Hacker News new | past | comments | ask | show | jobs | submit login
Traveling Wave Reactor (wikipedia.org)
76 points by jonbaer 10 days ago | hide | past | web | favorite | 66 comments

Nuclear designs are like quarterbacks, the best one is always the one "in development". The problems with all of these next-generation designs are not going to be with the physics, but the engineering. These reactors will be operating for 40+ years, and are complex collections of machinery that have wear-out and failure mechanisms that may not be discovered until years after full operation commences - when this machinery is now highly radioactive. Remember the pebble-bed reactor? Sound idea in principle, the Germans found out that its main problem was inherent in the engineering of the fuel - it turned out to produce large amounts of radioactive dust in the reactor.

The description of this reactor, which involves robotic fuel handling equipment operating within the containment constantly moving around fuel elements immersed in molten sodium metal seems doomed to create difficult repair problems as these mechanisms will fail and now need to be replaced/repaired within the hot zone of the reactor. It's quite easy to build a system which technically works, but ends up being uneconomic. To my knowledge there has yet to be a successful long-term operating power generation design of any kind using sodium coolant; the Monju plant in Japan was a failure largely because of problems with the coolant loop (including a leak and serious fire). Until these basic engineering issues are proven to be solved, the fuel physics seem kind of beside the point.

Elon would say that nuclear power has a "high PITA factor."

PITA factor is "Pain in the Ass" factor and refers to technologies that look great on paper but explode into a fractal of non-obvious design and operational difficulties when put into practice. He used the term to refer to hydrogen as rocket fuel for example.

Excellent points. In this case, we have Bill Gates providing essentially unlimited capital to work out the problems - even if that does take decades.

Looking at TerraPower's website, they're talking about a prototype by the mid-2020s, and they've already been working for 11 years. Would be nice to see this sped up. The Manhattan Project lasted only 4 years. We spend much more time on the ethical, safety, and proliferation considerations than we did back then, of course, and with good reason. But I wonder if this project could use a little Elon Musk hell-or-high-water drive to move it along. I'll probably get down-voted for that, but seems like a discussion worth having.

It's shocking to me that the best solution we have for climate change gets downvoted on a site that's generally pro-science. But it seems to happen with every discussion about nuclear energy.

Also surprising how even Bill Gates gets ignored on this issue. He's talked repeatedly about why nuclear[1][2][3], yet doesn't get a lot of traction.

1. https://www.gatesnotes.com/About-Bill-Gates/Year-in-Review-2...

2. https://youtu.be/d1EB1zsxW0k?t=520

3. https://ourworldindata.org/what-is-the-safest-form-of-energy

It's shocking because you are laboring under the misconception that nuclear is the "best solution".

It's the best known solution we have today.

The misconception is that wind and solar are able to replace fossil fuels. They can't as of now, that's just the unfortunate reality. They're a good supplement, but they don't produce 24/7, don't recycle well, take up tons of space, and batteries are orders of magnitude too expensive to be a complete solution. See the video I linked to where Gates explains this.

Batteries are orders of magnitude too expensive if you expect to cover seasonal differences with batteries. But no one in their right mind would do that. You are engaging in a kind of strawman argument I call the Argument from Bad Engineering.

Instead, batteries would be used for short term leveling (diurnal, perhaps a few days). Beyond that, leveling would be by demand dispatch, transmission, dispatchable demand, and long term chemical storage (for example, hydrogen or ammonia). In the shorter term, covering the last 10% with natural gas would be workable, even if the CO2 had to be captured and sequestered.

I didn't set up that strawman. They're too expensive for a lot more than that.

Seriously, watch Bill's response to this and I'd be interested if you or anyone has a better answer: https://youtu.be/d1EB1zsxW0k?t=520

A utility scale battery system might impose a cost of $200-300/MWh (source: Lazard 2018)

You might say "nuclear is only $110-180/MWh!". But that's for every MWh that nuclear pumps out; batteries are only on the hook for (some of) the times wind and solar aren't covering demand. So the levelized cost of stored and direct power from renewables will be cheaper.

Batteries are also declining rapidly in cost. Any nuclear plant started today will come online when storage costs could be half of what they are today, if this trend continues.

