Japan Earthquake: Nuclear Plants at Fukushima Daiichi

[rmattila]

Safety would mean isolating each reactor, so that if a catastrophic failure occurs, you don't risk losing all the reactors.

Remember that the ability to use the power from unit 6:s EDG was what saved unit 5 at Fukushima.

 

[htf]

Please provide the evidence to support this claim. I don't believe that there was concern about profit in the Soviet Union.

There was concern about profit in the Soviet Union. The disease was not called "shareholder value" but it was the race between capitalism and socialism.

There are many sources that analyse the real reasons of the accident. In the beginning the operators were blamed. This is what I call pilot error mechanism. This is the most convenient way to explain an accident.

There is a GRS report that summarizes the real reasons:

http://www.grs.de/sites/default/files/pdf/GRS-121_Deut.pdf

The summary of reasons for the accident:

Aus heutiger Sicht waren die Hauptursachen des Unfalls
- gravierende Mängel der reaktorphysikalischen Auslegung und der Auslegung der
Abschalteinrichtungen,
- ein politisches und organisatorisches System, welches nicht in der Lage war, diese
Mängel abzustellen, obwohl sie lange vor dem Unfall bekannt waren,
- ein sicherheitstechnisch unzureichend durchdachtes und geprüftes Versuchsprogramm,
- eine Betriebsführung und Bedienungseinrichtungen, die das Personal bei der
Wahrnehmung seiner Verantwortung für die Sicherheit überforderten.

In chapter 2.3 discusses the influence of the reactor condition:

Bestimmt und verschärft wurde der Unfallablauf durch:
- Ungünstige Wahl des Zeitpunktes - hoher Abbrand mit sehr großem positiven Voideffekt
von mindestens 5 b - für die Durchführung des Versuchs
- Nichtbeachtung der Erfordernisse der Reaktorsicherheit bei Aufstellung des Versuchsprogramms
- geringe Erfahrungen und unzureichende Beteiligung des Betriebspersonals an den
Vorbereitungen des Versuchs sowie Verstöße des Betriebspersonals gegen Betriebsvorschriften

The time line of the accident can be found in wikipedia:

Conditions prior to the accident
Schematic diagram of reactor

The conditions to run the test were established before the day shift of 25 April 1986. The day shift workers had been instructed in advance and were familiar with the established procedures. A special team of electrical engineers was present to test the new voltage regulating system.[20] As planned, a gradual reduction in the output of the power unit was begun at 01:06 on 25 April, and the power level had reached 50% of its nominal 3200 MW thermal level by the beginning of the day shift. At this point, another regional power station unexpectedly went off line, and the Kiev electrical grid controller requested that the further reduction of Chernobyl's output be postponed, as power was needed to satisfy the peak evening demand. The Chernobyl plant director agreed and postponed the test.

At 23:04, the Kiev grid controller allowed the reactor shut-down to resume. This delay had some serious consequences: the day shift had long since departed, the evening shift was also preparing to leave, and the night shift would not take over until midnight, well into the job. According to plan, the test should have been finished during the day shift, and the night shift would only have had to maintain decay heat cooling systems in an otherwise shut down plant. The night shift had very limited time to prepare for and carry out the experiment. A further rapid reduction in the power level from 50% was executed during the shift change-over. Alexander Akimov was chief of the night shift, and Leonid Toptunov was the operator responsible for the reactor's operational regimen, including the movement of the control rods. Toptunov was a young engineer who had worked independently as a senior engineer for approximately three months.[14]:36–8

 

[robinson]

Remember that the ability to use the power from unit 6:s EDG was what saved unit 5 at Fukushima.

That brings up an interesting thought. It seems that even a reactor that isn't running, that is in cold shutdown, is also an extremely dangerous thing to even exists. The loss of cooling, for any reason, and the inability to restore it very quickly, means big trouble and lots of it.

There is little doubt that natural disasters will come, and this disastrous situation will happen again. It's built into the very nature of a reactor, or a spent fuel pond. With out capable trained people, a support system, water, electricity and constant attention, any reactor/fuel pond is just a disaster waiting to happen.

