Hacker News Reader

Comment by credit_guy | Hacker News Reader

Here, I'll come up with a proposal. If Congress is serious about climate change, then they can ask (and allocate the budget) the Department of Energy to procure and operate a bunch of naval nuclear reactors. With whatever internal regulations they have, the US Navy has not had a single incident in their entire history of operating nuclear reactors. They are also quite cost effective, for example the cost of the 2 reactors A1B [1] that power a Gerald Ford carrier is about $1 BN. That comes to about $2BN/GW, which is about a tenth of what a civilian reactor costs. The US Navy builds about 1 carrier every 4 years so that comes to 1 reactor every other year. If the DoE gets the Congressional mandate to procure a few reactors per year, the cost is going to surely come down. Also these reactors don't need refueling for about 2 decades, while civilian reactors are refueled every 1.5 years.

[1] https://en.wikipedia.org/wiki/A1B_reactor

View on HN

This is not a very good idea for several reasons. Naval reactors require fuel that is much more enriched than normal reactors. They also produce significantly lower electricity. The Palo Verde facility produces 3GW of electricity and cost $11B in 2019 dollars. Each of the A1B reactors generates 125 MW. Life span of the reactor is not specified, but it's predecessor the A4W had a 23 year life span. By comparison, new nuclear plants are slated to last 50-80 years.

The net cost per GWh of electricity of the naval reactor is significantly worse than commercial plants. This is to be expected, because naval reactors are built to be compact and withstand the rocking of a ship at sea. Commercial reactors can leverage the efficiency of larger scale, and are built to be much more long lasting.

An A1B generates 125 MW electricity, but also 260 MW of additional thermal power used to power the propellers. If you convert the latter one to electricity at a 45% efficiency (typical efficiency for a generation IV nuclear power plant steam turbine), you get 117 MW, for a total of 242 MW. Two reactors could produce then about 0.5 GW. At a $1 BN cost, that's $2 BN / GW.

Palo Verde was brought online more than 30 years ago. If you look at Vogtle 3-4 (to be brought online in the next 2 years... if we are lucky) or Hinkley Point C, you'll see projected costs of respectively $25 BN for 2.5 GW and $32 BN for 3.2 GW. In both cases that comes at $10 BN/ GW. That is 5 times more expensive than the naval reactor.

Now, as you said, the cost of a naval reactor is very likely inflated by the exacting demands of its military usage. It needs to be compact, to work on a rocking ship, presumably it needs to be able to survive a certain amount of abuse that's to be expected if a ship/boat actually participates in combat, and I'm sure there are 100 other things that I'm missing here. All these factors make military devices absurdly expensive compared to the same devices intended for civilian use.

The logical conclusion is that if DoE wants to repurpose naval reactors for civilian use, then it can achieve significant cost savings. What I'm saying is that even not factoring these savings in, you still end up 5 times cheaper than the civilian reactors that are currently being built.

Edit: The lifespan of a Gerald Ford-class carrier is expected to be 50 years. The Nimitz aircraft carrier was launched 49 years ago. They do not replace their reactors. So, a naval reactor is designed to work for at least 50 years.

You also need to build a secondary containment vessel for the reactor, which is a significant expense. Because the cost of this containment is a function of surface area and generating capacity is a function of volume it's better to increase size. You also need to build steam turbines, heat exchangers, transformers, etc. The cost of the reactor is only a portion of the cost of the whole nuclear plant.

> Palo Verde was brought online more than 30 years ago. If you look at Vogtle 3-4 (to be brought online in the next 2 years... if we are lucky) or Hinkley Point C, you'll see projected costs of respectively $25 BN for 2.5 GW and $32 BN for 3.2 GW. In both cases that comes at $10 BN/ GW. That is 5 times more expensive than the naval reactor.

And by comparison you have the Taishan plant built for $7.5B with 3.5 GW generating capacity. If we want to go around cherry-picking examples we can also cherry-pick the cheap plants.

