Thread by @m_older

There was so much interest in this thread tracing the Fukushima Dai-Ichi crisis as it happened with reactor 1 that I'm going to do the same for reactor 2 and maybe 3. It's a lot, so I'll be adding to it sporadically. But it's an important story to tell.

https://twitter.com/m_older/status/1366324146901254146

I'm drawing from this report I researched and co-wrote for the French nuclear safety institute in 2013-2015 (note there's a page or two in French but after that it is in English) irsn.fr/EN/newsroom/Ne…

The report is drawn largely from official Japanese reports: Investigation Committee on the Accident at the Fukushima Nuclear Power Station (ICANPS); Nuclear Accident Independent Investigation Commission (NAIIC); TEPCO reports; & this book by Kadota amazon.com/dp/B0833Y1WX9/…

(Note that the book hadn't been translated at the time so I was, with a lot of kanji look up, working from the Japanese version. I haven't compared it to the English and I don't know if there are significant differences. Highly recommended if you find this thread interesting).

Like reactor 1, reactor 2 scrammed and shut down correctly when the earthquake hit. But nuclear reactors cannot, as they say, stop on a dime: they are very, very hot, and when they stop, they stop powering the functions that keep them cool (mostly circulating water, which, btw

is WHY nuclear power plants are close to water! Lots of people tried to dunk on Japan for even putting the plants there, but that's how nuclear plants are. The US has plants along the Mississippi River, notorious for flooding, as well as on the coasts).

Grid electricity went down due to damage to a switchyard breaker, but the emergency diesel generators started as expected, and the operators started the emergency cooling systems. Doubtless adrenaline was flowing, but shutdown was going according to emergency plans.

Then, the lights go out. The operators don't know what happened: they're in a windowless room. (What happened was a massive wave well > 20 feet high crested the seawall & swamped the plant, throwing around large debris & smashing things, esp the generators and electricy panels).

I mentioned before that operators for reactor 1 and reactor 2 shared a control room. Here's a picture of control room 3 that I got off the TEPCO site when I was doing my slides for this lecture last year:

kind of grim, right? you can tell the plant was first commissioned in the 70s. Ugh, hate to think of so many working hours in there, even without the disaster. But for the operators it was familiar, maybe comforting. Kadota quotes some of them speaking with affection about

specific reactors, like they learned and came up on one reactor and then switched to another but had fond thoughts about the first. And as the crisis unfolded more and more operators crowded into the control rooms to help and support.

but if you look at that picture, you'll also see a ton of switches, lights, dials. All of these were how the operators dealt with the reactors, how that sometimes affectionate working relationship was mediated. And after the emergency power went out, all of them were blank.

Reactor 1 & 2 shared a control room, but they were slightly different generations of reactor and had different emergency cooling systems. The reactor 1 system worked on condensate and theoretically could run without electricity. The operators had been turning on the RCIC system

of reactor 2, which then would automatically turn off when it reached a certain level, and then they'd turn it on again. They turned it on at 15:39; then they lost power and indicators. The operators assumed that the RCIC wasn't working. The reactor needed to be cooled before

the water evaporated, exposing the fuel rods, which would eventually burn through the containment, melting down. There was another cooling option, a high pressure coolant injection system, but they had no indicators for that either. By 16:25 the shift supervisor told the ERC

(that's the emergency response center where the bosses and other non-operator staff of the plant had congregated in an earthquake-resistant building) that they had no way of knowing the level of water in reactor 2; basically, that they thought it was on the way to meltdown.

But the operators neither ran away screaming nor gave up. They were trying to come up with alternative cooling solutions for both reactors. 1 idea was a diesel fire pump meant for fire control in reactor building; the pipes could be reconfigured to send the water into the core.

This pump was put into operation (without great success) for reactor 1, but when the operators went to check on the pump for reactor 2 the room it was in was flooded, so they couldn't get to it.

That pump in any case was in a different building - not the reactor building. The power plant superintendent, Yoshida, had been through a (less damaging) earthquake at a different power plant and remembered that outdoor pipes had been broken. He suggested connecting a fire engine

directly to a port on the outside of the reactor building in the hopes that the pipes in the building would have survived better. This was not in any of the emergency manuals, and it took some time to figure out how to proceed.

With little recourse in terms of getting water into the core, the reactor 2 operators realized the reactor would soon be leaking radiation, raising the pressure in the containment building and they needed to consider venting before it exploded. By 18:25 they were looking into

the emergency procedures for venting (not only had they never done it before, it had never before been done anywhere in Japan. There were procedures explaining how to do it, but they did not include how to do it without electricity.

Teams in the ERC were also trying to help by figuring out which valves for venting could be opened manually. They sent people to an admin building which had been declared
off-limits due to potential structural damage
from the earthquake to look for diagrams.