 

[jim hardy]

htf are you fluent in the language of those gts.de links?

There's something I've wondered about for over twenty five years - Perhaps you'd know.

""A special team of electrical engineers was present to test the new voltage regulating system.[20] ""

Early reports said purpose of the "test" was to try and keep power to some new piece of equipment after turbine trip. To that end they disabled reactor trip when turbine trip occurs from low power... that left reactor making power with noplace to send the heat so it began to boil. In that obsolete RMBK (scaled up US 1940's design stolen by spies in Manhattan project), boiling causes runaway.

Do those articles say what was the piece of equipment they were testing? That line quoted above is first description I've ever seen. "Voltage regulating system..." - voltage for what ?? Main generator? Plant computer?

thanks, old jim

 

[Astronuc]

Safety would mean isolating each reactor, so that if a catastrophic failure occurs, you don't risk losing all the reactors.

Impact of one reactor/unit on other units is a consideration in NPP design.

If properly designed, one should have a catastrophic failure of a single unit. As pointed out earlier, there are mandatory General Design Criteria that quite clearly emphasize the safety aspects.

http://www.nrc.gov/reading-rm/doc-collections/cfr/part050/
http://www.nrc.gov/reading-rm/doc-collections/cfr/part050/part050-appa.html

Standard Review Plan for the Review of Safety Analysis Reports for Nuclear Power Plants: LWR Edition
http://www.nrc.gov/reading-rm/doc-collections/nuregs/staff/sr0800/

http://www.nrc.gov/reading-rm/doc-collections/reg-guides/power-reactors/rg/

There are thousands of pages of rules and regulations all aiming to ensure safety of power reactors.

The expectation is that the emergency core cooling system is operable, even if there is loss of off-site power. However, this assumes that the ECCS systems are not placed in vulnerable locations, and that the primary system and ancillary cooling systems are more or less intact.

Had the seawall been 15 m, i.e., had the tsunami been prevented from inundating the Fukushima Daiichi site, or given that, that the oil tanks been buried underground, or perhaps located west of the units, and had the diesel generators and switch gear been located in a more secure locations, then the site would most likely have survived. Similarly, had the operators vented earlier, had the duct work to the stacks held, and/or had the secondary containment been of a more robust design, then the consequences of the accident wouldn't have been as bad they were.

 

[alpi]

... Without capable trained people, a support system, water, electricity and constant attention, any reactor/fuel pond is just a disaster waiting to happen.

No. The fuel elements or what is left of them just have to be submerged in water and the lost water (boil-off) be replaced.

Edit: Lack of submerging in water wouldn't be dangerous if it was possible to cool the fuel elements by spraying water.

 

[joewein]

Remember that the ability to use the power from unit 6:s EDG was what saved unit 5 at Fukushima.

Good point and one that also occurred to me when I read the comment you replied to.

If twinned reactors share 3 or 4 EDGs or redundant grid connections as in the case of the F-1 units, if you were to instead build those reactors at isolated locations, would they still be built with the same number of redundant units? Of course one would hope so from a safety point of view, but the economic realities are such that there would be more pressure to cut corners and not double costs.

You would have to give each unit the full number of diesels, transformers and power lines that each pair now has, instead of being able to share via a local bus for redundancy.

 

[Astronuc]

Good point and one that also occurred to me when I read the comment you replied to.

If twinned reactors share 3 or 4 EDGs or redundant grid connections as in the case of the F-1 units, if you were to instead build those reactors at isolated locations, would they still be built with the same number of redundant units? Of course one would hope so from a safety point of view, but the economic realities are such that there would be more pressure to cut corners and not double costs.

You would have to give each unit the full number of diesels, transformers and power lines that each pair now has, instead of being able to share via a local bus for redundancy.

Still, one has to avoid common mode failure. Fukushima Daiichi - Units 1-4 were all affected by the same event - the tsunami. They all lost offsite power, a common element, and they all got their turbine buildings, EDGs and switch gear flooded. Having redundancy (which is common) is no good if the redundant systems have a common vulnerability. Redundant system must also have independence, usually through different locations and trains/routes.

