5 ms·

Wow! I wish the thorium enthusiasts would all read this.

DennisP 7y ago
Small molten salt reactors were actually built. They weren't just paper reactors, even in Rickover's day. Now there are companies working on MSRs for commercial use. There's more computer simulation than running reactors so far, partly because governments are much more cautious now. But I'll also point out that Rickover in 1957 had no idea what computers in 2020 would be capable of. I think it's reasonable to suspect that we didn't find the best possible reactor design six decades ago.
- brandmeyer 7y ago
  The part that the computers help with - nuclear and thermodynamic physics ODEs - isn't the hard part. A couple of sharp grad students can do that with OpenFOAM and a modest workstation. The hard parts are the operations, maintenance, building for robustness, and enormous amounts of empirical field experience with the materials. Rickover and his contemporaries judged - rightly - that those issues would make MSR and metal-cooled reactors too finicky to be of practical military value.
  - perl4ever 7y ago
    ...but there actually was a liquid sodium reactor sub built in the 50s, from what I've heard.
    - brandmeyer 7y ago
      Only one. She was a contemporary of the Nautilus. It was their experience with that vessel that killed liquid metal coolant for the navy. It might help to understand something special about the nuclear navy in those days - prototypes in the lab don't count. Prototypes on land don't count either. The only prototypes that counted were ones that actually went to sea. All the way through the Skipjack generation, they were regularly doing crazy-eyed one-off experiments. Nautilus and Seawolf were both prototypes.
      - willvarfar 7y ago
        Agree. One Soviet 'November' class (K27) and all 7 'Alfa' class used liquid metal. It wasn't developed further or used after.
        DennisP 7y ago
        But Russia does have two sodium cooled reactors on their power grid right now.
  - DennisP 7y ago
    Maybe we shouldn't base our civilian energy structure on what has practical military value, even if it made sense as an expedient in 1957. Operations is hardly an advantage of LWRs over MSRs, especially if you stick with simple designs instead of LFTR (as almost all real projects are doing). Those also make maintenance pretty simple. Building for robustness is a lot easier when you don't need build lots of active safety systems or an oversize containment dome due to high-pressure coolant. And some designs use existing nuclear-rated materials.
- dredmorbius 7y ago
  And comining on sixty years later, still not decomissioned, because dealing with both chemically and radioactively volatile materials is a nightmare.
Animats 7y ago
Yes. Most of the problems with reactors have come from parts that are glossed over at the PowerPoint level. Many exotic reactor designs have been built, and most have had problems. Everything with complexity inside the radioactive area has been troublesome. The AVR pebble-bed reactor in Germany had a pebble jam, and still can't be decommissioned. Often the problems involve plumbing. Sodium reactors have leaks and sodium fires. Helium gas-cooled reactors have helium leak problems. Ft. St. Vrain was a really good idea which had both helium leak and corrosion problems. Three Mile Island was a valve problem. Fukishima was a failure of the cooling system after a tsunami. This is why some of the new designs are viewed skeptically. Molten salt reactors involve a chemical processing plant that works on radioactive materials. Radioactive chemical plants have always been troublesome. They tend to become toxic waste sites. (See Pantex, PUREX, etc.) Modular reactors with multiple reactors sharing the same cooling pool are at risk if anything leaks and contaminates the pool. Commercial reactors have to be built for a 40-50 year lifetime to pay off. Anything inside the radioactive zone is essentially unmaintainable for that period. Complexity inside thus just doesn't work. Which is why we're stuck with water and pressurized water. Simple mechanically and chemically.
- WalterBright 7y ago
  > Fukishima was a failure of the cooling system after a tsunami. The fixes to that design would be straightforward. For example, putting the backup generators on a platform so they wouldn't get flooded.
  - dredmorbius 7y ago
    The placement of generators at the base of the containment structures was already a mitigation to another risk: that ground movements common to the area might damage or destroy the generators. A complex environment is one in which multiple risks constrain options and operations. Generators can be placed neither high nor low without assuming risks or requiring additional mitigations. If a solution space is a search through a complex multidimensional topology, constraints place bounds on that space.
    - 7y ago

