

edit: removed confusing statement about fission/fusion

[1] https://www.iaea.org/publications/magazines/bulletin/21-1

[2] "Perhaps the most damaging criticism was that the study was inconsistent in applying the methodology to the various energy technologies. For example, while Inhaber considered materials acquisition, component fabrication, and plant construction in his analysis of unconventional energy sources and for hydro power, critics have claimed that he did not follow the same approach for coal, nuclear power, oil, and gas. Furthermore, critics claimed that the labor figures for coal, oil, gas, and nuclear power included only on-site construction, while those for the renewable energy sources included on-site construction, materials acquisition, component manufacture."

https://www.researchgate.net/publication/267817006, page 31.

my point in short is we don't have time for this right now. we need to de-carbonize our energy system a decade ago. this tech. is not operational and will take a decade to be so - that's not time we have to avoid catastrophic climate destabilization.

If the objection is that nuclear can't be built quickly enough, then the renewable options need to be judged against that same standard.

Suppose everyone in the world already agreed to go a hundred percent renewable in 10 years. Effectively replacing the entire world's energy infrastructure is going to take a huge amount of raw materials and production capacity. Is there going to be enough?

It's possible the answer is yes, but I wouldn't take it for granted. There will need to be a lot of battery storage, which means there needs to be a lot of lithium. Are there enough lithium mines? Have geologists identified enough deposits to cover that massive quantity? And are there enough factories that can build batteries? For wind, there will need to be neodymium permanent magnets. For solar, you need fabs to build it all.

Maybe all those challenges can be solved, but maybe it's also possible to streamline the process of building nuclear fission power plants.

> There will need to be a lot of battery storage, which means there needs to be a lot of lithium.

No, if we actually want this, pumped-storage hydroelectricity is a proven solution that works. There is actually no uncertainty regarding energy storage: we can do it.

There is only limited number of sites where building pumped hydro is possible, and lots of them are already taken.

Are there enough lithium mines?

No need to - there are 230 billion tonnes of the stuff in seawater.

Just one percent of that is enough to create 25,000TWh worth of li-ion batteries - that's roughly the amount of electricity produced by the world annually.

Also other chemistries like zinc-air are gaining traction and serve as a good, cheaper alternative.

Can you point to a commercial operation recovering lithium from seawater? Demand for lithium is high and expected to go much higher, so a seawater extraction plant should at least be in development if it's feasible to do that.

If it's purely a future technology, just like zinc-air batteries, then we should be allowed to look at future nuclear technologies as well, some of which are very much in development.

Can you point to a commercial operation recovering lithium from seawater

No, because it's exactly like with tar sands: it will happen when it's commercially viable. At this point it's much cheaper to "mine" the brines in Chile.

If it's purely a future technology, just like zinc-air batteries

Thing is - it isn't.

Here's a company making grid-scale Zinc-Air batteries:

https://eosenergystorage.com

They figured it out in 2015 and they're already sold-out for 2019 and secured funding for a production facility.

And this is just one of many companies doing the same.

I'd be happier with reference to EOS if I had been hearing more about them. They seem to have fallen off the map the last couple of years, which does not fill me with confidence.

They tend to do that - a rapid surge in press releases and after that up to a year of silence.

I've stumbled upon them years ago back when they were promising 1MW/6MWh modules, 10k cycles and usage in EVs.

They backed down on these promises somewhat since, so I guess what they have now is a real product.

I expect some more news late summer when the construction of the plant will be - hopefully - already in progress.

Thanks, didn't know zinc-air was available for grid batteries.

Btw uranium extraction from seawater has been demonstrated, and isn't used just because it's 5X more expensive than mining. That's not a showstopper because fuel is a small portion of nuclear energy cost.

Unless you have a very cheap and very scalable storage solution, that would be a mistake.

France took their grid from zero nuclear to 80% nuclear in two decades. We don't have a comparable example for wind/solar.

That’s a function of how recently solar and wind became cheap. China has a 31x increase in wind power over the last 10 years. The next 10 will be very interesting assuming prices continue to fall. When having a 30% oversupply of wind / solar still costs less than any alternative and they are only gettting cheaper eventually noting else comes close.

