[stephengillie]

It's possible to make a nuclear reactor fail-safe. It's only poor engineering that lead to Fukushima being fail-dangerous.

From here it moves from being a technical question (how to do this) to being a political question (who does it and why). How do we convince non-technical bureaucrats to convince reactor architects to make these be fail-safe?

[pdonis]

> It's only poor engineering that lead to Fukushima being fail-dangerous.

Poor engineering (more precisely, site selection) of the backup diesel generators, yes.

Poor engineering of the reactor itself? Absolutely not. The reactor was engineered extremely well; if it hadn't been, the radiation release due to the extended loss of backup power would have been much worse.

Yes, it's true that today we know how to design reactors that don't need backup power in the event of an emergency shutdown. But we didn't know how to do that when the Fukushima reactor was built. You can't fault the engineers who designed and built it for not doing something that nobody knew how to do at the time.

> How do we convince non-technical bureaucrats to convince reactor architects to make these be fail-safe?

I would say it's more a question of how do you stop non-technical bureaucrats from preventing reactor architects from making them fail-safe. From what I've seen of the decision-making processes that were at work at Fukushima, I strongly suspect that it wasn't technical people that made the decision to put backup diesel generators where they could be flooded by a tsunami. It was bureaucrats who were too ignorant to understand the issues involved.

[stephengillie]

> Poor engineering of the reactor itself? Absolutely not. The reactor was engineered extremely well; if it hadn't been, the radiation release due to the extended loss of backup power would have been much worse.

I've always had trouble understanding why the nuclear reactors weren't powering their own pumps. What sort of engineering situation led to them requiring yet another generator? Were they powered from the power grid instead of directly from the reactors?

> It was bureaucrats who were too ignorant to understand the issues involved.

This is so frustrating. But we can't expect elected officials to be experts on every subject matter.

[pdonis]

> I've always had trouble understanding why the nuclear reactors weren't powering their own pumps.

They do in normal operation, but they can't (obviously) if they're shut down. That's what the backup generators are for, to power the pumps so coolant flow is maintained for decay heat removal if the reactor has to shut down. Newer reactor designs don't require pumps to keep running after shutdown for decay heat removal; but those designs didn't exist when the Fukushima reactor was built.

[Zigurd]

> It's only poor engineering...

Poor policy, really. We should not be extending the operation of old plants.

The engineering is relatively poor based on what we now know about building and siting nuclear plants, but it was state of the art at the time.

[mikeash]

I don't think putting the emergency generators for a flood-prone site in the basement can be considered wise even given the knowledge of the time.

[tsotha]

GE reactor plans called for the batteries and generators to be in the basement, and TEPCO decided not to deviate from the plans because they were worried any changes could introduce subtle failure modes.

[Zirro]

"It's possible to make a nuclear reactor fail-safe"

Are any such "fail-safe" reactors in operation today? You make it sound like a very simple problem.

[stephengillie]

You make a great point. In retrospect, I think this is in the same state as remote work policies - techies can't see any reason it wouldn't work in theory, but practical implementation eludes us today.

[crocal]

Conceptually, Thorium reactors are an answer known to be workable. To put it simply the reactor is like a bowl of radioactive soup with a plug at the bottom made of ice and self refrigerated using the power produced by the reactor. If it stops working, plug melts, soup flows out of the bowl and reactor stops. The soup is not too radioactive, and thorium is more Common.

This design is not without challenges, but certainly not bigger than fast reactors and nuclear reprocessing. There was even working prototypes in the USA very early in the atomic age. However, Thorium reactors do not produce plutonium, and so, cannot be the technology selected in a cold war context where you need to build atomic weapons by the score.

So, no such reactors are in service today, but they are being studied, especially in India if I recall properly.

The reactors you see today were designed to produce bombs primarily. They only produce energy as an happy side effect. Safety was important, but not the main design goal.

To fix this, you need to overcome two problems.

1: you must convince governments to pile billions on developing this technology, with no military goal or economical incentive in sight.

2: you must overcome resistence to change of existing industry, who basically must turn their known design upside down. Good luck with that. See what AREVA has done with its stupid EPR...

So not a simple problem, but maybe complexity is not where you thought ^^

[dogma1138]

That's a working Liquid Metal or Liquid Metal Salt reactor not necessary a Thorium Reactor.

The US prototype "Nuclear Jet Engine" used a similar design to prevent a run off.

Reactors today also weren't designed to produce bombs, this was true for the 50's, the US isn't running breeding reactors any more, the US for the most part doesn't produce much if any fissile material any longer, the Netherlands actually produces more enriched Uranium per year than the states.