For example, the United Airlines DC-10 (Flight 232) had a catastropic failure of it's tail engine. One of the compressor blades severed the hydraulic lines, which while redundant, were not sufficiently independent. All three sets were routed together such that a common failure mode was possible.
http://en.wikipedia.org/wiki/United_Airlines_Flight_232

Redundancy and independence of safety-related systems is mandatory.

 

[joewein]

htf are you fluent in the language of those gts.de links?

I guess if it was good enough for Einstein it's good enough for him ;-)

 

[rmattila]

Good point and one that also occurred to me when I read the comment you replied to.

If twinned reactors share 3 or 4 EDGs or redundant grid connections as in the case of the F-1 units, if you were to instead build those reactors at isolated locations, would they still be built with the same number of redundant units? Of course one would hope so from a safety point of view, but the economic realities are such that there would be more pressure to cut corners and not double costs.

You would have to give each unit the full number of diesels, transformers and power lines that each pair now has, instead of being able to share via a local bus for redundancy.

I've never really thought through the redundancy requirements concerning twin units. The way things are here in Finland is that a general N+2 requirement is applied for all first-line safety systems (maintenance/repair + random single failure), so each unit must have two own diesel generators in addition to the minimum number required for fulfilling the safety function. Building twin units does not relax this safety requirement, but the practise has been to enable cross-connections between the diesel busbars to provide additional defence-in-depth in case of a common-mode failure of more than the assumed number of diesels (=2 per unit). The cross-connections are included in the emergency operation procedures, and they are modeled in the plants' PRA models, but since all diesel generators are still similar, this cross-connection does not reduce the calculational risk very much (common-cause failure of 8 EDG:s is not very much more unlikely than CCF of 4 EDG:s), but it may be helpful if one unit is still able to provide power e.g. from its own turbine.

There are three important principles to be applied in the design of safety systems: 1: redundancy, 2: diversity and 3: separation. Redundancy works against random failures, diversity (=a whole alternative system to fulfil the safety function in case of a CCF) works against common cause failures (both internal such as faulty materials and external such as a flood), and in both cases, the redundant subsystems must be physically and functionally separated in order to limit the failure to one subsystem, and the diverse subsystems must be functionally independent from each other so that the diverse system can be trusted to be available in all cases where the first-line system has failed due to a CCF.

Af Fukushima, there was good redundancy (2+2 EDG:s per twin unit) and pretty good diversity (RCIC, isolation condensers), but an obvious shortcoming in both the physical separation of redundant systems (flood knocked out several EDGs) and in the functional separation of diverse systems (RCIC failed either after losing battery backup or after S/C heatup due to loss of the heat removal chain from the containment). That is what made the accident disastrous for the plant - what made it an environmental disaster was the lack of defence-in-depth preparations for a severe accident. This is largely due to the old containment design, which is apparently very difficult to backfit to contain the radioactivity in case of a severe accident.

 

[joewein]

I've never really thought through the redundancy requirements concerning twin units. The way things are here in Finland is that a general N+2 requirement is applied for all first-line safety systems (maintenance/repair + random single failure), so each unit must have two own diesel generators in addition to the minimum number required for fulfilling the safety function. Building twin units does not relax this safety requirement, but the practise has been to enable cross-connections between the diesel busbars to provide additional defence-in-depth in case of a common-mode failure of more than the assumed number of diesels (=2 per unit).

I remembered incorrectly when I talked about 3 or 4 shared EDGs per pair at F-1. Each of the units had two EDGs, except for unit 6 (a BWR5) which had three. So it was really 4 or 5 shared EDGs per pair if one counts the bus for power exchange.

There was some diversity in EDGs, as each pair had at least one-air cooled EDG (in units 2, 4 and 6), but the one in unit 6 was the lone survivor.

That is what made the accident disastrous for the plant - what made it an environmental disaster was the lack of defence-in-depth preparations for a severe accident. This is largely due to the old containment design, which is apparently very difficult to backfit to contain the radioactivity in case of a severe accident.

