[DennisP]

I doubt anyone ever claimed that Chernobyl was nuclear Done Right. It was probably the most unsafe reactor design ever put into production.

The best modern designs have a strong negative feedback: as the fuel gets hotter, the reaction rate slows down, just due to the physics of the fuel and coolant. Chernobyl did the opposite: as the fuel got hotter the nuclear reactions happened faster. It relied on human operation of control rods to keep it from blowing up. And on top of that it skipped the containment shell.

[varjag]

Actually, it relied on two complimentary safeguard circuits to restrict the rod movements within its calculated performance envelope. They could not be overriden from the front panel, so electricians were dispatched to cut them off, in order to conduct an "experiment" in the wake of International Labour Day. How safe are the other reactor designs when malicious human intervention is considered is still a very open question.

I would also remark there was a deserved stream of criticism from the West both for the reactor design and the containment effort that followed. The overwhelming opinion throughout the 1990s-2000s was that the Western designs are safe and catastrophic scenarios like Chernobyl are implausible. Then Fukushima happens, the three reactors blow up and the only saving grace for Japan was the wind blowing oceanside. And the containment is done by low wage workers in sneakers, while Asimo and its advanced robot colleagues play robosoccer in Tokyo.

The narrative is now changed to other designs being "inherently safe", although it's not very plausible how any concentrated, massive release of energy can be made inherently safe. Yes you can reduce probabilities and exclude some catastrophic scenarios, but this is a thing with crises: they always come unexpected. I'm sure when the plug in a molten salt reactor fails for some reason (say tectonic shift from a quake), there will no doubt be another design touted as safe.

Now consider that fission power atm is what, 10-15% of compound world power generation? As of now it's focused in a handful of nations, most of whom are known for safety culture. As it proliferates worldwide, the average quality of maintenance will decline and the number of stations would naturally multiply, so perhaps we'll see a major nuclear accident every couple years instead of decades. And we haven't even started with waste disposal and non-proliferation considerations here.

I'd really really rather wait for fusion.

[pdkl95]

> Then Fukushima happens

Which was nowhere close to as bad as Chernobyl. There is a lot of fear-mongering going around that tries to imply they were somehow similar, but most of that relies on people not understanding the difference between a steam explosion in Chernobyl's reactor core (which had no containment building) blowing finely pulverized radioactive graphite and fission products across the landscape, and Fukushima's hydrogen explosions outside the containment chamber, which caused far less damage and left most of the fuel still contained in the reactor cores.

Nobody has died from radiation at Fukushima Daiichi, and there is a good chance any future cancer risk is extremely low. In fact, the way the reactor survived the worse earthquake and tsunami on record without causing a radiation hazard like some places around Chernobyl shows how safe nuclear power is, even in old designs. This was confirmed at the nearby Fukushima Daini reactor, which survived the earthquake and tsunami.

> perhaps we'll see a major nuclear accident every couple years instead of decades

Extrapolating from two incidents isn't useful. Since one of those data points wasn't significant nuclear accident, you're now trying to extrapolate form a single data point, which is not the kind of math we should base policy on.

The bigger problem with this kind of fear is that you're giving a pass to the coal industry (among other industries) which have far, far worse disasters regularly. Even if you count every single nuclear associated disaster (including Chernobyl), there number of people injured/killed and the land impact of those disasters is insignificant noise compared to what the coal (or chemical) industry has done.

> we haven't even started with waste disposal and non-proliferation considerations here

So you haven't looked at any of the advances in nuclear power that have been made over the last sever decades? Modern breeder designs don't have the waste problem. Most technologies advanced a lot over the last 40+ years - why would you assume that nuclear never advanced past the 60s?

[varjag]

> Which was nowhere close to as bad as Chernobyl.

Both were Level 7 accidents on INES scale. Fukushima release was about 1/6th of Chernobyl, but ended up blowing 4/5ths of its contents over the ocean. At some point, evacuation of Tokyo metro area was considered. Yeah no, both were pretty bad and the same magnitude events.

> The bigger problem with this kind of fear is that you're giving a pass to the coal industry (among other industries) which have far, far worse disasters regularly.

Fair point, but there's not much coal use in Belarus (mostly natural gas), nor in Norway where I am now (clean energy from the dam to my car's plug). For Belarus, the impact, in health, land and culture was incomparable to that of other energy accidents.

