3 ms·

No One Knows If Decades-Old Nukes Would Actually Work

16 points by fortran77 4y ago

Ancalagon 4y ago
Well let’s not test them to find out
techdragon 4y ago
There is a kernel of truth to this… and it’s simple. We used to actually blow them up and really test if they still worked. Now we don’t. So we simulate with supercomputers, we perform partial tests on sub components, we validate everything as best we can… except for the part where we can’t really test if the whole combination works together at the critical moment. (pun intended) In particular there’s a reasonably big thorn in the all sealed up and “ready to go” thermonuclear devices, the tritium in the devices has a pretty short half life and needs regular replacement… same but on a longer schedule for neutron sources. While a gun type fission device is incredibly simple and has a shelf life of probably well over a hundred years… we don’t use them… I don’t think anyone still actively fields that type in their arsenal. So it’s all complex implosion type devices with neutron primers, and even more complex thermonuclear implosion devices with tampers and fillers (Remember when the USA forgot how to make FOGBANK) and tritium and lots of circuits and electronics that while designed specifically for this environment… are sitting in proximity of radioactive material (just because the aren’t emitting radiation at a level that makes them unsafe for military personnel to handle doesn’t mean they emit none and the electronics are on the inside behind layers of material blocking some of the radiation from people) It’s not really a situation that should be surprising. What’s surprising is that we have managed to see the nuclear test ban treaty hold up as well as it has. I suppose it comes down to the calculated game theory style gamble that you can be sure enough of your own nukes working with billions of extra dollars on things like supercomputers and exhaustive component lifecycle testing programs and lots of hard work but it’s obvious that this also makes it harder for your nuclear armed adversaries to be sure theirs work. Cold War game theory type realpolitik. Edit: having written this all off the top of my head from the (kinda “no duh”) headline I’m pretty happy with how much I covered the actual topic vs the “but Russia and politics and maybe he might use nukes” subtext that is pretty much the rest of the article and the part I didn’t talk about. They even mentioned FOGBANK by name.
credit_guy 4y ago
This is so naive. Yes, there were no nuclear tests with bombs that go "boom" since the early 90's (except for the newest member of the nuclear club, North Korea). But you can test your nukes so you can have an extremely high degree of confidence that they'll work as designed when the time comes. How can one do that? We'll get to that, but first: if you couldn't do that, the nuclear powers would not have stopped testing. Why would they ever do that? Nobody can coerce a nuclear power to stop doing nuclear tests, just look at North Korea. With that out of the way, here's how you test. The easiest thing is to perform "hydronuclear tests" [1]. In such a test, just like in an actual nuke, you wrap the plutonium pit and detonate a shell of high explosives around it and create an implosion. The "UAT" implosion is slightly less powerful than the "production" one. Either critical mass is not achieved at all (i.e. sustained chain reaction) or if it's achieved, it's only infinitesimally supercritical, so the chain reaction goes on for only a few generations (in an actual explosion there are about 40 generations, with 99% of the energy coming from the last 2 or 3 generations). If the explosion is correctly tuned, you end up melting the plutonium pit (hence the name "hydro") but not transforming it into hot plasma. Hydronuclear tests are virtually undetectable. CIA thinks (with what they claim is a high degree of confidence) that Russia carries out hydronuclear tests. With a hydronuclear test you can determine how well you know your nukes: if you predict that with this amount of detonation, plutonium pit will melt in proportion of 10%, and it happens just as you say, then you know what you are doing. If 80% or 0.8% melts instead, then you have work to do. You tweak all the knobs that you can tweak, check once, twice, three times everything that can be checked, and eventually you get it. Or not. If you don't get it, you restart full scale tests. But considering that no nuclear power have restarted the tests, they probably got it. The US is quite advanced, it does not need to perform hydronuclear tests. It uses high power lasers, and who knows what. But the US is spending $15 BN on the nuclear stockpile stewardship ([2], p182). It's quite unlikely this money just goes down the drain. [1] https://www.globalsecurity.org/wmd/intro/hydronuclear.htm https://www.globalsecurity.org/wmd/intro/hydronuclear.htm [2] https://www.energy.gov/sites/default/files/2022-03/FY%202022%20SSMP%20March%202022.pdf https://www.energy.gov/sites/default/files/2022-03/FY%202022...
- sublinear 4y ago
  > The "UAT" implosion is slightly less powerful than the "production" one. funny
hulitu 4y ago
> No One Knows If Decades-Old Nukes Would Actually Work There is only one way to find out.