Even if they could not do much about the Mark 1 containments of unit 1-5 (and did not want to retire the units), and they knew the small containments would have trouble dealing with the large amount of hydrogen produced in a zirconium reaction on a prolonged station blackout, presumably they could have built some kind of "overflow containment" away from the unit for venting into instead of directly into the stack? At the very least they could have inserted some kind of wet scrubber before the stack, to remove as much Cs and I as possible without generating too much back-pressure. They weren't exactly short of space on-site, if you consider they're now setting up tanks for several 10,000 m3 of water and two separate water decontamination systems within walking distance of the wrecked units.

 

[rmattila]

I remembered incorrectly when I talked about 3 or 4 shared EDGs per pair at F-1. Each of the units had two EDGs, except for unit 6 (a BWR5) which had three. So it was really 4 or 5 shared EDGs per pair if one counts the bus for power exchange.

There was some diversity in EDGs, as each pair had at least one-air cooled EDG (in units 2, 4 and 6), but the one in unit 6 was the lone survivor.

In the old plants, it was common to have 2x100 % redundant safety systems, i.e. 2 EDGs per plant unit. This allows for a single failure, but does not enable scheduled maintenance or repair of broken components during operation. The newer plants typically have 3x100 % or 4x50 % systems, which gives more tolerance for preventive maintenance during operation (enabling shorter outages and ascertaining enough time to do the maintenance carefully, when it's not on the critical path) and allows repairing broken equipment without losing the single failure tolerance needed in case of an accident. I guess the interconnections between 2x100 % twin plant units might enable such repair/maintenance operations on one train of a twin plant (4 x 100 % EDG capacity when 2 x 100 % is needed) without sacrificing the single failure tolerance, if you could assume only one random failure per two plant units. My first instinct would be to say this isn't enough and you really need N+1 on each of the units separately, but as I said, I haven't ever had to really think the logic through for twin plant units with common systems.

And, again, this is from the Finnish point of view - I am not familiar with the American / Japanese thinking and requirements regarding the failure tolerance requirements in the old plants.

 

[rmattila]

Even if they could not do much about the Mark 1 containments of unit 1-5 (and did not want to retire the units), and they knew the small containments would have trouble dealing with the large amount of hydrogen produced in a zirconium reaction on a prolonged station blackout, presumably they could have built some kind of "overflow containment" away from the unit for venting into instead of directly into the stack? At the very least they could have inserted some kind of wet scrubber before the stack, to remove as much Cs and I as possible without generating too much back-pressure. They weren't exactly short of space on-site, if you consider they're now setting up tanks for several 10,000 m3 of water and two separate water decontamination systems within walking distance of the wrecked units.

The Swedes built a facility called FILTRA at their Barsebäck two-unit BWR site in the early 1980's. Here's a really thorough and well-written progress report of the project that resulted into a 10 000 m3 gravel bed being built next to the the units. That might be one approach to improve the capacity of old containments; however, it won't remove the problems related to preventing core-concrete interactions if a molten core falls to the bottom of the containment.

EDIT: Photograph of the Barsebäck site, with the FILTRA facility on the foreground.

 

[htf]

htf are you fluent in the language of those gts.de links?

I think so - but don't ask my German teachers.

There's something I've wondered about for over twenty five years - Perhaps you'd know.

""A special team of electrical engineers was present to test the new voltage regulating system.[20] ""

The report says that the test was classified as an electro-technical test because no impact on the reactor was expected. The test was planed by an electrical engineer because the electrical systems should be tested. The test should demonstrate that a loss-of-coolant accidents with simultaneous loss of normal power supply can be mastered. Therefore it had to be demonstrated that the main feedwater pumps can be powered by the the rotational energy of the down spinning turbo generator. (I remeber a report that said it was a timespan of ~1 minute that had to be bridged until emergency power is available). This test is part of the initial startup procedures and required by Soviet regulations. The test failed and Chernobyl #4 was nevertheless put into operation in December 1983. It was decided to postpone the test and repeat it during next regular shutdown. Meanwhile modifications to the generator were made - part of these modifications was the new voltage regulating system that should be tested by a team of electrical engineers.

I hope this helps.