Small molten salt reactors were actually built. They weren't just paper reactors, even in Rickover's day. Now there are companies working on MSRs for commercial use. There's more computer simulation than running reactors so far, partly because governments are much more cautious now. But I'll also point out that Rickover in 1957 had no idea what computers in 2020 would be capable of. I think it's reasonable to suspect that we didn't find the best possible reactor design six decades ago.

brandmeyer 7y ago
The part that the computers help with - nuclear and thermodynamic physics ODEs - isn't the hard part. A couple of sharp grad students can do that with OpenFOAM and a modest workstation. The hard parts are the operations, maintenance, building for robustness, and enormous amounts of empirical field experience with the materials. Rickover and his contemporaries judged - rightly - that those issues would make MSR and metal-cooled reactors too finicky to be of practical military value.
- perl4ever 7y ago
  ...but there actually was a liquid sodium reactor sub built in the 50s, from what I've heard.
  - brandmeyer 7y ago
    Only one. She was a contemporary of the Nautilus. It was their experience with that vessel that killed liquid metal coolant for the navy. It might help to understand something special about the nuclear navy in those days - prototypes in the lab don't count. Prototypes on land don't count either. The only prototypes that counted were ones that actually went to sea. All the way through the Skipjack generation, they were regularly doing crazy-eyed one-off experiments. Nautilus and Seawolf were both prototypes.
    - willvarfar 7y ago
      Agree. One Soviet 'November' class (K27) and all 7 'Alfa' class used liquid metal. It wasn't developed further or used after.
      - DennisP 7y ago
        But Russia does have two sodium cooled reactors on their power grid right now.
- DennisP 7y ago
  Maybe we shouldn't base our civilian energy structure on what has practical military value, even if it made sense as an expedient in 1957. Operations is hardly an advantage of LWRs over MSRs, especially if you stick with simple designs instead of LFTR (as almost all real projects are doing). Those also make maintenance pretty simple. Building for robustness is a lot easier when you don't need build lots of active safety systems or an oversize containment dome due to high-pressure coolant. And some designs use existing nuclear-rated materials.
dredmorbius 7y ago
And comining on sixty years later, still not decomissioned, because dealing with both chemically and radioactively volatile materials is a nightmare.

Animats 7y ago

Yes. Most of the problems with reactors have come from parts that are glossed over at the PowerPoint level. Many exotic reactor designs have been built, and most have had problems. Everything with complexity inside the radioactive area has been troublesome. The AVR pebble-bed reactor in Germany had a pebble jam, and still can't be decommissioned. Often the problems involve plumbing. Sodium reactors have leaks and sodium fires. Helium gas-cooled reactors have helium leak problems. Ft. St. Vrain was a really good idea which had both helium leak and corrosion problems. Three Mile Island was a valve problem. Fukishima was a failure of the cooling system after a tsunami. This is why some of the new designs are viewed skeptically. Molten salt reactors involve a chemical processing plant that works on radioactive materials. Radioactive chemical plants have always been troublesome. They tend to become toxic waste sites. (See Pantex, PUREX, etc.) Modular reactors with multiple reactors sharing the same cooling pool are at risk if anything leaks and contaminates the pool. Commercial reactors have to be built for a 40-50 year lifetime to pay off. Anything inside the radioactive zone is essentially unmaintainable for that period. Complexity inside thus just doesn't work. Which is why we're stuck with water and pressurized water. Simple mechanically and chemically.

WalterBright 7y ago
> Fukishima was a failure of the cooling system after a tsunami. The fixes to that design would be straightforward. For example, putting the backup generators on a platform so they wouldn't get flooded.
- dredmorbius 7y ago
  The placement of generators at the base of the containment structures was already a mitigation to another risk: that ground movements common to the area might damage or destroy the generators. A complex environment is one in which multiple risks constrain options and operations. Generators can be placed neither high nor low without assuming risks or requiring additional mitigations. If a solution space is a search through a complex multidimensional topology, constraints place bounds on that space.
  - 7y ago