Hacker News Reader

Hacker News

Comment by acidburnNSA | Hacker News Reader

Parent Full thread

acidburnNSA

Author here, AMA.

This post got me introduced to Ira Flatow, and we discussed it on NPR's Science Friday a while back.

https://www.sciencefriday.com/segments/floating-nuclear-powe...

I mentioned this in a thread yesterday when someone asked me what my favorite advanced reactor was.

https://news.ycombinator.com/item?id=31463673

The huge crane being sold to China in 1990 is the sad ending of the story:

https://www.tampabay.com/archive/1990/11/05/huge-crane-sold-...

View on HN

Interesting article, but the relative cost of nuclear vs solar/wind/storage is why utility managers around the world are not that interesting in expanding their nuclear footprint.

Often you'll see people argue that excessive regulation is what is driving costs up, but without a very high level of maintenance, nuclear power systems will inevitably have an unacceptable rate of failure. These safety margins need to be higher than those for units like oil refineries, coal-fired power plants etc., as the consequences of failure are so long-lasting (fires at oil refineries are relatively common, for example, but they don't result in 100-year no-go radioactive exclusion zones).

Therefore, to make nuclear accidents highly unlikely, maintenance costs amd engineering standards will inevitably be much greater than in any other power supply system. This means nuclear will always cost significantly more than any other system - and this also applies to storage of too-hot-to-handle 'spent' nuclear fuel, as well as to the costs of eventual decommission. In addition, the nuclear fuel supply chain (high grade uranium ore -> fuel rods) is fairly complex, and subject to geopolitical and economic price spikes.

All in all why would a utility manager these days looking at all the options for non-fossil-fuel power go with nuclear?

On cost, the simple answer is that if fossil and biofuel had to pay for the air pollution health effects they cause (4.2 M deaths/year [1]) and the climate change damage they cause, then every utility would switch to nuclear for baseload immediately. The markets are incorrect in not capturing these costs. Nuclear is dramatically cheaper than fossil and biofuel on this basis alone in a real market.

And of course the advantage of nuclear over wind and solar is that nuclear has a 92% capacity factor with fairly regular well-planned outages.

The wind and solar costs you've seen, while stellar, do not include the costs of the backup generators and batteries, additional transmission to remote site, overcapacity required to charge the batteries at scale, etc etc. Nuclear is certainly on par with them in a fully decarbonized system. In any case, building nuclear alongside wind and solar makes the whole system cheaper [2].

Regarding 100 year no-go zones, the hottest areas I see around Fukushima today measure about 30 mSv/yr, which is about 2x higher than Denver's natural radiation. Japan is lifting even more evacuation orders as we speak [3]. If they gave people the proper information about the health risks (e.g. that you need more than 100 mSv in a year before the smallest measurable cancer increase can be seen) then people would have all moved back 10 years ago.

[1] https://www.who.int/health-topics/air-pollution

[2] https://doi.org/10.1016/j.joule.2018.08.006

[3] https://sp.m.jiji.com/english/show/19741

I think we want to compare nuclear and solar/wind/storage, where the base goal is to have a steady power generation rate over time (adjustable to demand within certain limits). This tends to produce a closer cost outcome for nuclear and s/w/s as storage costs (batteries, inertial storage, pumped hydro, etc.) are probably going to at least double or triple the installed costs of the s/w/s power system.

For example, if you want to provide relatively steady 1GW of power 24-7, 12 m/yr, this can be done with a nuclear plant running at its optimal rate (*optimal for fuel rod lifetime and steady burn rate). With a solar/wind system, you'd likely need enough PV panels to produce say 2GW at noon, with 1GW going to grid demand and another GW going to storage, to meet demand at midnight etc. Balancing seasonal demand in subpolar regions would require a lot more power and storage as well, so nuclear may have an advantage there.

Clearly both systems are possible, but both have supply chain issues (uranium ore and fuel elements and reactor construction, vs. battery components and PV panels and wind turbines and storage infrastructure). However, I'd guess long-term maintenance costs are going to be much lower with solar/wind/storage systems (with much lower fueling/waste storage issues and liabilities, which would really just be battery maintenance and replacement costs).