We have already tried using maritime nuclear reactors for grid generation. The first nuclear plants brought online for grid generation were maritime reactors repurposed for grid production. Larger purpose-built reactors won out.

Vogtle and Hinckley aren't cherry picking expensive plants, it's cherry picking middle of the road.

VC Summer is expensive, many billions spent and nothing to come of it ever.

Where do your cost numbers from Taishan come from? How do you come to costs that are believable from massive Chinese construction, or at least a cost that might be transferable at all to the rest of the world?

The history of nuclear is very clear: keep on increasing costs throughout construction, just enough that, taking into account the sunk cost fallacy, it makes sense to soldier on. VC Summer overshot that, and had massive corruption in the auditing of all parts of the project. Somehow Vogtle continues.

We literally do not know how to build nuclear in a cost effective manner any more. We can't structure contracts in the right way, we can't perform engineering to a high enough degree to make constructive plans. At Vogtle they literally poured the wrong concrete, and had to go back and get the design recertified with the NRC, because the original design was impossible to build, and on site they just plowed ahead with what they thought they could build. This is the level of incompetence, ball dropping, and bad contract structure.

Perhaps this sort of thing is fixable, but not on any reasonable timeline. The management is rotten from the top, so there's nobody that we can even order a nuclear reactor from.

Suppose you had $7.5B and wanted 3GW of nuclear at one of the many sites in the US that would welcome nuclear and its jobs. Who do you even bring that money to in order to build it? Rosatom? Are they going to meet NRC standards?

Great by the same logic you should use a commercial reactor over a naval reactor, you should also just use a different power source over nuclear.

What other power source generates carbon-free energy without intermittency or geographic dependency?

The construction, fuelling and cleanup of a site is far from carbon zero. There is also a geographic dependency, or should be.

Nuclear power puts out more CO2 than solar or wind according to Nature (hydro isn’t mentioned for some reason).

“carbon emissions ranged from 1.4 grammes of carbon dioxide equivalent per kilowatt-hour (gCO2e/kWh) of electricity produced up to 288 gCO2e/kWh. Sovacool believes the mean of 66 gCO2e/kWh to be a reasonable approximation.”

https://www.nature.com/articles/climate.2008.99

Solar + wind with storage and a grid. The parts are all there, and it's cheaper than nuclear today.

No, the storage part is not there. Hydroelectric storage is expensive, takes a long time to build, and is geographically dependent to boot. Only ~5 minutes of global electricity storage can be provided with batteries using all known lithium deposits. Only 19 minutes worth of storage is available with all the lithium we can mine with today's equipment [1].

This is why plans for a solar and wind grid assume that some silver bullet is going to provide dirt-cheap and nigh-infinitely scalable storage.

1. https://dercuano.github.io/notes/lithium-supplies.html

These are not particularly relevant or helpful comparisons for knowing whether lithium ion is ready to deploy now (it is), or whether storage will be achievable with lithium ion and other chemistries (it will).

This is only looking at currently known reserves, a number which has doubled in only a few years. It also compares it to total energy consumption, a meaningless comparison for the coming decades.

Further, the same industrial capacity for lithium ion batteries also works for sodium chemistries. We have only focused on lithium because the primary applications are in mobile things at the moment: cars and mobile devices, where the weight advantage of lithium is important.

For grid storage, weight and specific energy are not important, and sodium chemistries will be ideal. There are also entire classes of flow chemistries that are in their infancy.

But what is mature and cost effective is lithium ion storage. The only place where we have open data about the feelings of investors, the PJM and ERCOT interconnection queues, storage is being deployed in GW comparable to new natural gas GW. This number alone, the GW and not the GWh, tells us that investors think this new tech is ready and deplorable. And it is falling in cost exponentially. Other battery tech is following and dropping in cost too, but lithium ion is benefitting from having existing markets that can fund massive learning.