I will add that DC-coupled storage in PV fields should be cheaper than this, because the batteries don't need their own inverters or grid connection.

Gates claims Tokyo needs 23GW for 3 days a year during a monsoon.

Tokyo has less cars per household than many Japanese cities but there's still about 3 million of them.

If those cars each had a 50KWh battery then that would be enough power to run the city for about 1/3rd of a day.

But since replacing all the ICE cars with electrics is forecast to save money and lives and help make cheap renewables even cheaper, it seems like they could work something out to tap into that when necessary and have solved a fair chunk of the problem without spending any money (in fact saving quite a lot which they can put towards this problem).

Gates appears to think providing that 23 GW with nuclear, at $10,000/kW, is better than providing backup combustion turbines at $400/kW.

> best solution

For some values of "best". There is no such thing as a one-size-fits-all power source.

While I sympathise with your sentiment, viable is enough. We aren't exactly sd spoilt for choice

There are many other things we could do that are as viable, if not more - we barely started using wind or solar, and while wind and solar may be poor replacements for coal, hydro or nuclear, it's excellent for cars and trucks and they are a significant portion of our energy matrix.

We will eventually need to increase the fission base, if fusion continues to be 10 years away for much longer, but that's something that can wait until we have better designs than the current generations.

What nuclear really excels at is at replacing coal. It's pretty much a plug-in replacement: nuclear generation can be built pretty much wherever coal can and, from the grid's perspective, it looks a lot like coal. This, along with industry subsidies (both direct and as military expenditure) has always shifted economics towards it.

Why will we eventually need to increase the fission base? I see no good justification for that.

> What nuclear really excels at is at replacing coal

Coal is being displaced, but not by nuclear. Nuclear is actually rather bad at displacing coal. People were building coal fired power plants for decades when they could have been building nuclear plants, because nuclear failed at being attractive enough to replace coal.

Nuclear is dense and constant, far more dense than wind or solar.

BTW, burning coal for energy should be considered a crime against humanity. It should be replaced by anything, regardless of energy output ;-)

"Density" only matters insofar as it affects cost. And wind and PV come out much cheaper than nuclear, so density is clearly not very important. One could have figured this out by looking at the cost of land vs. the cost of wind/solar equipment put on that land.

As for constancy... this is actually a negative for nuclear, since the business case for nuclear assumes it can sell its output at a high average price, 24/7. And increasingly that market isn't there.

Density will eventually become relevant. What is the maximum human population of the Earth? How much energy would a planet-sized city need? How much space for cultivation or food manufacture?

Agreed about the current market factors, but constant is easier to manage.

If density becomes relevant for solar, then nuclear is no solution. At that very high power usage, the raw thermal pollution from nuclear will limit it to below what solar can provide.

Of course, this will be at an energy usage that is so incredibly high that it has no relevance to anything we do today. So it's a silly point of no relevance to current decisions.

It’s political. The American left has taken an anti-science, moralistic stance on climate change.[1] Fixing climate change through technology, without changing the economy or politics, doesn’t serve desired ends. While most people here are scientifically minded, they’re also generally left leaning, and the opposition to nuclear power is part of that platform.

[1] That’s not to suggest the right has embraced science by denying climate change is happening. As far as I can tell, the right’s solution to climate change is “god won’t let anything bad happen.” The left’s isn’t any better or more realistic. I’m at a Costco in the Maryland exurbs. A hugely diverse crowd (black, Hispanic, Asian, white) is doing their weekend shopping. Families with 2+ children in tow. No joke, some just shouted out “living the suburban dream” while we were eating froyo. Then they’re going to get into their SUVs and go home to their, on average, 2,000+ square foot house.

The idea that we’re going to solve climate change by having all these people move to 500 square foot apartments in the city, going only where the government decides to build public transit, and limiting themselves to one kid, is utter fantasy. In fact the opposite will happen. Chinese, Indians, Bangladeshis, Africans... they’ll get Costco and SUVs. Hoping otherwise is just as unrealistic as thinking God will fix things.

I'm fairly pro nuclear. There was a time when nuclear was probably the best solution to stop climate change. But that time was 5, maybe 10, years back. Since then, solar and wind have caught up significantly in terms of efficiency.

