Thread by @m_older

(I have other things I should be doing but much easier to live-tweet a report I wrote 7 years ago so) let's talk about reactor 3!

https://twitter.com/m_older/status/1367083083086909440

procrastination is not the real reason I'm doing this (although it might be the reason I'm doing it *now*...). It's a story and people want to hear what happens next, that's one reason. I want more people to know about this, that's another. It's been ten years, and the impact

of the accident on the nuclear industry is still felt, but maybe it is starting to fade en.wikipedia.org/wiki/Nuclear_p…

It has also occurred to me as I kept writing tweets when I really did have other things to do that maybe this is somewhat therapeutic for me too. I am among the least affected by this accident but I do remember how terrifying it was, deciding to go to Japan after it happened.

I remember flying in on Swiss Air, this was maybe 2 weeks after the earthquake? and the announcement as we approached: "The situation in Tokyo is" [pause] "normal."
[it was not normal. it was also not complete chaos.]

I remember how terrifying it was to eat, or breathe.

and, to be clear, I was nowhere near Fukushima! I was in Tokyo for a bit and then up north of Sendai, working on tsunami response. as I said, among the least affected. but my fears weren't completely unreasonable, either. There's a lot of uncertainty about these risks.

honestly I was so traumatized that when they offered me the job writing the report, I didn't want to take it. I didn't want to think about it any more.

Yes, of course I'm glad I took it. There are very few circumstances in which more knowledge is not better.

But there also wasn't anything in that knowledge that made me feel like "oh, that's all right then" or reconciled me to nuclear power. It didn't make it all better.

So that's a hint at my personal perspective on this. Let's go on with the knowledge. Reactor 3 started off in better circumstances than 1 & 2: the instruments were working and so was the emergency cooling! Very good, relatively!

Like reactor 1 & 2, reactor 3 shared a control room with reactor 4, but 4 was off-line, as were units 5 & 6. Here's a diagram of the layout for 1&2 (and there's a picture of 3 in the 2 thread, sorry to do it backwards)

This from Wikipedia gives you a sense of where the reactors are located (the picture is from 1975, 6 is still under construction)

Since 3 was in a manageable situation, all the attention was on 1&2 for the first day or so. But at 11:36 on March 12, the RCIC emergency cooling system of reactor 3 stopped working. The operators went to the room and tried to restart it, but it stopped again.

The water level started to drop, and the other emergency cooling system, the High-Pressure Coolant Injection (HPCI) started up. Good! Things are working! [dun dun DUN] The operators continued to manage these cooling systems.

at 15:36 that afternoon there was an explosion in reactor building 1 that hampered the efforts to inject water there (see thread 1). Even though reactor 3 was still doing well, Yoshida realized that many normally simple tasks, such as the preparation of a venting line,

would take far longer than expected under these very difficult circumstances, so he ordered the preparation of the vent line just in case (really expecting that it would be necessary sooner or later and easier to prepare sooner).

Around 20:36, the 24-volt DC source of electricity for the water level gauge of reactor 3 runs out of charge, so the operators can no longer track the level of water in the core. They respond by increasing the flow rate of the HPCI to make sure they don't run out of water, also

they can monitor the pressure inside the reactor and the pressure of the HPCI to make sure the water can be injected. Meanwhile, the ERC gives them 13 2-volt batteries (THIRTEEN! TWO-VOLT! BATTERIES!) to try to hook the indicator back up.

The operators were particularly concerned about not being able to track the water level because they had adjusted the way they were using the HPCI, modulating the flow so the standard stop and start wouldn't drain the batteries.

They decided they would open the safety relief valve to depressurize the reactor, then reconfigure the line to use the diesel-driven fire emergency pump (remember, this was an idea for both 1 & 2) instead of the HPCI; they thought this would be more reliable. [dun dun DUN]

They told the ERC about this plan but somehow did not quite tell everyone at the ERC who might be interested, which is not surprising, everyone was incredibly stressed and busy and almost certainly missing at least one night's sleep, communications were still limited.

At 2:42 in the morning operators left the MCR to change the D/DFP configuration from the spray to injecting water into the reactor. Remember, they had no way of communicating back to control room without returning. While they were out, the operators who remained in the MCR

stopped the HPCI. Then they tried to open the safety relief valve by pushing the button on the control panel.
The "open" indicator didn't change.
They tried again.
The "open" indicator didn't change.