16 points by fortran77 4y ago

Well let’s not test them to find out

There is a kernel of truth to this… and it’s simple. We used to actually blow them up and really test if they still worked. Now we don’t. So we simulate with supercomputers, we perform partial tests on sub components, we validate everything as best we can… except for the part where we can’t really test if the whole combination works together at the critical moment. (pun intended) In particular there’s a reasonably big thorn in the all sealed up and “ready to go” thermonuclear devices, the tritium in the devices has a pretty short half life and needs regular replacement… same but on a longer schedule for neutron sources. While a gun type fission device is incredibly simple and has a shelf life of probably well over a hundred years… we don’t use them… I don’t think anyone still actively fields that type in their arsenal. So it’s all complex implosion type devices with neutron primers, and even more complex thermonuclear implosion devices with tampers and fillers (Remember when the USA forgot how to make FOGBANK) and tritium and lots of circuits and electronics that while designed specifically for this environment… are sitting in proximity of radioactive material (just because the aren’t emitting radiation at a level that makes them unsafe for military personnel to handle doesn’t mean they emit none and the electronics are on the inside behind layers of material blocking some of the radiation from people) It’s not really a situation that should be surprising. What’s surprising is that we have managed to see the nuclear test ban treaty hold up as well as it has. I suppose it comes down to the calculated game theory style gamble that you can be sure enough of your own nukes working with billions of extra dollars on things like supercomputers and exhaustive component lifecycle testing programs and lots of hard work but it’s obvious that this also makes it harder for your nuclear armed adversaries to be sure theirs work. Cold War game theory type realpolitik. Edit: having written this all off the top of my head from the (kinda “no duh”) headline I’m pretty happy with how much I covered the actual topic vs the “but Russia and politics and maybe he might use nukes” subtext that is pretty much the rest of the article and the part I didn’t talk about. They even mentioned FOGBANK by name.

credit_guy 4y ago

This is so naive. Yes, there were no nuclear tests with bombs that go "boom" since the early 90's (except for the newest member of the nuclear club, North Korea). But you can test your nukes so you can have an extremely high degree of confidence that they'll work as designed when the time comes. How can one do that? We'll get to that, but first: if you couldn't do that, the nuclear powers would not have stopped testing. Why would they ever do that? Nobody can coerce a nuclear power to stop doing nuclear tests, just look at North Korea. With that out of the way, here's how you test. The easiest thing is to perform "hydronuclear tests" [1]. In such a test, just like in an actual nuke, you wrap the plutonium pit and detonate a shell of high explosives around it and create an implosion. The "UAT" implosion is slightly less powerful than the "production" one. Either critical mass is not achieved at all (i.e. sustained chain reaction) or if it's achieved, it's only infinitesimally supercritical, so the chain reaction goes on for only a few generations (in an actual explosion there are about 40 generations, with 99% of the energy coming from the last 2 or 3 generations). If the explosion is correctly tuned, you end up melting the plutonium pit (hence the name "hydro") but not transforming it into hot plasma. Hydronuclear tests are virtually undetectable. CIA thinks (with what they claim is a high degree of confidence) that Russia carries out hydronuclear tests. With a hydronuclear test you can determine how well you know your nukes: if you predict that with this amount of detonation, plutonium pit will melt in proportion of 10%, and it happens just as you say, then you know what you are doing. If 80% or 0.8% melts instead, then you have work to do. You tweak all the knobs that you can tweak, check once, twice, three times everything that can be checked, and eventually you get it. Or not. If you don't get it, you restart full scale tests. But considering that no nuclear power have restarted the tests, they probably got it. The US is quite advanced, it does not need to perform hydronuclear tests. It uses high power lasers, and who knows what. But the US is spending $15 BN on the nuclear stockpile stewardship ([2], p182). It's quite unlikely this money just goes down the drain. [1] https://www.globalsecurity.org/wmd/intro/hydronuclear.htm https://www.globalsecurity.org/wmd/intro/hydronuclear.htm [2] https://www.energy.gov/sites/default/files/2022-03/FY%202022%20SSMP%20March%202022.pdf https://www.energy.gov/sites/default/files/2022-03/FY%202022...

sublinear 4y ago
> The "UAT" implosion is slightly less powerful than the "production" one. funny

hulitu 4y ago

> No One Knows If Decades-Old Nukes Would Actually Work There is only one way to find out.