And given the dearth of new nuclear in the us, safely bringing up a reactor, if we started building tomorrow (ignoring the political challenges of where), would take the better part of a decade to be operational. Solar and wind brings faster returns.

Really we just need to stop subsidizing coal and similar.

It’s not just a matter of efficiency. Investment in nuclear has major collateral effects that investment into wind and solar does not have. You won’t have wind powered spaceships. You can’t use wind power to for weapons or aircraft carriers. Nuclear fusion is going to be the bedrock of the next generation of world superpowers. Wind and solar are technological dead ends.

Spaceships? That's the supposed payoff for an energy source that is 4x the levelized cost of the competition?

I tell you what: if someone wants nuclear spacecraft, let THEM pay for it.

If we’re going to subsidize a technology we should think about the whole range of benefits.

It's not a benefit, it's a joke. Nuclear spacecraft were never valuable enough for NASA to spend much money on; why should they justify the enormous cost penalty of new nuclear power plants?

(BTW, the reactors in power plants would be useless for spacecraft.)

There are lots of uses for nuclear technology in space other than propulsion. Nuclear has huge power to weight advantages, critical for space. Nuclear would also be vastly cheaper if we deregulated the industry and enabled mass production of components, as with nuclear and solar.

(As it is, China will ship us subsidized solar while developing thorium technology itself, enabling it to become the dominant technological superpower.)

> Nuclear has huge power to weight advantages, critical for space.

The most powerful space nuclear reactor ever launched, the Soviet Topaz-I, had a specific power of 16 KW(e)/kg.

The PV system on the NASA Juno spacecraft, at 1 AU, had a specific power of 50 kW(e)/kg.

Perhaps you are talking about aspirational PowerPoint engineering nuclear concepts (and ignoring similarly advanced PV concepts), or maybe systems operating in the outer solar system. But otherwise, no, solar has the weight advantage over nuclear, not the other way around.

The point is these are niche applications that cannot carry the weight of justifying civilian nuclear power.

These are not “niche applications.” Increased energy production to enable better weapons, better transportation, and increased production is what has historically enabled leaps in civilization. The first country to perfect nuclear fusion is going to have a huge advantage over everyone else. Conservation and wind power aren’t going to take you in that direction.

Wait, you were talking about space, not "energy production" in general. The space applications are niche, because space itself is niche.

Your claims there with respect to terrestrial energy production are not justifications for nuclear. Wind and solar are much cheaper than nuclear, and rapidly getting even cheaper.

Space is niche right now, and will always be niche until we get off this rock and have space colonies.

Oh, they were. NASA was working very hard on nuclear propulsion, and even managed to get support from Congress in the budget, something they never had for Shuttle.

Took direct presidential order to dismantle the project and essentially bury nuclear space propulsion.

The point I am making is that when push comes to shove, NASA doesn't rate nuclear propulsion work highly. It's not what's keeping us out of space. The value to the nation of nuclear rockets is reflected in what $$ are budgeted to them.

The way things are budgeted in USA doesn't reflect much on value the nation puts on anything.

That said, US Congress did fund it, on a rather non-pork-barrel project, even. It just happened to have enmity pf Nixon's administration, to the point that they arranged for NASA to scrap Space Nuclear Propulsion Office and laid off everyone involved, despite 100m budget and the project being well on track.

Investment in fission technology has no practical impact in fusion development.

Practically speaking, better solar will be more helpful to spacefaring than better fission.

In the early days of nuclear power we got things done a lot faster. Now in the U.S. the government makes things unnecessarily difficult; the NRC requires a couple hundred million dollars worth of design work before they'll even take a look at a new reactor. Then they give a flat yes or no, and if it's no then you're out of business. You're pretty much forced to do a big design up front; you can't do a series of experiments and adjust as you go.

Things are different in Canada, and that's why reactor startups like Terrestrial Energy are making good progress there. But the U.S. DOE has a lot of infrastructure that could really help.

The Manhattan Project also cost something like $40 billion in today's dollars and had the benefit of conscripting some of the most brilliant minds in the world (including many Europeans who fled the war) to work on it for modest salaries.