They couldn't depressurize the reactor.
With high pressure inside the reactor, the D/DFP couldn't push water into it, even after it was turned on.
At 3:37 and again at 5:08, the operators tried to get the RCIC to work again, but they couldn't restart it.

It took a while for this dire information to trickle up the ERC. There were some people on the operations team in the ERC who knew it, but "were too busy" (I'm quoting ICANPS here) to report to the head of their team.

and imagine that feeling, where everyone has been trying in absolute desperation and for two brutal days to keep 1 & 2 from exploding, and the one reactor that seemed to be fine is, maybe, not, and all because you were trying to make it more fine.

At 3:55 they finally tell the ERC operations team leader. It then gets reported up to Yoshida and Tokyo HQ. Tokyo HQ want to know if the HPCI was stopped manually or automatically, and the team at the plant lets them think it was an automatic stoppage.

Yoshida on the other hand, now that he knows, tells them to focus on constructing a line for water injection via fire engine and finding batteries to open the safety relief valve for venting.

At 5:08 the shift team started the suppression chamber spray manually. They found that the handle of the suppression chamber spray valve was extremely hot. NOT GOOD.

Yoshida estimates that the level of water lowered to the top of the fuel around 4:15. Reactor 3 has gone from doing very well to really not.

The ERC is trying to connect electricity through the one remaining power center to a high pressure water system for reactor 3, but they had trouble with connections &
cables, which are not straightforward or easy & running a power plant is different from electrical wiring

AND there was a metal door which had been warped by the tsunami to the point of being impossible to open, delaying the work.

They get the line for seawater injection set up by around 7am, but they won't be able to get the water in there until they've depressurized the reactor. That required 120 volts of electricity, and there were NOT ENOUGH batteries on site.

This is what I'm talking about when I say crises manufacture scarcity. Not having enough batteries seems like a ridiculous statement to make about a country like Japan - they have all the batteries they could want! But getting the right batteries to the right place in time when

communications are mostly out, roads are mostly blocked, planes are grounded and trains are stopped, everyone within 10km has evacuated, and you're also trying to source generators, air compressors, fire engines, AND there are ongoing aftershocks...

Other TEPCO plants were trying to help; they sent 50 batteries to Fukushima Dai-ichi, but they were all 2-volt, and it wasn’t practical to connect 60 in a series to reach 120 volts. NOT PRACTICAL AT ALL.

The team gathes 12-volt batteries from private vehicles (those of staff and sub-contractors) in the parking
lot, bring them to the MCR, and try to connect them despite already high levels of radiation in the control room.

Opening the SRV also required compressed air, but they were hoping there was enough air left in the system for it to work.

Meanwhile, the politicians are on their bullshit again: gathered in the PMs office, hearing about the plans for injecting seawater, they tell the plant they should exhaust all freshwater resources first. And Yoshida says, sure absolutely. But the thing is, that meant that while

all the freshwater was reserved for reactor 3, reactor 2 went straight to seawater. Even for reactor 3, this meant two water lines: one getting freshwater into a cistern near 3, and one from that cistern into the port for reactor 3

Anyway, at 9:08 the restoration team connects enough batteries to reach a 120-volt charge and opens the
safety relief valve. Around 9:25, the injection of fresh water was started, using the fire engines.

So that's good! They managed something very difficult! But around 9:28, the pressure in the drywell of unit 3 went up: the depressurization wasn't working. They inspected the site and found an air leak caused by a faulty connection, which they repaired temporarily with TAPE.

There was enough air remaining in the air cylinder, so they found a new cylinder but did not change them, leaving the new one to the side to be ready.
Seeing white steam in the building and registering high levels of radiation, they very sensibly evacuated.

Around 11:17 pressure begins to rise again, and they go in to change the cylinder. Temperature & levels of radiation are very high, so they work in relays of 15 minutes each, wearing full protective equipment. At 12:30, more than four 15-minute shifts later, they confirm

the depressurization valve is open. Meanwhile freshwater is running out. Yoshida tells them to get ready to quickly change the injection to seawater. Around 12:20, the fresh water runs out. Even though they were prepared, they can't start seawater injection until around 13:12.

thanks, politicians far away in Tokyo.

Radiation is going up. As with the operators in the 1&2 MCR, operators are now crouching on the side of the room with lower levels, in this case the unit 4 control panels. The pressure in the reactor drywell is rising again; they decide to switch out the air cylinders for an

air compressor to try to keep the valve open. They find the air compressor at the office of one of the sub-contractors, since naturally TEPCO doesn't have their own available, and hook it up around 19, then refuel it periodically.

Now, the seawater for the injection at this point is coming from that backwash pit they found where some tsunami water got left behind. But in the middle of the night - remember, no electricity, so dark - as well as cold, has I mentioned it was cold? - the water level in the pit

seems too low, injection stops for a bit, then they find some places where it's still deep enough to suck up water.

Meanwhile, the pressure is still rising. They're trying different things: a small generator to keep the valve open, a new air compressor sent over from Fukushima Dai-ni, opening an additional vent. But the pressure is still going up.

Again, extreme complexity & processes that we collectively don't have that much experience with. Nobody can determine, mechanically or through calculations, what the issue is here: not with limited data they have, not with all the possible data. Certainly not under this stress.

What I mean is, it's not like they, oh, didn't do the math right, or weren't experts, or whatever. Nobody knows enough about the processes of nuclear power plants, no one has enough experience in them going wrong, to understand all of this or deduce it from the data we have.

Yoshida's worried. It seems to him that reactor 3 at this point is showing the same signs as unit 1 had just before its reactor building exploded 2 days earlier. Concerned for the safety of the workers, after a discussion with HQ, between around 6:45 Yoshida orders

that workers in the field evacuate to the earthquake-resistant building. That means, though, that they can't keep working on things like connecting electricity and alternate water lines.

The explosion doesn't happen. Between 7 & 7:20 drywell pressure even goes down slightly. There's still danger, but construction of a line to refill the backfill pit is urgent so that water injection to all three reactors can continue. At 7:30 Yoshida rescinds the evacuation order

Around 9am -we're up to March 14 now- the workers completed a line from the sea to the backwash pit they were using as a source for water injection, making it possible to refill the pit. Good!

Around 10am 7 water trucks from the self-defense forces (Japanese military, only allowed to be self-defense b/c post WWII treaty, very good on disaster response) arrive at the plant with 35 tons of water. Two are sent towards the backwash pit. Good!

At 11:01 there's an explosion in the unit 3 reactor building.

Some TEPCO staff and sub-contractor staff are injured, all are evacuated back to the ERC. Also some soldiers are hurt, and all 7 trucks leave without unloading the water. (I still say the SDF is pretty good on disaster response. There's a story I skipped earlier when some SDF

arrived earlier, on March 12, with no idea of the status or what they were supposed to do, and have to wait around before they're given tasks because the PM's visit interrupts everything).

(But mainly I appreciate the SDF because they were going around the tsunami evacuation centers with a large mobile laundry truck. also portable bath houses. Also they figured to send females soldiers to assessments so women would feel comfortable talking about needs).

When the explosion happens Yoshida is struck with guilt. Initially 40 staff were unaccounted for, and he is sure that this time someone has been killed, because he made the decision to let them go back out and work.

But little by little people trickle back to the earthquake-resistant building, and although there are injuries, at the end everyone is accounted for and alive.

More good news: The indicators for reactor 3 show that pressure in the drywell just after the explosion was
above atmospheric pressure, suggesting the explosion occurred outside the containment and the containment is probably intact. phew.

Around noon, Yoshida orders the teams to restore the water injection lines for reactors 1, 2, and 3 to attempt to stabilize the reactors.

Most of the fire engines were damaged, the hoses
unusable. The debris made it impossible to reconstruct the same lines. The team decides to construct a line directly from the sea to the reactor, via fire engines. The line for 3 is complete & the water injection started at 16:30.

And that is pretty much stabilization for reactor 3. Again, it is years of clean up and decommission. But the acute crisis is over.

If you want more information, there's a lot in the report: irsn.fr/EN/newsroom/Ne…

There's also a conference going on now at @BelferCenter about nuclear energy; I'm speaking as part of a roundtable tomorrow 11:30 eastern but most of the other speakers are far more knowledgeable about the topic than I am! belfercenter.org/event/nuclear-…

(that's not false modesty: this is kind of a side topic for me, not my main area of expertise. I just happened to have a job that required me to look very, very closely at one specific part of it)

(still have opinions tho)

I appreciate everyone who followed and supported these threads. Think about the energy you use and where it comes form; think about the people on the frontlines when there's an accident; think about your politicians and the choices they're making and vote and agitate accordingly.