ATM there are pretty much 2 countries which use breeder reactors exclusively that's India and Russia, Japan and France still operate some but they are also phased out.

As for the Thorium crap, does a thorium reactor works? yes, is it better than other LM/LMS reactor designs no. Is it safer than other WAS reactor designs no. Is it less radioactive than other reactor fuels, thorium yes, the actual LM/LMS fuel for a thorium reactor nope, for both slow and fast thorium cycles you are going to mix in some really nasty isotopes with the salts to kick off and control the reaction and the burn-off rate. Thorium reactors will still require breeder reactors to produce fuel elements for the actual fuel.

While I also like to believe that some YouTube video can solve all of the problems in the world and that everything is a government conspiracy the world doesn't really work that way.

Yes in principle you can design a safer "Thorium" liquid metal salt reactor than a 1950's fast breeder reactor but they aren't necessarily safer than modern 4 generation reactors like BN-1200 (BN-800 can be considered a gen 3.9 since it's a faster breeder liquid metal cooled reactor which is kinda of a prototype for the Gen 4 BN-1200 one) which is a sodium cooled fast breeder reactor, unlike the (classical Shippingport) Thorium design the BN-1200 (and the 800) design can be safety used with (since the coolant can reach 500c+ in the heat exchanger without being under any pressure) modern high efficiency gas-steam turbines, competing designs which use Fluoride Slats and Helium as coolants also offer WAS operations in many cases with coupled with high efficiency.

Also an important thing to remember is that while that pipe-dream Thorium reactor might be safer on paper we have 70 years of experience with dealing with Uranium fuel cycles and reactor design. Those 70 years of experience, procedures, known and understood risks and issues are quite important when designing things that might kill 100000 people when they go boom, saying well i found this old design from a US project to irradiate the upper atmosphere with nuclear jets doesn't really inspire much confidence in anyone who ever handled even a simple risk assessment.

[crocal]

Let me sum it up, removing the usual layer of intimidation and mockery: - This is crap but humf yeah it can work on principles - Reactors are not designed to make bomb but humf yeah a couple of country have these around. Oh you know, details on the map like India, Russia, France,... - And anyway, we have 70 years of experience doing this, trust us.

I have witnessed this attitude all too much as I grew up in the heart of French Nuclear industry. Arrogance is the reason why after 70 years, despite superb technological feats, Nuclear industry has failed to convince anyone that it is safe.

Probably you will not care, but your last statement is frightening and wrong. If your Tech can kill 100000 people, then 70 years of procedures and experiences will not be sufficient to convince any assessor applying recent safety certification standards. If you have learned so much about the dangers of using this technology, then design one machine that is fail safe under very adverse Human behavior. And demonstrate why in understandable words. Gen 4 reactors so far have failed to do that and they are not even working yet.

[dogma1138]

And again what does this has to do with Thorium? On one hand you propose a liquid metal salt reactor which doesn't have to be using thorium, on the other hand you insist on using older designs with slower burn rates which require the use of water as coolants and cannot be coupled with super-critical boilers to run gas-turbines.

No nuclear technology is perfect, no nuclear technology is intrinsically safe even if you have a WAS (Walkaway Safe) reactor design.

The idea of reactors like the BN-800, BN-1200 and other ->/= Gen 4 reactors is to couple efficient electricity production, multiple fuel cycles to reduce waste, a walkaway safe design paradigms to reduce the likelihood and the impact of a run away reaction while building on top of 70 years of experience.

I'm not an expert on the French nuclear industry and considering that the French pretty much proliferated nuclear tech to half the planet i would say their reactor designs is the least of their sins.

But picking up some competing design from the 50's and saying that it's the answer for all of our problems is just criminally naive there are tons of competing designs for 4 Gen reactors and reactors that cannot be even tied to a specific evolutionary generation some of them are safer than the "proposed" Thorium reactor, some of them give better answer to nuclear waste, some of propose to build the reactor as a nuclear waste containment vessel and pretty much have like a roman candle like design that can be then left in the ground after all 3 fuel cycles have been depleted, some offer very high thermal to electrical wattage ratios so saying X is the best is just scientifically wrong.

[crocal]

I never wrote Thorium was the one final solution but "a" solution. And I also wrote that there would be challenges. I replied to a question about whether such design existed. If other fuels can be used than Thorium, that is fine with me. My main points are not about Thorium at all. I know there are other avenues (though certainly not as well as you do).

So then let's not use words like "criminal" when people just consider alternate designs, new or ancient, maybe?

[elektromekatron]

Give less weight to relative hours of previous service in the safety assessment. The reason crap old designs are still built today is because they score higher on the risk assessment than safer designs due to the accumulated hours of service.