> Modern breeder designs don't have the waste problem.

Right, they have the weapon proliferation problem that you can't design away.

[DennisP]

Breeders aren't necessarily proliferative. The IFR for example ends up with a mix of plutonium isotopes which is more difficult to weaponize than natural uranium ore. The main problem is the need for a high fissile load at startup, but the same mix works for that.

Thorium breeders are another possibility, as long as you're not isolating protactinium.

Another route is Transatomic's design. It's not actually a breeder, and runs on uranium enriched as low as 1.6%, but burns up almost all the transuranics because it leaves them in the fuel mix for a long time, removes fission products, and has really good neutron economy.

Even if you use breeders with potential proliferation issues, one way to go would be to keep those reactors in weapons states, and use them to dispose of the waste from non-weapons states.

[burnte]

> shows how safe nuclear power is, even in old designs.

More than that, had the reactors been gravity-shutdown capable ones and gravity fed cooling (rather than pump driven) and the seawalls been higher as originally proposed, none of us would be talking about Fukushima. So like you said, safety isn't unattainable. Just the opposite, it's usually sabotaged by people who have no place making such calls.

[DennisP]

I'd be happy to wait for fusion, if catastrophic climate change were not bearing down on us like a freight train.

And also if coal were not killing many thousands of people every year, and we didn't have so many large dams past their designed lifespan.

[varjag]

Well frankly, converting to fission in near term (say 15 years) worldwide is pure speculation on our side. This is just not going to happen. I just hope natural gas, hydro and the increasing green generation component can help us get by and perhaps reduce the coal usage.

[DennisP]

Let's look at the worst possible case for molten salt reactors and imagine that somehow the reactor gets cracked wide open.

With a solid-fueled reactor, you need enough fuel for 18 months (the normal refueling cycle). With liquid fuel you can add a little bit of fissionable material as often as you want, so your fuel has barely enough reactivity to keep your plant operating.

With solid fuel, the fission products build up over that 18 months. With liquid fuel, three things happen:

- Noble gases bubble right out of the fuel to an outgas system. You get a tank of xenon and krypton. You can take that away to secure storage as often as you like.

- Some metallic fission products plate onto collectors.

- The major fission products we worry about are cesium, strontium, and iodine. In conventional reactors they're gases, but in MSRs they bond very strongly into fluoride salts. Optionally we can remove them as we go.

So in our imaginary catastrophe there's not enough reactivity for a big excursion, there's less decay heat than in a conventional reactor, and most of the radioactive stuff is liquid that quickly cools to solid rock. There's nothing to drive any kind of explosion, since it's all chemically stable and at atmospheric pressure.

It's worth noting that Fukushima's reactor survived a 9.0 earthquake without significant physical damage. It was the loss of electric power that took it out, and that's a problem molten salt reactors don't have at all.

[burnte]

> How safe are the other reactor designs when malicious human intervention is considered is still a very open question.

Nothing is safe from malicious action, absolutely nothing. You can cut someone with a piece of paper, drown them with a glass of water, or stab them in the throat with a pencil. You can even simply punch someone with your hand in the back of the neck and kill them. We simply make malicious action difficult and unlikely through design and rules, exactly how we manage cars and coal plants and nuclear weapons.

[varjag]

Exactly. And I think we all can agree that a sharp pencil does not have the same worst case scenario potential as an automatic rifle or nuclear weapon. It is really hard to sabotage a coal plant with the same national impact as an INES L7 nuclear accident. I think among the "conventional" energy sources only some of the hydro sites come close.

[PhantomGremlin]

I'd really really rather wait for fusion.

Over 40 years ago, when I was checking out colleges to attend, the tour at one school took us thru a fusion research lab that iirc had a Tokomak (or some similar contraption).

IMO no way that practical fusion happens even 40 years from now. I won't be around to see it.

Maybe the Chicago Cubs win another World Series before we see practical fusion? Could that happen? :)

[varjag]

I never expected to see space rockets landing down on their feet, like in 1950s sci-fi novels, yet here we are :)

[kristianp]

"I'd really really rather wait for fusion."

This is like the computer company that loses business after announcing the date for its next model, except worse, because there's no date for fusion. The threat of fusion is sucking the air out of fission.