> This is only looking at currently known reserves, a number which has doubled in only a few years.

False. It is estimating at the total amount of accessible lithium, not just the known reserves.

> For grid storage, weight and specific energy are not important, and sodium chemistries will be ideal. There are also entire classes of flow chemistries that are in their infancy.

Feel free to cite this as an option once sodium batteries actually become available at scale. Until then this amounts to, "hope some future solution solves storage."

> But what is mature and cost effective is lithium ion storage. The only place where we have open data about the feelings of investors, the PJM and ERCOT interconnection queues, storage is being deployed in GW comparable to new natural gas GW.

This is not even remotely true. We don't even have 1 GWh of battery storage [1]. Sure, we're not deploying "new" natural gas because energy demand is decreasing and we already have existing natural gas plants. But the point is that

> And it is falling in cost exponentially. Other battery tech is following and dropping in cost too, but lithium ion is benefitting from having existing markets that can fund massive learning.

Cost is a function of supply and demand. If you actually try to use lithium ion batteries for grid storage, this will create massive demand and thus increase cost. Again, there is insufficient accessible lithium to provide even half an hour of energy storage.

1. http://css.umich.edu/factsheets/us-grid-energy-storage-facts...

False on all counts.

The GitHub estimate is only using known resources and reserves, a number which goes up every year as we discover more. It is not an estimate of total accessible lithium. Lithium resources, the type where we get most of our lithium, increased from 40M tons to 80M tons from 2016 to 2020 estimates, and will continue to increase:

https://en.wikipedia.org/wiki/Lithium#Reserves

> This is not even remotely true. We don't even have 1 GWh of battery storage [1].

I don't know where that number comes from on that page, but it's wrong. More than 2GWh were connected to the US grid in Q4 2020 alone:

https://pvbuzz.com/woodmac-new-battery-storage-systems-q4-20...

And even if your number were right, it doesn't address the core point that battery storage deployment is growing at an absolutely incredible pace. In cost-competitive grids, it's replacing natural gas:

https://rmi.org/clean-energy-is-canceling-gas-plants/

> Cost is a function of supply and demand

This is just bad economics. These all affect each other. As production costs fall for lithium ion batteries, demand is growing, as shown by that RMI document. The cost of batteries is not falling because the demand is falling, the cost of lithium ion battery is primarily determined by manufacturing costs at the moment. The input costs of lithium is not going up because there's not enough lithium. And if supply of lithium does get constrained in the future, then there are alternative chemistries that are not supply limited.

> The GitHub estimate is only using known resources and reserves, a number which goes up every year as we discover more. It is not an estimate of total accessible lithium.

Yes, it is. 5 minutes is the amount provided by known reserves. 19 minutes is what can be provided with all accessible lithium. This is known reserves, plus the amount we expect to find later.

> I don't know where that number comes from on that page, but it's wrong. More than 2GWh were connected to the US grid in Q4 2020 alone:

Which amounts to a whopping... 14 seconds worth of energy storage.

> And even if your number were right, it doesn't address the core point that battery storage deployment is growing at an absolutely incredible pace. In cost-competitive grids, it's replacing natural gas:

17 GW of natural gas was constructed in Texas alone. In fact, not even all of Texas, just the part serviced by ERCOT. Your claim "storage is being deployed in GW comparable to new natural gas GW" is not even remotely true, and your own sources prove it.

> This is just bad economics. These all affect each other. As production costs fall for lithium ion batteries, demand is growing, as shown by that RMI document. The cost of batteries is not falling because the demand is falling, the cost of lithium ion battery is primarily determined by manufacturing costs at the moment. The input costs of lithium is not going up because there's not enough lithium. And if supply of lithium does get constrained in the future, then there are alternative chemistries that are not supply limited.

