Hacker News Reader

Comment by acidburnNSA | Hacker News Reader

I spent a month earlier this year writing up the history of nuclear power development in the USA [1]. If you like this it might be right up your alley. Things that actually were built include a truck-mounted nuclear reactor (ML-1), nuclear reactors in space (SNAP-10A), a nuclear-powered merchant ship (NS Savannah), an ice-base powered by a nuclear reactor under Greenland (Camp Century), dozens of highly exotic power reactor prototypes, and (of course) the modern nuclear power plants.

[1] https://whatisnuclear.com/reactor_history.html

I didn't cover tanks but did touch on nuclear-powered aircraft, which we spent about $1B (1950s) dollars on before ICBM progress obsoleted the concept (for urgent military needs at least). The work led to the molten salt reactor concept, which is much loved today in the internet nuclear circles with civilian power production in mind.

View on HN

> "[...] before ICBM progress obsoleted the concept."

I'm afraid that anti-ICBM defenses and the availability of reliable auto-pilot technology made the concept of a nuclear powered planes relevant again: https://en.wikipedia.org/wiki/9M730_Burevestnik

First of all that's a very important article people should be aware of: around the ABM treaty, nuclear _powered_ cruise missiles, and maybe some deaths developing that tech. This is just last year, Putin driven.

Second of all, it's not clear why it needs to be nuclear powered. Our 1970's subsonic cruise missiles are terrain-following for 2500km at 30m AGL. You can throw them from trucks or ships or planes and go well into someone's territory. A fair assumption is that everyones' are better now in all axes. The point is you don't need the price/risk of unlimited range, unless your goal is to terrorize and poison everything on the way to your target.

We had some nuclear scramjets in the Project Pluto era. I think one nice thing about them in the military sense is that they can go REALLY fast for a pretty long time.

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

They were doomed. The issue was heat. At supersonic speeds air friction heats the missile. Without fuel being burnt there is no way to dump the heat off the missile. Regardless of heat from the engine, after a few hours air friction would see the entire thing glowing red, the "thermal thicket".

There's enough energy in kerosene though that regular cruise missiles are plenty dangerous...

Depends entirely on what you want to do. If Russia wanted to hit a missile silo in Nebraska with a [supersonic] cruise missile they'd need to sneak an Oscar up the Mississippi (which is implausible if not impossible.) Granits have less than a thousand kilometers range, so realistically the only way Russia could strike at the middle of America is with ballistic missiles. ABM tech calls even that into question, hence their recent 'superweapon' projects like Skyfall and Poseidon.

Project Pluto's nuclear ramjet bomber/missile sounds like it came from the imagination of a science fiction novelist or a screenwriter:

a locomotive-size missile that would travel at near-treetop level at three times the speed of sound, tossing out hydrogen bombs as it roared overhead. Pluto’s designers calculated that its shock wave alone might kill people on the ground. Then there was the problem of fallout. In addition to gamma and neutron radiation from the unshielded reactor, Pluto’s nuclear ramjet would spew fission fragments out in its exhaust as it flew by.

Not as efficient at killing as ICBMs, I guess, but a much more terrifying concept.

What's fascinating about nuclear reactors is that you can do them in so many ways that are very different.

If you think of combustion engines, there's only a few viable cycles or configurations: otto, diesel, brayton. Some small variations on these with supercharging and valve timing etc. Wankel engine as one different configuration.

But for nuclear power there's hundreds or thousands potential ways to produce energy. Different nuclei to start with, different fuel cycle, breeding. Neutron speeds, moderated and unmoderated. Fuel can be solid, dissolved, molten, gas. As can moderator. Pebble bed.

As far as I understand, even traditional reactor rod scheduling is a hard computational problem that a lot of resources is used on even by today's massive computing standards.

There's so much "blue ocean" left in fission reactors.

I've thought a lot about this as well. In fact, I made a random reactor concept generator [1] that is a Python script making combinations. There are 1.3 million different options possible and that is still not all of them.

[1] https://whatisnuclear.com/random

More info including the full list here [2].

[2] https://whatisnuclear.com/random_info.html

Hah, that's awesome!

As they say, old airplanes were designed entirely by hand. At some point computers could do some calculations. In the nineties, the computers could analyze CAD designs. Later, the computer could iterate through designs and do dimensioning. Maybe currently, generative design of parts. In the future they will be able to do configuration.

>> I didn't cover tanks but did touch on nuclear-powered aircraft

IIRC, this idea resurfaced in the early 2000s for long-lived drones, but were terminated due to the concern that there may be radioactive material/debris if they were shot down. Is that true?

Not sure about the resurfacing, but the concern is legitimate. When you're splitting atoms to get heat, the leftover smaller atoms have what's called nuclear afterglow. This afterglow is most hazardous the moment the reactor stops and slowly becomes less hazardous with time. In any case, a nontrivial fission power reactor vaporizing in atmo would be akin to atmospheric nuclear weapons tests from a radiological hazard POV. We've done stuff like that before in remote areas, but no one thinks it's a good idea.

To do it again, we'd have to basically convince ourselves that no matter what shot the plane down, the radiation containment system would survive the hit and the crash.

At #the-sodium-deuterium-reactor-program you quote this:

> The principal problem in the SDR is a psychological one. Many people have a feeling that sodium and water react explosively; they do not react explosively. Hydrogen and oxygen do react explosively. The pyrotechics that you may have seen on sodium combining with H2O is principally one in which hydrogen has reacted with oxygen in the air. The reaction rate is of the order of microseconds; it is truly an explosive reaction. Sodium and water can be handled quite safely and quite easily, as I will show you later.

Well, so, sodium and water do not explode because of dissociated hydrogen reacting with oxygen so much as through coulombic explosion:

https://www.youtube.com/watch?v=8PEVmflpUCo https://cen.acs.org/articles/93/i5/Sodium-Potassium-Really-E.... https://www.nature.com/articles/nchem.2161 https://en.wikipedia.org/wiki/Coulomb_explosion https://www.bbc.com/news/science-environment-30983141

What happens is that as sodium loses electrons to the water it becomes more and more positively charged, and this causes tendrils to grow as the sodium's surface pushes on itself to reduce the electric potential, but this increases the surface area in contact with water, which speeds up the rate at which the sodium loses electrons, which speeds up the rate at which tendrils form, and so on. This reaction is spectacularly fast. It does also produce hydrogen, yes, which could react with oxygen, but even without that the tendril formation is explosive.

Plain sodium will not necessarily explode in water because a) it has oxides on its surface that reduce the rate of reaction with water, and b) hydrogen formed at the surface does the same. But with the right setup you can get an explosion and it really is incredibly fast. In one of the videos on the topic clean sodium (no oxide) is dropped through a noble gas atmosphere into water, and what they get is an explosion.

Apparently this was only properly understood very recently, around 2014.

So yeah, knowing nothing more about SDR, the thought freaks me out! :)

Reply on news.ycombinator.com