In this interim please enjoy this artwork by @kenjulyonetani of uranium chandeliers sized proportional to the number of nuclear power plants in each country, which I was lucky enough to see in Singapore in 2013 kenandjuliayonetani.com/en/works/cryst…

So the operators and the ERC were trying to figure out the valves for manually venting, since the station blackout meant that they couldn't open it by just pushing a button. Nobody knew this information, which is not entirely surprising since it's a very rare emergency procedure

made more rare by the need to do it without electricity. BUT! It seems that there were people who, because of their jobs, might know which valves - but those were not TEPCO staff. They were subcontractors, and apparently not on site because they couldn't reach them until 12th.

Meanwhile, the operators decided that even if the pump for reactor 2 wasn't working, they should try to reconfigure a water line into the core, in case they found another source of pressurized water. That took most of the evening.

And all this time they were sure that the reactor was creeping - or rather, boiling - closer and closer to meltdown. At 21:13 ERC staff estimated that, the water
would reach the top of fuel at 21:40, and they
report this estimate to the authorities.

But, as you may remember from reactor 1, the ERC has had people scrounging car batteries from the parking lot and poring over diagrams to try to connect them. At 22:00 they get the indicator of water level to light up, and it shows a level of 340 ml above top of the fuel, good!

The cooling might be working! OR, the indicator might NOT be working ¯\_(ツ)_/¯. But at 23:25 they manage to light up an indicator for the pressure around the reactor and, unlike reactor 1, it's well within specifications.

So attention swings hard to reactor 1. But reactor 2 is far from in the clear; not only that, they still don't really understand what is going on in there. Between 1 and 2 am operators set out for the room with the engine for the emergency cooling system.

Radiation levels are rising around the plant because of leaks from reactor 1, obviously it's dark, and there are debris - like crushed-car-size debris - scattered around by the tsunami. They make it to the room, but it's flooded to the tops of their boots and they don't go in but

they do hear a sort of metallic clanking, so maybe the mechanism is working? Remember they can't communicate with the control room from the field so they go back and report in. At 2:10 they go back. They water is even higher, but they go into the flooded, dark room.

They hear the metallic sound again and the pipes are vibrating, but when they leave they still can't confirm or deny whether the cooling system is working.

So they go to the reactor building, because as we see throughout this story the impulse to "know", or measure, is powerful, and compare pressure in the reactor and discharge pressure of the RCIC and conclude it is working. Very good news!

But if it's working that means it's using resources that need to be replenished, and indeed at around 4 am (nobody is sleeping. would you?) they confirm that the water in the tank that feeds the cooling system is dropping. so the team decides to change the source for the RCIC

(that's the cooling system, short for Reactor Core Isolation Cooling, all the acronyms are in the report) to the suppression chamber tank, apparently there's also a risk of that increasing too much, I don't remember the details of why.

but mainly they want to save some of the water in the condensation tank that's the normal source for the RCIC, because it's not easy to move water around under these conditions. So from 4:20 to 5am, March 12, these operators are in the basement of the reactor building,

moving valves around in the RCIC room with water up to the top of their boots. Now, this change also meant that the temperature & pressure of the suppression chamber would gradually rise as the water but <foreshadowing> operators did not check on these until 4:30 on 14 March.

Why not? They had a lot going on, and remember all of this is desperate cobbled-together plans that don't have any basis or instructions in the emergency procedures, because THERE WERE NO EMERGENCY PROCEDURES FOR STATION BLACKOUT. Also, nuclear plants are complicated.

Like, you can't necessarily figure out what the impact of a change will be until you've done it complicated. Lots of things going on. Lots of moving parts & deadly temperatures & deadly radiation & pressure & steam & oh yeah, in this case aftershocks & dead fish & crushed cars.

Meanwhile, the reactor 1 situation is rapidly worsening, the radiation levels IN THE CONTROL ROOM are rising so that the operators are crouching on the reactor 2 side of the room and the ERC sent over some (bulky, awkward, necessary) protective equipment for them.

Also around this time superintendent Yoshida in the ERC gets in a fight with TEPCO HQ because they suggest that only staff < 40 should take iodine pills, while Yoshida doesn't want to make the distinction, wants everyone to take them. (they end up with required <40, optional >40)

Most of March 12th is taken up with the very difficult tasks of manual venting and then setting up seawater injection for reactor 1 (see other thread). Around 17:30, Yoshida orders preparation for the venting of reactors 2 & 3. They're trying to get ahead of things,n

not have to deal with the dangerous rush that they had with reactor 1. Meanwhile, two of the operator shift team have just been evacuated because they went over the legal radiation limit while attempting to vent reactor 1, so they all want to do it before it's too late.

So that you know: there were no fatalities from acute radiation poisoning from this accident. Long term damage is obviously much harder to track.

So operators manually open the valves creating a venting route for reactor 2; there's a rupture disk blocking the release of radiation that will automatically break and allow the venting when the pressure gets too high. But.