> the costs of the backup generators and batteries

Just to put an extra-fine point on this...the backup generators are typically fossil fuel fired, so the unpaid externalities apply to those as well.

What am I misunderstanding/miscalculating? The US uses ~4000 billion kilowatt-hours [1] and these $1 Billion reactors produce 8GW. 8GW is 8 million kilowatt-hours. 4,000,000 MKWh / 8 MKWh = 500,000 reactors. To provide 20% of electricity we need 100,000 of these reactors at a cost of $100 Trillion? A quick google says:

> One hundred trillion dollars. That’s the total net worth of all Americans in 2017*.

[1] In 2021, net generation of electricity from utility-scale generators in the United States was about 4,116 billion kilowatthours (kWh) (or about 4.12 trillion kWh). https://www.eia.gov/energyexplained/electricity/electricity-...

Im on mobile so will check you math later but our 100 GW nuclear fleet makes 20% of our electricity. So you need 500 GW-class reactors to power all electricity in the USA, not 500,000.

UPDATE: So yeah power is rate of energy delivery, measured in Watts or Joules per second. Energy itself is measured in Joules or (equivalently) Watt-seconds. Units that are like kW-hour have power × time, so those are energy units. So if the USA uses 4116 billion kWh of energy in a year, that comes out to 469 GW of power averaged over the year. So you need about 500 GW-class nuclear reactors to generate the country's electricity.

Meanwhile electricity is only 40% of our total energy usage. The rest is transportation (almost all oil), building heat (natural gas), and industry (mostly natural gas). Those have to be decarbonized too, and if we switch them to electric alternatives we can expect our total electric demand to roughly double.

PS I always use GNU units for these kinds of unit conversions (`units "4116 billion kW*hour/1 year" "GW"`)

Thank you for sharing! What a great article. As I read it I formulated several questions only for them to be clearly resolved as I read further. This is so often not the case.

I would be interested in learning more. Is there one of the recent development projects that is more mature and likely to succeed than the others?

Does the science Friday episode have additional information not included in the writeup?

Thanks for saying so. I'd say the writeup is more informative than the Science Friday piece, honestly.

There are a few good modern reports about using his kind of idea for making ammonia and hydrogen, but nothing too serious is going on outside of Russia and China on this topic.

[1] https://www.lucidcatalyst.com/hydrogen-report

Most of the people talking about it in advanced reactor space are using very low readiness reactors rather than PWRs, which led me to make this extremely fringe starwars prequel meme: [2].

[2] https://i.imgur.com/bDaVjRY.png

The Koreans have the PWRs and the ship industry in place, they'd be the best to make this happen again. But so far they've just done a few small MOUs here and there. Again, nothing too serious is happening from what I can tell. I'd be happy to be wrong.

Thanks again!

A small notes, with the premise that I favor nuclear in general, not only future hoped fusion if we will ever get it: we use electricity mostly on-shore, also deep inside land. Surely a water-born NPP can travel so can theoretically adapt to different needs, theoretically only since it still demand a land-based distribution network, surely a water-born NPP can be sunk easily in case of an emergency so it's safer BUT while I agree that due to climate change and social changes derived we need flexibility that means going via air and water instead of land since infra on land can't really be built and changed in quick enough time at a low enough costs with a low enough needs of natural resources and yes we can easily move stuff on the sea that on land, BUT we still live on land...

Also for defense needs, where possible, we want critical industries, those who most need nuclear, deep inside land to better protect them to direct military attacks...

Such floating power plants can be useful for a colonization where land can't be used like in the now melting Arctic areas where we can't safely built on soil due to it's continuous changes and we need to quickly develop infra there in a race for natural resources, they can be useful in some very seismic areas or for emergency usage in remote areas IF small enough and able to travel quickly enough (witch is unlikely) avoiding being too near the coastline to be safe in maritime terms but I fail to see other realistic usages...

It seems a bit like the '70s nuclear-powered aircraft experiments, formally nice, but only in engineering terms, and they fails as expected...

My understanding is that naval reactors mostly (if not all?) require near-weapons grade uranium. Though I assume this is to allow a compact design. How highly enriched would it need to be in this case? Any more than current land-based civil designs?

Reply on news.ycombinator.com