The assumption that the price of lithium won't go up if we try to use it for grid storage is bad economics. Let me put the staggering mismatch between battery supply and storage demand in perspective:

* The US alone uses 500 GWh of electricity each hour. The world uses 2.5 TWh of electricity every hour. * The entire world produces ~300 GWh of lithium ion batteries annually [1].

If we actually tried to provision one hour's worth of electricity storage the price of batteries would skyrocket, because there isn't enough supply to meet demand. We could provision one hour's worth of storage even if we bought every single lithium ion battery produced anywhere in the world for a whole year.

And this issue is going to become even worse as we switch from fossil fuels to electricity for heating, transportation, industrial chemical production, and so forth.

1. https://cleantechnica.com/2019/04/14/global-lithium-ion-batt...

Nuclear power plants are thermal power plants and that means they need cooling. The power density of nuclear power plants is so high that most of them can't be placed near rivers because rivers have a variable flow rate.

If the flow rate is too low you risk killing aquatic life in the river ecosystem so instead the nuclear plant is turned off. You can avoid this by placing the nuclear power plant near the ocean. That's what the Japanese did with the Fukushima power plant even though it's a tsunami prone area.

What gives you the idea that nuclear power plants can't be placed near rivers? Almost all that aren't on the coast are near rivers.

And they don't need to use potable water. The Palo Verde plant uses wastewater.

Because humans need water to survive, all population centers are built with access to water. Thus, cooling is available pretty much anywhere one would want to build a nuclear plant.

So you can build nuclear in a tsunami zone? in a seismic zone? in an area without cooling?

So you can mine and enrich uranium without carbon?

Nuclear does none of the things you fantasize it to do really.

> So you can build nuclear in a tsunami zone? in a seismic zone? in an area without cooling?

Yes, you harden the structure against tsunamis and earthquakes. That's part of why nuclear plants are so expensive.

Atmospheric cooling can indeed be done anywhere. It's typically easier and more efficient to use water cooling. And humans need water to survive, and thus population centers are built near sources of water, water cooling is almost always an option. Also nuclear plants can be cooled with seawater.

This is in stark contrast to hydroelectricity which needs both a river and a valley to be viable. Geothermal power needs magma near enough to the surface to heat water into steam.

> So you can mine and enrich uranium without carbon?

I don't see why not. Use electricity produced by nuclear plants to drive centrifuges. Also use said electricity to power mining equipment.

And you didn't answer my question: What other carbon-free sources provide energy 24/7, besides ones that need very specific geography like hydroelectricity and geothermal power?

> What other carbon-free sources provide energy 24/7

Wind + solar + biofuels + waste + batteries.

Batteries are primarily for peak usage, and it could be car batteries (V2G).

Biofuels are primary for seasonal usage (e.g. winter).

Nuclear is too expensive if you take into account the risks, which are currently externalized.

There is not enough accessible lithium to provide nearly enough storage [1]. 5 minute with known deposits, and 19 minutes estimated to be accessible with current mining techniques.

Biofuels are low energy density, and don't provide nearly enough power. Not to mention they aren't carbon-free. Burning biofuels releases carbon into the atmosphere that would otherwise be trapped.

1. https://dercuano.github.io/notes/lithium-supplies.html

Your source seems to be low by about 2 orders of magnitude on the energy density of lithium. They assume ~100% of a battery is made of lithium. There are only 200-300g of lithium metal per kwh in a lithium ion battery[0,1], or 12-18MJ per kg.

[0] http://www.meridian-int-res.com/Projects/How_Much_Lithium_Pe... [1] https://www.researchgate.net/deref/http%3A%2F%2Fgreet.es.anl... page 10

Battery: it doesn't have to be lithium (even thought, currently all planned ones use lithium-ion). Sodium-sulphur would be an option as well.

Biofuels are low energy density: this isn't about aviation or transportation, so that's not a concern at all.

Biofuels don't provide enough power: citation needed (are you moving the goalpost again?) - note that most energy will come from wind and the sun, so there is relatively little need for biofuels.