 

[Most Curious]

Sadly, the Chernobyl engineer in charge ordered the ECCS locked out - literally padlocked - to preserve test integrity!

The up and down changes in reactor power were against written safety requirement and resulted in nearly ALL of the control rods out of the core to attempt power increase. They got their power increase allright, but it was WAY too much and WAY too fast.

The RBMK has a positive void and positive temperature coefficient making control at low power levels VERY difficult and unstable.

The 1st few feet of the control rods were graphite rather than absorber material so a SCRAM or even orderly insertion of control rods INCREASED power until the absorber portion of the rod was in the core.

All of the above and a few more items made the RBMK a ticking time bomb, just waiting for a few operator errors to cause a disaster.

 

[joewein]

The Swedes built a facility called FILTRA at their Barsebäck two-unit BWR site in the early 1980's. Here's a really thorough and well-written progress report of the project that resulted into a 10 000 m3 gravel bed being built next to the the units. That might be one approach to improve the capacity of old containments; however, it won't remove the problems related to preventing core-concrete interactions if a molten core falls to the bottom of the containment.

Thank you, rmattila! That was a very inspiring read. Tepco had almost 30 years to follow the Swedish example, but didn't.

Perhaps the Swedes had an extra incentive to think about cleaning up venting gases in case of a melt-down. They had built the Barsebäck BWRs a mere 20 km from Copenhagen, the Danish capital (metropolitan population: 1.9M), and the Danes kept calling for its shutdown. I found a study ("http://130.226.56.153/rispubl/reports/ris-r-462.pdf"") published just around the time the Swedes decided to build their filter system.

I think that, had there been an effective filter system for the containment venting at Fukushima Daiichi, the reluctance to vent the containment sooner, while there was still less pressure and less hydrogen would not have been as great.

 

[I_P]

This document appears to indicate the operators aligned the venting from both drywell AND S\C. No scrubbing from the drywell venting.

The document also indicates very high dose rates in the buildings and onsite well before the venting even took place. To my laybrain that seems odd - was there containment failure before they even got to vent?

source : http://www.tepco.co.jp/en/press/corp-com/release/betu11_e/images/110618e15.pdf"


On the SGTS\HVS contamination - why the high dose rates in the Unit 1 Turbine building early on? And why does SGTS even go into the turbine building? Why does the SGTS appear to be HEAVILY contaminated, it shouldn't have even been possible for it to be working after loss of power. So SGTS just opens itself up on loss of power? WTF. The more I read about the design of these systems the less I want to know, kind of.

Initial reports from workers at the plant described considerable damage resulted from the earthquake, prior to the tsunami. Recent interviews with workers present at the time bolster this. Radiation alarms were tripped at the #1 reactor prior to the tsunami. It is believed by a number of Japanese experts that cooling systems failed initially due to earthquake damage - breaks in the primary cooling system that initiated emergency spraying systems to kick in.

TEPCO has considerable stake in claiming that the meltdowns were entirely the result of an unforeseeable, rare event (huge tsunami) rather than being initiated by earthquake damage from shaking that was within or just barely exceeded the plant design basis.

 

[nikkkom]

The 1st few feet of the control rods were graphite rather than absorber material so a SCRAM or even orderly insertion of control rods INCREASED power until the absorber portion of the rod was in the core.

This is a simplification. Only many years later I found a more detailed description of the graphite tip problem. Here it is:

RBMK reactor's active zone is 7 meters high. Therefore, control rods' absorbing section is 7 meters high too. Since control rods sit in water filled vertical tubes, if rod would have only absorbing section (without any additional tips below it), with the rod in fully retracted position the whole tube will be water-filled. Water absorbs neutrons, which is bad for neutron balance. Therefore, adding a 7 meter long graphite tip below absorbing section improves neutron balance: graphite doesn't absorb neutrons (or more precisely, it absorbs them less than water).

However, this requires an additional 7 meter long tube extension under reactor's active zone, in order for the tip to have a place to move into when control rod is fully inserted. This in turn will require deeper basement etc. IOW: it will cost money.

So designers decided to make tips shorter than 7 meters. Say, 5 meters. Which means that if all control rods are fully retracted, the lowest 2 meters of control rods' tubes are water filled.