It's also a function of not yet having really cheap, scalable storage that's not limited by geography.

In any case, it's clear that nuclear can do the job if we're smart about it, because France did it.

Right now, France is running almost carbon-free: 83% nuclear, 14% wind and 10% hydro:

http://www.gridwatch.templar.co.uk/france/

But that works in part because they're also exporting 15% of their generation to the UK, Switzerland, Italy and Spain. The export is reducing the need for gas plants in these neighbours, which have higher marginal costs. If France was an isolated grid, or if the neighbours had a similar generation mix, it would be harder to do this without a lot of storage, or throttling the nuclear plants.

Exactly, Hawaii is an example of an independent grid, France does not operate an independent grid any more than Ohio does.

The majority of people in power will operate by the ends justify the means. I've always wondered if this is true, why humanity will not readjust population numbers and where the renewables actually can sustain the population. Seems better for increasing the length in time for humanity to exist on the planet and come up with a solution to energy where the population can progress to infinite. Compared to the alternative outcome of everyone dies by ruining the planet. Maybe it's already too late for humanity and thus people don't care.

Brilliant solution! Since you thought of it, perhaps you'd like to be 'readjusted' first?

Sure, legalize suicide and make it possible to receive in a humane medical setting.

Why do you need it to be legalized? After all, it's a one way trip, you don't care about being convicted post mortem, do you? :)

I care about life insurance payout to my relatives in case I become terminaly sick and want some dignity.

Statistically failure rate is high and where the very real outcome of brain damage is possible. Wouldn’t want to be stuck in a state causing more resources spent and unable to attempt the act again. I am seeking it medically for the record and I assume most people are just irrational against letting people die which is why this world is in jeopardy anyway.

Interestingly and sadly a rather bright young man set his mind to exactly this problem and engineered a solution. https://imgur.com/a/Z5mEB#0bVIYlf

Yah I've seen it before and personally I know someone who had success with a shotgun but I would prefer something guaranteed by society.

Please seek help

Online: https://suicidepreventionlifeline.org

Phone: 1-800-273-8255

What makes you think I haven't attempted to use all the available resources yet? I'm seeking the next best solution after rationally doing all that could be done.

Wrong title. Fusion is easy, plasma physics isn't.

Fusion is likely to be a nonstarter even if the plasma physics issues are completely solved. It has fundamental engineering and cost problems.

It's hard now because we don't know how to do it well. In the future, we'll learn, then it will be easy. History tells us this.

Mankind is currently in possession of a practical fusion technology.

It might be worthwhile to remember that Ivy-Mike fission-fusion technology worked the very first time it was tried in 1952. Mike technology was the basis of the first thermonuclear weapons in the US arsenal. Adapting Mike technology to be pure hybrid DT-DD fusion opens up many new applications in economical power generation.

In 60 years, no other fusion technology (Magnetic Confinement or Inertial Confinement) has ever produced any net energy (more energy out of the fusion reaction than it takes to get the fusion plasma to fusion conditions).

In 60 years, all existing MCF and ICF fusion systems have never worked (in the sense that they have not produced more energy from fusion than it took to get the fusion plasma to fusion conditions).

Mike fission-fusion technology worked the first time it was tried and produced huge amounts of net energy (and has never failed).

Rather than placing our faith in scaling laws while we build ever larger and more expensive Magnetic Confinement fusion experiments (tokamaks and stellarators) while trying to achieve break even energy generation - why not go back to the field and adapt technology that has never failed to finally find success in fusion?

They looked at this. It's grossly uneconomical as a power source.

"In a 1975 review of the various Plowshares efforts, the Gulf University Research Consortium (GURC) considered the economics of the PACER concept. They demonstrated that the cost of the nuclear explosives would be the equivalent of fuelling a conventional light-water reactor with uranium fuel at a price of $328 per pound."

So, how do you turn a bomb into a practical source of power that doesn't destroy the surrounding area the moment you turn it on?

Reply on news.ycombinator.com