Hacker News Reader

Comment by baseten | Hacker News Reader

So they've finally managed to solve the little problem of the extremely high temperature AND highly corrosive nature of this fuel?

View on HN

The high temperature limits the materials we can choose from. We've elected to stick with known materials (stainless steel) for the first generation to keep our schedule short. As someone noted below highly corrosive depends on the material you are in. The fuel salt is actually pretty modestly corrosive as long as the chemistry is kept right. Specifically, we have to keep the fuel salt reducing - we don't want any free fluorine running around. We keep a balance between UF3 and UF4 (roughly 99% UF4). It's like keeping the chemical balance in your swimming pool. Under those conditions the vessel will last a very long time indeed (>60 years). But it takes a long time to prove this and we have to swap out the graphite anyway so we swap out all critical components every four years. They get disassembled, cleaned, and normally put back into service with a new graphite load. This is kindof like your laser printer cartridge.

Stainless steel. Could you go into more depth about who's making your pipes and fittings? What surface treatments will you be using, etc? Will you be putting a cooling jacket, insulation, or anything over the pipes? Very interesting in the metallurgy you're using since it appears to be the BIGGEST problem of the MSRE at Oakridge.

Also, what happened to that experimental reactor? Didn't Obama send it to Norway or something?

The challenge with metallurgy for MSRE was two fold. First was the neutron interaction with nickle forming helium that migrated to the grain boundaries. We avoid this problem altogether by having a single fluid design with a protective shield of B4C absorbing neutrons before they hit the wall.

The second problem was with tellerium penetrating the Hastalloy and weakening it at the grain boundaries. This isn't a problem with stainless steel.

The stainless steel planned is SS316 which is available from multiple sources.

The primary loop does not have insulation surrounding it but it does have a 1m thick graphite reflector to bounce most of the neutrons back to the core then a layer of B4C to absorb the rest before they get to the vessel.

The MSRE was shutdown decades ago and recently had its fuel salt removed. The vessel and piping are still in Oak Ridge inside its concrete silo.

Great, grain boundaries are really what it gets down to.

I only have two things to say: 1.) Good luck testing and gathering data on these ideas. You'll need it. 2.) Are you hiring mechanical engineers?

What happens when something goes very wrong and corrosion happens? Does the fuel leak into groundwater?

The primary loop (that contains all the fuel salt and fission products) is inside a can so if we get a leak in the primary loop the salt will be contained inside the can. The can drains to the fuel dump tank (FDT). Both the can and FDT get passively cooled by the membrane and all entrance pipes come in from above so there is no weakness at the bottom.

These in turn are inside a silo, which is inside the silo hall. All total there are four barriers to break through before we get radioactive release.

Further, the fission products are combined with fluoride upon forming (when you fission UF4 the uranium is split into two fission products and the F4 is available for recombining with the fission products). Most of the fission products like to stay in the salt so even if the salt gets spilled the fission products won't wander farther than the salt spill.

And what will happen to the solidified salt spill? Could it be safely cleaned up or would it be too radioactive to approach?

It will be rather intensely radioactive. A person cannot approach it. A cleanup like this would have be done robotically. If the spill is contained in the can then the whole can is designed to be sealed and shipped back to the can recycling center where there are facilities for cleaning up the can. The main point though is that the spill is contained within the building and does not spread to the environment. A worst case accident becomes an economic loss of the reactor rather than a mass evacuation.

When it comes to corrosion, the question is always "corrosive in what?" A couple years ago I looked through the Oak Ridge documents on MSRs, and they appeared to be pretty confident that the alloys they were using were sufficiently durable.

It looks like this prototype is based on the Oak Ridge MSR experiment, which ran successfully for years didn't it?

Yes it is based on their work. MSRE was a successful experiment that gathered a lot of data. MSBR and DMSR were follow paper designs based on MSRE. We've combined elements of each. A major difference though is that we are not trying to be a breeder yet. Our first priority is to be very safe and lower cost than coal as soon as possible. We'll leave breeding and absolute minimal waste production for a second generation.

Nope.

Molten salt corrodes stainless at a rate of 0.025 mm / year. The solution: make the walls thicker, change out and inspect the primary vessel and plumbing every 4 years; you get 40 years per millimeter.

I think their solution to that is to replace all of the parts every 4 years.

This makes me think I do not understand what "walkaway safe" means.

You don't have to handle the decay heat with active cooling after stopping fission. You can simply "walkaway safe". They're using physics and engineering design to replace failsafe systems.

If you've got your thorium spread out and not reacting it should no longer be high temperature well before the 4 years are up.

This means that in the event of a problem the reactor both shuts down and passively disposes of the decay heat so that it has not chance of going critical (starting up) again (which was a risk at Fukishima without active cooling).

At that point you can just let it sit there "forever" if you want. Not that you would of course, you'd want to clean it up.

The overall concept is good, but there could be things that go wrong in the real world. How fast the heat ramps up vs the failsafe plug in the vessel melting, or the chance of some material failure shedding debris which could potentially plug the drain path.

Actually if you read the Oakridge reports those cases are considered. For reference this reactor would have both survived and shut down cleanly in both the Chernobyl (over driving the reaction) and Fukishima (30m tsunmai + 9+ earthquake) situations.

The design has two passive cooling systems. The first uses natural circulation through the heat exchangers to the normal primary cold sink. The second uses radiative heat to the membrane to a pond intended specifically for decay heat. Either one alone is sufficient - so if something (like a tsunami) takes out the primary cold sink we can still cool the core. If despite this the core overheats then it drains to the fuel drain tank. There we have passive cooling using the membrane.

If the drain between the core and the drain tank fails AND the both other passive cooling methods fail to keep the core cool enough then yes eventually the core vessel will fail. In that case, the fuel escapes the first containment (the primary loop) into the can. The can drains to the fuel drain tank so we still provide cooling.

"Walkaway safe" means that in an emergency/failure situation, it transitions from "working reactor" to "broken reactor" without threatening lives and property with danger (at least, not threatening once you've exited the reactor complex proper -- you might in fact need to walk away, especially if there are fires and things that caused the problem).

The alternative is a Fukushima-style transition from "working reactor" to "unapproachable death trap generating clouds of hydrogen gas which can explode the containment dome, sending toxic radioactive smoke from uranium fires into the atmosphere (while other material leaches into the ocean via the water table)". Even Fukishima's spent-fuel cooling ponds required water be added for for weeks and weeks on end.

If you'd like it to keep the reactor working while walking away from it, that's another matter which is keyworded something like "unattended operation" not "walk-away safe".

Molten salt is way less corrosive than water.

Very black and white answer there, which is odd as it is a very relative question. Salt is actually incredibly corrosive to a lot of things, as is water, but what materials are we asking how corrosive they are against?

I'm assuming the grandfather post means against metals (stainless steel specifically).

Reply on news.ycombinator.com