If in this configuration all rods are lowered simultaneously, 5 meter long graphite tips (which are already in active zone, in its 5 upper meters) will start moving down, and in this lowest 2 meters of active zone neutron absorption will be reduced, and reactivity will increase.

If it so happens that reactor's local reactivity in its lower part is already higher than average, then this reactivity spike can be the last feather on the camel's back.

 

[Cire]

TEPCO has considerable stake in claiming that the meltdowns were entirely the result of an unforeseeable, rare event (huge tsunami) rather than being initiated by earthquake damage from shaking that was within or just barely exceeded the plant design basis.

The anti-nuclear community have considerable stake in claiming that the meltdowns were entirely the result of a foreseeable, common event.

You don't have a primary cooling failure and also need to vent the RPV due to excessive pressures 12,24,48 hours later.

 

[Morbius]

Sadly, the Chernobyl engineer in charge ordered the ECCS locked out - literally padlocked - to preserve test integrity!

The up and down changes in reactor power were against written safety requirement and resulted in nearly ALL of the control rods out of the core to attempt power increase. They got their power increase allright, but it was WAY too much and WAY too fast.

The RBMK has a positive void and positive temperature coefficient making control at low power levels VERY difficult and unstable.

The 1st few feet of the control rods were graphite rather than absorber material so a SCRAM or even orderly insertion of control rods INCREASED power until the absorber portion of the rod was in the core.

All of the above and a few more items made the RBMK a ticking time bomb, just waiting for a few operator errors to cause a disaster.

Most Curious,

Add to that, that they conducted their experiment in the middle of a Xenon Transient.

The operators had greatly reduced the power in preparation for their experiment, when the load controller in Kiev called and asked that Chernobyl Unit 4 remain online at the reduced power, since they really needed the capacity. It was several hours later that the load controller released the plant to go offline.

That put them right in the middle of the Xenon Transient from the initial lowering of reactor power. The reason they had so many control rods out was because they were attempting to compensate for the parasitic neutron capture due to Xenon-135.

However, if you get the reactor critical, and you start burning the Xenon-135, then you are burning away a neutron poison, so that increases reactivity, increases power, hence burning more Xenon-135...and you have yourself a positive feedback loop.

The fact that they conducted this experiment in the middle of the Xenon Transient is a real big factor in the cause of the Chernobyl disaster.

Greg

 

[Most Curious]

Thank you Greg! I had forgotten what they called that, an "Iodine" or "xenon well", IIRC - what you probably more correctly called a "Xenon Transient". In any event, there were procedures in the manual that forbid them trying to overcome the "poison" problem so rapidly. Further, the reactivity reserve mandated a minimum number of rods fully inserted in the core at all times - which they violated trying to burn up the Xenon.

Like almost ALL major accidents - nuclear, aircraft or others - a whole series of events pile up, leading to disaster. At Chernobyl, they had a poorly designed test, foolish cheif engineer (pad lock ECCS?), at least three positive feedback loops in the reactor - just to name a FEW. No wonder they lost control of it!

Fukushima is similar, in my uneducated opinion. Some questionable site design features, maybe an operator error or two, an unforgiving small containment, HUGE Earth quake and enormous tusnami not to mention political issues (Permission to vent before the containment ruptures?). Point being, no single even did them in but rather a SERIES of bad events and beyond design basis problems. In some ways I am in awe the situation did not turn out even worse than it has so far.

 

[jim hardy]

""Like almost ALL major accidents - nuclear, aircraft or others - a whole series of events pile up, leading to disaster.""

Bumping up against the philosophical thread now...



Ernest Gann's classic book "Fate is the Hunter" made that a theme, more so in early edition.

In my life as a plant troubleshooter it is exactly the truth - little things stack up like dominoes and eventually something pushes the first domino. Often it's a trivial initiating event and that's how the small things of the Earth confound the mighty.

Good maintenance consists of not letting the dominoes stack up. Japanese are very good at maintenance in fact our industry sent people over there to study how they achieved such good reliability.

