[erentz]

I wondered about this myself. As another commenter pointed out naval plants use a more enriched uranium which IIRC allows them to run the reactors for 20+ years without needing to refuel.

One other way of thinking about it though is that this is evidence we can have a program to make specialized small modular reactors fairly economically. So we shouldn’t be so skeptical about proposals we do it for civilian use.

[jabl]

> this is evidence we can have a program to make specialized small modular reactors fairly economically.

It is? I would imagine that the advantages of nuclear power for submarines are so great that militaries would gladly pay 100 times more than would be acceptable for civilian power without blinking.

(I do think SMR's are promising for civilian applications, I just don't think that nuclear submarines are a good argument for the potential economic benefits.)

[erentz]

Well I mean, one of the new submarines cost something like $2 billion. The reactor can’t be more than a 20% of the total cost. So no it’s not anything like 100 times more. It’s probably practically economically viable for civilian power as is. It just needs necessary modifications and housing for land based use. And permission to use the higher enrichment uranium. The fact you don’t need to refuel it for 20 years might be a bonus because the military can build them, lock them up, ship them to site, they produce power for 20 years, then they ship them back for refueling or decommissioning.

[jabl]

> Well I mean, one of the new submarines cost something like $2 billion. The reactor can’t be more than a 20% of the total cost.

Wikipedia claims the latest generation US attack submarines come in at $2.8 billion each. If the reactor is 20%, and the entire propulsion system 30% (to make a comparison to a power plant), shaft power is 30 MW, we get $28000/kW. For comparison, civilian nuclear power plant capital cost is around $1500-$6000/kW, depending on where it's built, the cost of capital etc.

So yes, not 100 times more expensive than civilian power, but still way too expensive to be economically viable. (And to be pedantic, I didn't say it was 100 times more expensive either, just that the advantage of nuclear propulsion for submarines are so great that navies are prepared to pay whatever it costs (and cost minimization is certainly not such a high priority for sub reactors as for civilian ones))

> And permission to use the higher enrichment uranium.

No. Just no. Look, I'm a huge proponent of civilian nuclear power, but HEU for civilian use flies in the face of decades of non-proliferation work (see e.g. the US RERTR program which has been running since 1978). And for good reasons too; given access to high grade HEU, even a relatively unsophisticated actor (such as a decently funded terrorist group) could make a gun-type nuclear weapon. In contrast, an implosion device using Pu is significantly more challenging, and is probably out of reach for non-state actors.

In fact, there is pressure to direct the US Navy to switch to LEU (that is, less than 20% enrichment), precisely for non-proliferation reasons. For example, French submarines run on 7% enriched fuel, and they refuel every 10 years. So with 20% fuel, US submarines could probably make do with a single mid-life refueling.

See e.g. http://fissilematerials.org/library/rr15.pdf for a report detailing various efforts to get rid of non-weapon uses of HEU worldwide.

[erentz]

Actually I was spitballing 20% of $2b as an upper bound for the entire plant. Does 30% come from something? For the thought experiment I’ll take 20% of $2.8b though. So $560m.

Not sure how shaft power is what we are considering here as this would be converted to purely electricity generation. These plants are estimated 210MW thermal, say a 35% conversion factor that’s more in the ballpark of 70MW electric. So $560m for 70MWe that doesn’t need refueling for 20 years. So more like $8,000 /KW. For a system that can be largely self contained, shipped to site, and not need refueling for its 20 year life span.

We could replace over 200GW of fossil generation sources in this country with 3000 of these at a price of $1.68t. Amortized over that 20 years fuel life span. That’s crude ballpark $84 billion a year.

US electricity production at the moment (11am EST) is about 490GW. With 99GW coal and 166gW of gas.

So we could effectively turn the country’s electricity system fossil free for a cost of $84 billion per year.

That’s seems dirt cheap. It’s less than this years military budget increase. For that we clean our grid. The more I think about it the more it seems it’s something we could totally do without breaking a sweat if we actually took climate change seriously.

[jabl]

> Does 30% come from something?

No, I just added a bit on top to get an estimate for the entire power conversion system, assuming your estimate of 20% for the nuclear parts were in the right ballpark.

> Not sure how shaft power is what we are considering here as this would be converted to purely electricity generation

The power to turn the prop shaft would roughly be what's available for turning the generator if the thing is used to produce electricity rather than propulsion.

Sure, there's the "hotel load" to run the rest of the sub, but then again a generator isn't 100% efficient either so I think roughly 30 MW electrical output is about right.

Details are classified, but from publicly available sources my understanding is that naval reactors tend to work at lower pressure and temperature than civilian land based PWR's, so efficiency takes a (substantial) hit. So I do think my estimate of 30 MWe is closer to the mark than your 70.

> doesn’t need refueling for 20 years

So it's a life of ship core which is supposedly 33 years. But, that's 33 years of average use. AFAIU subs spend most of their time creeping along pretty slowly in order to stay silent, so on average they'll be using a small fraction of max power. Civilian power reactors otoh want to run at full power as much as possible to amortize capital costs.

Not to mention that HEU just won't fly for civilian use, as I tried to explain my previous message.

As for economics, yes I think that decarbonizing the electric grid with series produced reactors would, in the end, be pretty cheap. France did that some decades ago, for about $1500/kW in today's money. AFAICT that would require the government to handle it (like in France), both to decide on one reactor model (to get on the learning curve by series producing it) and to have access to cheap capital.

If one wanted to do it with SMR's rather than large reactors like France, I suggest e.g. Nuscale would be a better starting point than submarine reactors.

[squiggleblaz]

> One other way of thinking about it though is that this is evidence we can have a program to make specialized small modular reactors fairly economically. So we shouldn’t be so skeptical about proposals we do it for civilian use.

In Australia we occasionally have a discussion about our submarines. I think we have the largest and longest range conventionally powered subs. Should we get nukes? The argument against is we have no industry. (Although, I think, ex prime ministers are significantly more likely to argue for nuclear subs than current prime ministers. That means something.)

I guess they both play off each other. Life works best with nuclear subs and nuclear plants, or without either.

[erentz]

One option that was floated (can’t remember by who) is for Australia to lease new Virginia class submarines from the US. This would essentially mean the US would own the lifecycle aspects of the power plant. Don’t know if it’s a good idea or not.