Burning biofuels releases carbon into the atmosphere that would otherwise be trapped: No, it would be released anyway (well, unless if you burry it really deep).

The problem with nuclear power is cost, due to high risks. And even then, the insurance (which is really expensive for nuclear plants) doesn't cover all the risks. The biggest risk is externalized: if e.g. a power plant in Switzerland would blow up, almost the whole country would be become un-inhabitable. And there is no insurance company paying for that.

> Battery: it doesn't have to be lithium (even thought, currently all planned ones use lithium-ion). Sodium-sulphur would be an option as well.

Right: we assume some other form of energy that has yet to be commercialized will provide cheap storage. Get back to me when this solution actually demonstrates feasibility.

> Biofuels are low energy density: this isn't about aviation or transportation, so that's not a concern at all. Biofuels don't provide enough power: citation needed (are you moving the goalpost again?) - note that most energy will come from wind and the sun, so there is relatively little need for biofuels.

Biomass provides 1MWh per ton of dry wood [1]. On average, forests have 38 tons per acre [2]. The US consumes 11.5TWh of electricity daily, so this works out to 319,444 acres per day. The US has ~750 million acres of forest. So we have 2,343 days worth of biomass energy. Or about 6 years.

Sure, forests grow, but they take longer than 6 years to grow. Also the figure of energy was in raw BTUs, so the actual electricity generated is only about ~50% of that.

> Burning biofuels releases carbon into the atmosphere that would otherwise be trapped: No, it would be released anyway (well, unless if you burry it really deep).

It would be trapped in the form of trees and vegetation. If burning biofuels doesn't release carbon into the atmosphere why are people concerned about deforestation?

> The problem with nuclear power is cost, due to high risks. And even then, the insurance (which is really expensive for nuclear plants) doesn't cover all the risks. The biggest risk is externalized: if e.g. a power plant in Switzerland would blow up, almost the whole country would be become un-inhabitable. And there is no insurance company paying for that.

This is not even remotely true. The plants in Switzerland have secondary containment. Even Chernobyl, which had no secondary containment, created an exclusion zone of 40x40km. "Almost the whole country would become un-inhabitable" is laughable. It really just demonstrates that aversion to nuclear is not based on rational thinking.

1. https://www.nacdnet.org/wp-content/uploads/2016/06/AppendixA...

2. https://www.nrs.fs.fed.us/fia/maps/nfr/descr/xlivebiohw.asp

> we assume some other form of energy that has yet to be commercialized

Both sodium-sulphur and lithium-ion are commercialized and widely used already (currently pumped storage is a lot more widely used, but it's not possible everywhere). [1]

Biofuels: as I wrote, it is only needed to fill the gaps [3], e.g. in winter, not to power 100%. It is already widely used, for example in Europe [2]. And it's not wood (CO2 is trapped in wood for some time, but not in vegetation). This doesn't displace forests.

> The plants in Switzerland have secondary containment.

So did Fukushima. There were many problems with nuclear plants in Switzerland, e.g. [4]. There is no 100% safety. In Switzerland, most people live in cities... sure, you could still live in the mountains, right.

> It really just demonstrates that aversion to nuclear is not based on rational thinking.

Actually, it is based on rational thinking. As the catastrophic events in Fukushima and Chernobyl, and the near catastrophes elsewhere have shown, nuclear power is dangerous. The population has to bear that risk. The companies would just get bankrupt. The insurance would only cover a small part of the costs.

1. https://en.wikipedia.org/wiki/Battery_storage_power_station 2. https://www.iea.org/data-and-statistics/?country=UK&fuel=Ene... 3. https://www.iea.org/articles/how-biogas-can-support-intermit... 4. https://en.wikipedia.org/wiki/M%C3%BChleberg_Nuclear_Power_P...

Biomass generates ~10% of the electricity from one country in Europe. Biomass is useful in countries like Brazil where extensive farmland means biodiesel is a viable automobile fuel. But for grid generation, the watts per acre is insufficient.