It sounds like a lot to me, but it's worth noting that there are individuals with more money than that (and certainly plenty of tech companies who could invest that much). The Gates Foundation alone could fund that investment.

Per the Wikipedia article the prototype is dead:

> However, in January 2019 it was announced that the project had been abandoned due to technology transfer limitations placed by the Trump administration.

It is a heat source for thermo-electric generation - basically just a way to boil water[1] not the "unlimited" power its always promoted as promising. We don't have unlimited water resources[2].

Elon Musk decided big disposable rockets where just not worth building - well big thermo-electric plants are just not worth building. Solar and wind are the paradigm change in power generation, any Manhattan project would be crazy not to focus on them now and supporting technologies.

[1] https://www.usgs.gov/mission-areas/water-resources/science/t...

[2] https://news.ycombinator.com/item?id=21156408 - "Global groundwater extraction a ticking time bomb”

Given sufficient power and seawater, we have unlimited drinking water. Reactors provide sufficient power.

So I don't think "running out of water" is a valid concern here. (Though some plants have had issues from lack of water flow in rivers, that's more a temperature issue than a lack of water one)

More magic unlimited thinking I'm afraid. There are costs to desalinating seawater, cost of the plant, cost of locating plants to suitable coastlines, cost of disposal of concentrated brine, cost of introducing gigawatts of heat into water courses and local region - in an already heat stressed climate.

Its big resource intensive thermoelectric generation, with or without unlimited fuel source.

You seem to misunderstand the issue. Reactors do not "consume" freshwater, primary and secondary coolants runs in a closed loop, water never "leaves" or runs out. There are also variants that use liquid metals

The powerplant requires a cold reservoir to discharge heat, as does any heat engine. Cooling towers that evaporate fresh water are one of several methods for doing that.

Coastal powerplants use seawater to discharge heat.

Reactors in megawatt range usually use passive cooling and air cooling, you could operate them in the middle of sahara.

> Reactors do not "consume" freshwater, primary and secondary coolants runs in a closed loop, water never "leaves" or runs out.

That is untrue. They have the same cooling requirements as all thermoelectric power plants do, and some like eg. Diablo Canyon do straight up boil water away. Others heat up lakes or coastlines - which costs (more plant, land and maintenance) and can only be done in some places [1]. Plants cooling requirements are a big part of their overall cost, even with the legacy environmental protection which has led to our current crises.

[1] - https://www.ucsusa.org/clean-energy/energy-and-water-use/wat... Types of cooling

( https://iopscience.iop.org/article/10.1088/1748-9326/7/4/045... )

The part you are quoting as 'not true' talks about primary coolant, which is radioactive as that's the water flowing thought the reactor core. I assure you it never leaves the station.

I think the overall thrust of your post is in general agreement with mine - reactors have to discharge heat into seawater, or use cooling towers which evaporate water. This the issue is of correctly designing the cooling of, and sighting of, the powerplant.

Furthermore, this is not an inherent quality of nuclear power, it's a product of deploying it in the form of >3000 MWt power-plants.

If we adopted the approach of small module reactors, they can be passively cooled.

Sorry I was a bit short to simply say "untrue" but I had not raised the prospect of radioactive release and am aware of closed layers of cooling. I pointed out all thermoeletric plants consume significant amounts of increasingly scarce water resources and it seems only misleading to reply "Reactors don't consume water...". Its like saying people don't consume gasoline (their machinery does) Reactors don't consume water (their electrical generation systems do)

> Furthermore, this is not an inherent quality of nuclear power

I never suggested it was - although nuclear plants due to having more complicated safety requirements than other thermoelectrics, do tend to have lower thermal efficiencies and higher cooling costs per energy unit produced.

My point was those costs are significant at any practical scale of generation. The heat output itself of thermoelectric plants can be environmentally significant, so the idea of reactor breakthroughs promising "unlimited" generation is, well, its basically a marketing line, for research funding.

After watching the Bill Gates Netflix documentary I was wondering what HN would think of the nuclear safety debate in regards to TerraPower's traveling wave design.