There had been an explosion in the reactor 1 building earlier that afternoon. At this point people (in the ERC and perhaps HQ) are theorizing that it was due to a hydrogen leak. Worried this could happen to reactor 2 as well, the ERC tells them to close that venting outlet again.

Meanwhile they are still trying really hard to get electricity connected again on the plant, because that will make everything so. much. easier. If you read reactor 1 thread, you know that part of the problem was the power centers that managed transforming voltages were damaged.

But around 20h on March 12 they found an undamaged power center and try to connect a cable. This means clearing roads, cutting through steel doors bent by the force of the tsunami, not to mention sourcing cables and connections. 40 staff involved, work ongoing for days.

March 13 around 8 operators go BACK to reactor 2 building to reopen valves they had to close the evening before. (did anyone sleep that night? I certainly hope they did. you would think they would have to. but also hard to imagine). Pressure still not high enough to break disk.

They also start preparing to inject seawater into reactor 2 in case the RCIC stops working. But there is a LOT going on with reactor 3 at this point. On March 14 around 11am there's an explosion in reactor 3 building. Then at noon, the water level in reactor 2 drops.

By 13:25, Yoshida is convinced the RCIC has stopped. They are trying to inject seawater, which is now going to come straight from the sea, but construction of the pumping line is interrupted by frequent aftershocks and evacuations.

(I have to imagine, too, that by this point, THREE DAYS after the earthquake, everyone is just absolutely wrung-out exhausted).

So the fire engines aren't ready to begin pumping until 16:30, which is right when they estimated the water would reach the top of the fuel.

So now there's a complicated engineering-related situation. They need to get the water in. But the pressure is too high now in the core for the amount of pressure the fire engines can exert in pumping. They need to vent, but remember that foreshadowing from before?

The water from the RCIC has been cycling through the suppression chamber, heating it up and raising the pressure there. Yoshida thinks they need an escape route for this pressure before they can risk venting. But the head of the Nuclear Safety Commission favored venting+water 1st

Yoshida won, but when they tried to vent the suppression chamber, the air compressor they had wasn't strong enough to open the large compressed-air valve to do so. This brought up that discussion again, and the TEPCO president decided they couldn't wait and sided with NSC

so at 16:30 the line was connected and 4 minutes later they connected 10 batteries in a row to open the safety relief valve. But it didn't open right away, didn't stay open when it did, and because of the heat and pressure in the suppression chamber, depressurization was slow.

At 17:17 the water in the reactor reached the top of the fuel, at 18:22 it was 3,700 ml below the top of the fuel, exposing the fuel rods completely. At 18:50 the indicator of water level passed off the scale.

The pressure did not get down to a level which would allow water injection until 19:03.
(wait for it...)

Then at 19:20 the firefighters realized that the two
engines which should be pumping the water in had run out of gas.

This happens again and again: extremely complex feats of engineering, undone by a very mundane lack or mistake which, in the midst of this increasingly desperate catastrophescape, is not mundane or trivial at all.

They restarted the fire engines at 19:54 and 19:57, leaving at least 37 minutes when the injection was not happening. After that, the injection team set up a schedule of relays for checking the fuel regularly.

They managed to get the water going in, and by 21:30 it was only (!!!) 3000mm below the top of the fuel. But the pressure in the reactor was high, and it was possible the injection wasn't pressurized enough to keep pumping the water in. With fuel exposed, and no water going in

it was only a matter of time until a meltdown. At this point Yoshida was talking to Tokyo HQ about evacuating staff, although apparently this meant only nonessential staff: he and others were almost certainly planning to stay.

At around 1am the pressure in reactor 2 stabilized, postponing any talk of an evacuation. But in the morning - March 15 -there was another explosion. Given the situation, they assumed it was from reactor 2; also, the pressure indicator for the suppression chamber showed zero,

which was impossible. An explosion in reactor 2 at that point meant much more danger, and Yoshida ordered the evacuation. 650 staff left at 7am; 50-odd people who remained included Yoshida, high-level staff, & necessary operators nominated by the heads of the teams in the ERC.

At 8:11 some worker who had been near reactors 3 and 4 were able to get back to the ERC and report that the explosion had in fact been in the reactor building of unit 4; this was followed by a fire. The high levels of radiation prevented assistance from local firefighters, but

by around 11:00 the ERC could confirm that the fire was out. Yoshida decided that the conditions allowed for the staff to return to the plant, starting with managers. Reactor 2 was more or less stabilized, although the long, costly, dirty, decommissioning process remained.

I'll continue with reactor 3 tomorrow, in a separate thread. In the meantime, if you would like more absolutely ridiculous stories of nuclear disaster and desperate heroics, here: en.wikipedia.org/wiki/Windscale…

also your reminder that there is a conference on these topics running at @BelferCenter this week: belfercenter.org/event/nuclear-…