Globally, biomass is used for 0.7% of total energy demand [1]. Almost all of it for fuel, it doesn't even make it on the chart for electricity generation.

> So did Fukushima. There were many problems with nuclear plants in Switzerland, e.g. [4]. There is no 100% safety. In Switzerland, most people live in cities... sure, you could still live in the mountains, right.

And the secondary containment in Fukushima meant that most of the radiation was contained. Fukushima is already being resettled. You harbor this skewed perceptions where nuclear catastrophes render massive swathes of the earth uninhabitable, "almost the whole country [Switzerland] would be become un-inhabitable". No it would not. Even an uncontained meltdown resulted in a 40x40km exclusion zone. An a contained one is much less drastic. Three Mile Island didn't even result in any permanent exclusion zone.

Easy, there’s many: 1) Solar + transmission lines. It’s always sunny somewhere

2) wind + transmission lines, always windy somewhere

3) use some energy produced by hydro to manufacture some concrete river beds and reservoirs

4) use some of the energy produced by 1-3 to dig real deep for geothermal everywhere

5) Ocean thermal energy conversion

Don’t get me wrong , I’m not anti nuclear , I’m a huge fan of the big reactor in the sky it produces all we need with perfect reliability there’s no reason to do something as dumb as trying to build terrestrial reactors

> 1) Solar + transmission lines. It’s always sunny somewhere

> 2) wind + transmission lines, always windy somewhere

These don't produce power consistently. That's why one would need to build redundancy. Also it's not always sunny somewhere, unless you build transcontinental transmission lines. And even then, there's a period of time where most sunlight is hitting the pacific ocean.

> 4) use some of the energy produced by 1-3 to dig real deep for geothermal everywhere

> 5) Ocean thermal energy conversion

Both of these are geographically dependent. Might as well has just said hydroelectricity.

> Solar + transmission lines. It’s always sunny somewhere

Unless it is sunny 24/7 in a given country or even group of coutnries (e.g. the EU) this is not viable.

Countries will not give up their energy security and put themselves at the mercy of the other side of the planet (where it is sunny) plus whoeever might want to damage those transmission lines and cripple a country. It is already an issue with oil and gas.

> So you can build nuclear in a tsunami zone? in a seismic zone?

Why not? It worked at Onagawa: https://www.sciencedirect.com/science/article/pii/S259012302...

It’s also failed. Why not just avoid it? That’s the approach taken according to your link.

The solution to the problems faced at Onagawa were to decommission the plant, and this process would take longer than the duration for which the plant actually ran.

“the 2011 events strongly influenced the decision to decommission the Onagawa Unit 1 early, brought to attention the length of the decommissioning process (which will surpass the operation stage)”

The decision to decommission the plant is political, not technical.

Onagawa was closer to the epicenter than fukushima and suffered no ill effects. It can be done, the main different between Onagawa and Fukushima is that they were owned by different companies and one company took safety seriously.

For more reading, see: https://thebulletin.org/2014/03/onagawa-the-japanese-nuclear...

and IAEA report: https://www.iaea.org/newscenter/focus/actionplan/reports/ona... which the article is based on.

The US stockpile of HEU would be depleted a lot faster this way, but enrichment could start again. I don't see major downsides to this proposal. Thanks for providing a specific and plausible idea!

Thats a great idea. I'm trying and failing to find gotchas.

There is a potential gotcha: proliferation potential. The naval reactors use highly enriched uranium; if it falls in the wrong hands, you can end up with someone being able to build a bomb. That's why I said such a program needs to be run by the Department of Energy, the same department that has to maintain the nukes. I don't have a personal objection to this, but a lot of people would be unhappy with an essentially military program to be established for a problem that is not military in nature.

Hopefully the DOE will continue to be led by a person who didn't previously campaign on disbanding it!

Reply on news.ycombinator.com