I am very wishful but its still stuck in the wishful thinking phase. Obviously the 'trade war' derailing it is a travesty, but I think its important at this point (post SR15) to recognize that wishful thinking for future tech is a form of soft/self-delusional climate change denialism. The graphs are very stark at this point. We need to turn the corner on emissions in 2020/2021. That means anything still in the R&D phase is not going to be part of the first decade of the solution. This is basic project management not even an opinion on nuclear safety/proliferation.

Yes, but if you believe that going from our current level of wind/solar to 80% wind/solar is a lot easier per gigawatt than going from 80% to 100%, then it still makes sense to develop advanced nuclear for the last part, while still rolling out wind/solar as fast as possible right now.

Not to mention, fast reactors like the TWR can destroy the nuclear waste that people claim to be so worried about.

And they are not the only kind of reactors than can. Fast breeders can too, as well as MSR, while producing usable isotopes (e.g. medical).

Waste is a solvable problem.

Solar has a really bad eroi problem and I say it as an enthusiast and an owner of a microinstallation on my roof. New nuclear is absolutely necessary but it just can’t come online for a million reasons.

Some countries manage better than others. Perhaps they'll profit from electricity or hydrogen exports in the future.


Solar has a good, and continually rising, eroi. But eroi isn't really a particularly useful metric:


EROI of solar is just fine. The notion that it's not is based on well-debunked flawed studies.

Eroi is a red herring. The question should be energy returned on manual labor.

How is EROI a red herring? Energy returned on manual labor may be interesting wrt. profitability, but EROI is literally whether there's any sense of doing it in the first place.

The thing is: if you can output (through highly automated processes) a device that yields even just 1% more energy than it required in production over its lifetime, you have the opportunity for an exponential growth in energy production at your hands. Higher EROI might indicate a steeper growth, but you can achieve the same by just scaling up production.

My theory is the documentary was actually a 3 hour commercial for the Traveling Wave Reactor.

An alternative: mix californium and natural uranium and you get a solid state subcritical reactor with no need for regulation.

Though, there will be no way to turn it off until Cf burns out.

This is a completely terrible idea.

Let's suppose we have a just barely subcritical reactor with k = .99. Then, you're going to need about 1% Cf in the core to sustain a chain reaction.

The core of a 1 GW reactor contains about 100 tonnes of fuel, and fissions maybe a tonne of fuel per GW(e)-year. So, at k = .99, it would consume 10 kg of Cf per year. At current market prices this would cost $270 B/year, assuming even that much Cf were available.

It would be BY FAR cheaper to provide the neutrons to drive a subcritical reactor by means of an accelerator, rather than with californium. 1 GeV protons on a uranium target will produce about 60 neutrons per proton, from spallation and fast fission. This sort of scheme has been repeatedly investigated, but has no sufficient advantage over conventional reactors to justify the extra cost and complexity.

Hmm, I think you have a miscalculation about neutronics here, and fission/capture ratios. Cf neutrons are in their majority come nicely in 1MeV-2MeV, so you have much more fission going per neutron, and much less thermal neutron captures.

If you have an intense neutron emitter dispersed in fuel + moderation + reflectors you should be able to go with much lower k to generate a meaningful amount of power.

You should me able to go down to single grams of Cf per megawatt. Yes, not that economical, but at least possible for things like reactors for space use and such. At that price, that will be cheaper than current RTG material.

Correct me if I am wrong.

Correcting away!

The energy of the Cf neutrons doesn't matter much, since after that first generation the neutrons will be ordinary U fission neutrons.

The comment about moderators/reflector is silly, since the purpose of those is to make k higher. One you know what k is, they are irrelevant.

The ratio of Cf burned to (Cf + U) is 1 - k, so unless k is very close to 1 you are going to use a hell of a lot of Cf. But if Cf is very close to 1, you need to control it anyway as fuel burns up, so you might as well just go critical. Cf buys you nothing.

Hi, I checked my math and found it completely unsound. It will still have to be very, very close to criticality, though with a bit lower critical mass.

It will also require regulation over the years as Cf burns out, and u238 fast fission products themselves product less neutrons than fission products of thermal neutron reactors.

Er, "if k is very close to 1"

I am sure the use of Californium will also generate some extreme NIMBY.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact