there are a number of questions that need answer regarding the situation at the Fukushima Daiichi nuclear reactor, aka Fuk-D.

As a starting point, here is an amateur online feed of radiation in Tokyo (Park18).

It is a geiger-counter, the normal count rate is 10-20 cpm, around noon on march 15th the count rate peaked at about 120 cpm, the counts then dropped before there was another broad rise on march 16th to about 40 cpm.

The first question is: why am I having to link to an amateur with a geiger counter?

The Japanese Atomic Energy Authority has four online environmental radiation monitors.
All went offline at 4 pm on friday march 11th and have not come back.

I understand they are busy, I understand that they don’t want to panic people, but this is stupid.

I don’t have the energy to get into the weaknesses of 40 year old GE designs for the Mark I Boiling Water Reactor.

BWR containment specs and age related safety management (pdf) – IAEA TECDOC-1181 has many of the gory details if you care – including descriptions of the generic containment layers.

Nuclear power has its place, I accept that Tokyo Electric Power Company might have reasonably spec’d magnitude 8 as a reasonable maximum earthquake to design for, and who knew a tsunami (in Japan of all places!) would wipe out all the backups, and that it was bad luck the 9th magnitude quake hit just before the oldest reactor was scheduled to be shut down (modulo the decade extension for pure profit operation).

So, riddle me this:

  • The reactors are supposed to scram when alerted to a major earthquake.
    Given the location, Fuk-D had about 30 seconds to scram.
    A Mark I BWR should be able to insert the control rods in 10 seconds.
    Did they?
    Was it automatic? Require manual override decision?
    The design has control rods go up into the core, not drop in. Did they go all the way in before the whole thing jumped? Or did they buckle?
    I’m sure the operators have been debriefed and there is an incidence officer who knows the timeline for the first couple of days.
    Tell it.

  • What has been emitted?
    It is known, even if the PR department has not been briefed, and journalists couldn’t tell us when it is spelled out for them.
    Radiation need proper units when reported: we need milliSieverts per unit time, total curies, date stamps and size of contaminated area.

    As y’all know, ambient dosage is about a milliSievert per year.
    Occupational limit doses are usually 20 milliSieverts or so for a year.
    So background is about 0.1 microSivert per hour (or One Banana Equivalent Dose – man that 40K is nasty).

    An acute dosage of a Sievert puts you in hospital.
    10 Sieverts and you are not coming out of hospital.

    So, if there are measurements of 400 milliSieverts per hour at the site, as was reported on the 15th, workers who volunteer can only last about an hour before having to leave, permanently.
    But, is that a single small spot due to direct exposure to fuel elements? Or a little burp of 16N – with half-life of 7 secs it is hot, but not for long?
    How big is the spot that is, or was, so hot?

  • What isotopes are being measured on site?
    Are they seeing 137Cs and 131I?
    Technetium? In what quantities.
    Quantity matters.
    Are the fuel rods compromised and venting?
    Is the containment leaking?
    You can “detect” anomalous radioisotopes in extremely low concentrations, what matters is how much material is dispersed and in what form.

Basic question: did the scram of any of the reactor cores fail, and are there therefore any of the reactors sitting there with critical sub-cores – chunks of uranium still undergoing sustained fission?

If not, did the loss of primary coolant occur early enough for substantial breach of fuel rods due to thermal stress, and has the resultant fuel dispersal lead to undesigned criticality?
Are the random blobs that fell out sitting there undergoing nuclear reactions.

TEPCO ought to know this by now, and it makes a very big difference whether these are just hot cores sputtering while residual heat is bled away, or whether they are undergoing fission without permission of the operators and without actual functioning cooling systems, or containment…

Hot damaged cores are bad.
Unscheduled fission in cores is catastrophic.

Oh, and please, someone, explain to journalists that “white smoke” is often the thing popularly known as “steam”.

Habemus fission?

Go to MIT NSE Nuclear Information Hub for actual authoritative information

PS: General Electric website with fact sheets and news – eg GE is shipping new mobile generators to Fukushima, say they offered, took TEPCO this long to take them up on it.



GE mobile turbine generators

Comments

  1. #1 Derek Fox
    March 16, 2011

    Suggest putting nuclear construction activities on hold until we get a better-educated class of journalists out there.

    Governments, not to mention energy companies, are supposed to try to keep secrets. Journalists, however, should not be so easily distracted / misled / misinformed / pacified.

    (Not the only area where this shortcoming has been apparent, of late. Maybe this is what Twitter is for.)

  2. #2 Ben
    March 16, 2011

    Steinn,

    Most of what I know is what I read in the papers, but: Any reactor is expected to scram on reaction to a large earthquake. Since the earthquake was quite a few miles from the reactor, I kind of doubt that the shaking was far enough outside design parameters to interfere with scramming. We didn’t, for example, see any obvious signs of seismic damage to the outsides of the buildings.

    But even once scrammed the residual heat is immense. The failure of the cooling system and of the backup power to the cooling system is almost certainly the major cause of all the problems. To paraphrase a well-known failure analyst, “I don’t think anyone anticipated the overtopping of the seawall” (around the diesel generators).

    In a BWR the water is also the moderator. “Moderator” is a terrible nomenclature since the moderator actually enhances the reactivity by slowing fast neutrons to where they can be captured by other uranium nuclei. Reactor fuel is not highly enriched, so it does not reach criticality without a moderator. As bad as a molten lump of uranium, cladding, and maybe concrete sounds, I am guessing that it is unlikely to maintain fission efficiently. But for God’s sake we want it to stay inside the pressure vessel. In TMI-2, a partial core melt was not catastrophic, as containment was maintained: http://en.wikipedia.org/wiki/File:Graphic_TMI-2_Core_End-State_Configuration.jpg On the other hand, at Chernobyl the “corium” lava flowed into the building. Until a couple of days ago I didn’t know corium was an actual word. Of course this reactor isn’t like Chernobyl’s.

    The tragedy is that it’s entirely possible that if the backup diesel generators had been 30-40 feet off the ground the failures would not have happened. I assume that the generators are very large and this would have been expensive. This is the real problem: nobody wants to pay the full cost of energy, so safety corners for unlikely events are cut, and we wind up paying the cost when oil rigs blow out or nuclear reactors fail, instead.

  3. #3 Eric Lund
    March 16, 2011

    The design has control rods go up into the core, not drop in.

    That accounts for something that has been puzzling me. According to sources I have seen, and I don’t know if they are reliable, TEPCO came up with the brilliant real estate saving idea of putting the cooling ponds (I’m forgetting the exact technical term here) for spent/extracted rods on the roof. This would be a bit awkward if you have drop-down control rods (which I would expect because gravity will help you if the power fails), but much more feasible if the control rods are push-up.

    Which leads to the obvious follow-up question, per Ben’s point, of why they didn’t put the backup generators on the roof instead. It couldn’t be any harder to lift the generators up there than the control rods, and you can leave them up there when you shut down for scheduled maintenance (as was the status of reactors 4-6 at Fukushima Daiichi when the earthquake struck). It also has the advantage of putting the generators above any possible typhoon storm surge–while the east coast of Tohoku probably doesn’t get many landfalling typhoons, western Honshu and Kyushu certainly do.

  4. #4 Steinn Sigurdsson
    March 16, 2011

    Shakemap at earthquake.usgs.gov has peak acceleration at 0.1-0.2g at Fukushima.
    Ground wave travel time is less than a minute from the epicenter, but earthquake monitors would be on shore most likely, so I’m guessing ~ 20-30 secs notice.
    Enough time if scram is automatic, maybe not if it needs an operator to manually authorize.

    The primary diesel generators are ~ 4MWe (6000 hp), I am told.
    That is diesel locomotive size.
    Apparently the replacement generators rushed there were not “plug compatible” – or so I read anecdotally.
    That seems farcical, any half decent engineer ought to be able to bypass plugs and jury rig a DC connection straight into the system.
    I’m figuring they had the original 40 year old generators, probably US spec and therefore no longer plug compatible with what is available in Japan.

    Having the cooling ponds four stories up, in a building with blowout panels seems a bit moronic. Given how critical cooling is for the rods, you’d think they’d want in-ground ponds with drainage towards them, not away.

    The old BWRs need active primary cooling – H2O has to be force circulated, the newer designs are supposed to be able to hang in there through convective cooling only. Don’t know how critically dependent that is on the condenser and secondary cooling though.

  5. #5 Craig Heinke
    March 16, 2011

    Everything I’ve read (e.g. World Nuclear News, MIT NSE, wikipedia, etc) indicates that the scram was automatic, and completed successfully. There is no longer U fission going on in any of the cores, the problem is the residual heat, which is problem enough.

  6. #6 Eric Lund
    March 16, 2011

    Having the cooling ponds four stories up, in a building with blowout panels seems a bit moronic. Given how critical cooling is for the rods, you’d think they’d want in-ground ponds with drainage towards them, not away.

    Well yes, there’s that. There is also the minor detail that if your rods start to melt,or if the building underneath them falls down, the remaining fuel in the spent rods tends to flow toward the fuel in the reactor. You might live to tell the story if containment holds, but it kind of compounds your problems in a worst-case scenario.

    There are only two advantages. One is that you need less real estate (a nontrivial consideration in Japan). The other is that in-ground ponds on oceanfront property might not be feasible, between risk of leaks and risk of introducing salt water due to a tsunami or storm surge. My guess is that they decided in-ground ponds had too high a probability of a minor radiation release into the ocean, and they underestimated the probability of catastrophic failure of their rooftop ponds.

  7. #7 Joseph
    March 16, 2011

    For what it’s worth, Daiichi means “first”. The two plants are Fukushima-Daiichi (Fukushima-I) and Fukushima-Daini (Fukushima-II). “Fuk-D” does not disambiguate successfully.

  8. #8 Steinn Sigurdsson
    March 16, 2011

    @Craig – the reports I’ve seen say automatic shutdown was initiated. ie command to put control rods in was issued and something spring into action.
    I have not seen an affirmative statement that the rods were seen to go in completely – ie they say power drop all the way down to idle or neutron flux go down to levels consistent with idle etc — not saying they didn’t, just haven’t heard it.

    The whole sequence of events makes more sense if the core didn’t shut down completely – the diesel generators didn’t fail immediately, the core ought to have cooled down quite a lot, for it to heat to the point where Zirconium was oxidizing suggests a significant power source, didn’t think the residual heat from longer lived fission products would heat it that much, that late, even with partial coolant loss.

  9. #9 Steinn Sigurdsson
    March 16, 2011

    k@J – so Fuk-D-1 or FukD-1?

  10. #10 Ben
    March 16, 2011

    Couple quick comments:

    I don’t think having the spent fuel storage in an upper story of the containment building interferes with control rods at all. The control rods have to be inside the steel pressure vessel. Here is a cutaway of a GE Mark I BWR: http://en.wikipedia.org/wiki/File:BWR_Mark_I_Containment,_cutaway.jpg The pressure vessel is the brown cylinder, the cutaway wine-bottle shape is the primary containment structure. See also http://www.nrc.gov/reading-rm/basic-ref/teachers/03.pdf

    Something I read also implied that having the spent fuel up there was more common than just being a TEPCO innovation. Having them overtopped by a tsunami also wouldn’t be great.

    The flow rates for cooling are really large. I haven’t found an estimate on the NRC or GE website yet, but wikipedia says typically 45e6 kg/hour of water. That is 3400 gallons/min which is consistent with other stuff I have read. When a reactor is scrammed it’s still hot and it still produces about 7% of the heat of operation from residual radioactivity. So the power requirement to pump coolant is still large and the generators must be big.

    My father (who has worked on properties of fuel rods) says that if the fuel becomes partially uncovered the zirconium cladding gets very hot, hot enough to drive the oxidation reaction (implication was ongoing fission not required). These things run at many atmospheres of pressure, so even under normal operating conditions the water inside the vessel is at of order 300 C.

  11. #11 Adam_Y
    March 16, 2011

    The primary diesel generators are ~ 4MWe (6000 hp), I am told.
    That seems farcical, any half decent engineer ought to be able to bypass plugs and jury rig a DC connection straight into the system.

    I’m a half decent engineer and I happen to think that jury rigging a 4 Megawawtt power supply is insane.

  12. #12 Lab Lemming
    March 16, 2011

    IIR correctly, The Japanese National Earthquake management system is designed to initiate shutdowns of all major industrial facilities on the arrival of the seismic P waves (of a certain size) at the closest station. The P waves travel at about 8 km/sec, while the surface waves responsible for the shaking move at 1-3 km/sec. So it is likely that the shutdown was completed before the shaking started.

    Even if they weren’t, the various reactors have now been filled with seawater and borax, which should surpress future fission.

    Everything i’ve heard is that they simply can’t supply water fast enough to keep pace with boil-off, so the fuel rods are exposed to steam and/or air, both of which oxidize Zr at high temperatures.

  13. #13 Andrew Foland
    March 16, 2011

    Link to all the Daichii reactor parameters you ever wanted.

    Just after shutdown, the decay heat power is on the order of a couple hundred megawatts in the core of one of these reactors. By a day, it’s down to O(10) MW. It falls off only very slowly after a day.

    After a day, an exposed and uncooled fuel rod’s cladding melts in O(1 hr). The Fermi-1 (different design) cladding melt happened within 4 minutes of the first problematic readings.

    A day or two to boil off, and then hours to melt down, is not at all inconsistent with the decay heat, or the observations.

    Also, google NUREG-1150 if you want to read about disaster scenario meltdown probabilities for a BWR.

    As for the spent fuel ponds, an NRC document I don’t have handy suggests they boil off in 100-140 hours without coolant. Again, consistent with timelines we’ve seen so far.

    You can read for some useful numbers as to what happens once the spent fuel ponds boil off.

  14. #14 Steinn Sigurdsson
    March 17, 2011

    @Adam – consider the alternative…
    You have a generator, and load, and the plugs don’t match.
    You have a machine shop and stock.
    Jury rig a a conductor to bridge this, even temporarily,
    or have a nuclear reactor meltdown.

    @Andy – thanks, very useful.
    The spent fuel pond scenario is worrying.
    I’d guesstimated 10,000 km^2 as the ballpark area to worry about with uniform dispersal of radiological load from a core’s worth of rods.

  15. #15 Ben
    March 17, 2011

    I will guess that if someone said spare generators weren’t “plug compatible” they meant they couldn’t plug and play for some larger reason, like incompatible voltages or impedances, not that the physical connectors were incompatible. I don’t think you hook this kind of generator up with a plug and jack – more likely a junction box and bus bars the size of your arm or leg.

  16. #16 Steinn Sigurdsson
    March 17, 2011

    I’m pretty sure they are DC…
    I’m sorry, but if you lose a generator, and you have another one running, you convince the current to go into the system, and ignore safety regulations – I agree you’d need bus bars like that – it is a 4GWe power station, they have machine shops and materials.
    Power up the primary coolant pumps using whatever you can, as much as you can.

  17. #17 Steinn Sigurdsson
    March 17, 2011

    Idle thought: the first hydrogen explosion looked like tens++ kg of H2 detonating. Zr has atomic number ~ 90
    So order 10 tons of zirconium metal oxidised to produce that hydrogen.
    The core is 100++ tons.

    So a significant fraction of the fuel rod cladding had to have oxidised. So a significant fraction of the fuel rods lost structural integrity.

    ++ungood

  18. #18 astro
    March 17, 2011

    WRT plugs… I’d guess: 1) something was lost in translation (like 50 workers retreating to safety not meaning that they evacuated the plant), 2) needed machinery/tools/supplies were swept away (look at the sliding photos on ABC.au site), broken or otherwise unavailable to cut large hunks of metal, or 3) the culture “Let’s follow protocols, discuss options and achieve consensus” won out over “I’m the one on the front line, so it’s my call to do whatever I think is best.” Hopefully, they’ll live long enough to tell the story.

  19. #19 astro
    March 17, 2011

    If the primary containment vessel is not damaged, then would that much H be able to escape into the rest of the building so quickly? I realize H is small and hard to prevent leaks, but still that’s a significant enclosure.

    If not, then wouldn’t that suggest that either: 1) the hydrogen in the original explosions came from spent fuel, or 2) the primary containment vessel is damaged enough that Hydrogen can escape. As for option #1, would there have been time for its water to boil off before the first explosion (at each reactor)?

  20. #20 Fred Magyar
    March 17, 2011

    This is the real problem: nobody wants to pay the full cost of energy, so safety corners for unlikely events are cut, and we wind up paying the cost when oil rigs blow out or nuclear reactors fail, instead.

    But, but, nuclear is *TOO* cheap to meter, right?!

  21. #21 Ben
    March 18, 2011

    Hydrogen probably was produced in the reactor vessel and escaped into the secondary containment building while venting steam from the cooling system. More sophisticated explanation than I’m capable of at http://mitnse.com/2011/03/15/explanation-of-hydrogen-explosions-at-units-1-and-3/

    Nobody sensible has claimed that nuclear power was cheap for a long time now. I am a card carrying bleeding heart environmentalist and I am deeply ambivalent about nuclear power. Fossil fuels are not really better, just less instantaneously scary. While a crisis like this focuses everyone’s attention on the drama, it is likely that the long term effects on people and the environment are still going to work out much lower than, say, the aggregate effects of mining and burning coal (which generates 45% of the electricity supply of the US). I can understand why countries that are not rich in fossil fuel resources have gravitated to nuclear power.

  22. #22 Roberto Mejia
    May 6, 2011

    It’s Jerry-Rig, lol

The site is currently under maintenance and will be back shortly. New comments have been disabled during this time, please check back soon.