The major dominoes there were placement of electrical equipment in basements on ocean side of plant, and failing to act in 1990's when better information regarding tidal waves came to light. I believe one personal letter from plant workers to an executive would have changed the outcome.

old jim

 

[jim hardy]

Thanks htf that does help.

powering the feed pumps briefly would help the turbine avoid overspeed on loss of grid, so it does make mechanical sense



old jim

 

[htf]

My understanding of the test is that not overspeed of the turbine should be avoided but to demonstrate that the rotational energy of the down-spinning turbine can be used to power the main feedwater pumps. These pumps are needed to cool the reactor during a loss-of-coolant accident in an initial phase (says the GRS report). The scenario hence was loss of normal power supply during an loss-of-coolant accident. It takes a fairly long time to bring up the emergency power system (almost 1 minute according to an other report) - presumably to long for the twofold fault situation. Therefore they wanted to show that the spinning turbines can serve as an emergency power system that is immediately available. This test had failed during entry-into-service.

Rumours say that the reason why the Soviets considered this fault scenario was the Israeli attack on the reactor in Iraq in 1981.

 

[rmattila]

As a side note, low startup speed of gas turbines has been one reason why emergency power for safety systems at NPP:s is usually arranged with diesel generators. Typically, you can start an EDG within 10 s or so, while a gas turbine may take minutes to start up.

The http://www.mhi.co.jp/atom/hq/atome_e/apwr/index.html by Mitsubishi Heavy Industries has some design features that according to its designer make emergency diesel generators unnecessary: the plant allows for a longer period of time before power is needed in case of a LOCA, and this combined with a emergency gas turbine designed for a very short startup time (of the order of half a minute or so) are said to fulfil all safety criteria without any emergency diesel generators. As far as I understand, from maintenance point of view gas turbines would be a tempting option as NPP emergency power generation units.

 

[tsutsuji]

http://www.tepco.co.jp/en/nu/fukushima-np/images/handouts_110811_01-e.pdf (11 August) "Leakage point from flexible hose in circulating cooling device for Unit 4 Spent Fuel Pool"

http://www.tepco.co.jp/en/nu/fukushima-np/images/handouts_110814_02-e.pdf A magnitude 6 earthquake occurred at 3:22 AM on 12 August off the Fukushima prefecture shore. This press release explains the consequences on the Fukushima Daiichi plant: boiler stops at the desalination facility, injection rate into unit 1 reactor declined to 3.2 m³/hour, and one air-control compressor breaks down at unit 1 at 5:06 AM. The small leak at SFP4 cooling system was found at 5:27 AM.

http://www.jiji.com/jc/c?g=soc_30&k=2011081200844 SFP1 temperature reached 39.5°C at 11:00 AM on 12 August, down from 47°C on 10 August when the SFP cooling system was started. Two more leakage points were found at the SFP4 cooling system, bringing the total number of leakage points to 4. Each leaked quantity is small, like bleeding. Damaged hoses will be replaced. An alarm rang at 6:15 PM on 12 August, and the water treatment facility stopped. As no abnormality was found, it was started again at 11:30 PM. A wrong alarm temporarily shut down the facility on 11 August too.

Fukushima Daiichi unit 1:

At 7:36 pm on August 13, we adjusted the rate of water injection through
reactor feed water system piping arrangement to approximately 3.8m3/h as
we confirmed decrease in the amount of water injection to the reactor.
http://www.tepco.co.jp/en/press/corp-com/release/11081401-e.html

http://sankei.jp.msn.com/affairs/news/110814/dst11081413210003-n1.htm In the morning of 13 August, 6 tons of sodium carbonate (a chemical agent preventing foreign bodies from sticking to the pipes) leaked at an evaporation equipment in the desalination facility. One of the two desalination systems was stopped, bringing down the desalination capacity by one half. It will be started again on 15 August. The cause could be the hose band being loosely fastened, or the rise of temperature inside the tent.

http://www.tepco.co.jp/en/nu/fukushima-np/images/handouts_110814_01-e.pdf Pictures of the unplugged sodium carbonate hose and subsequent repairs.

http://cryptome.org/eyeball/daiichi-npp16/daiichi-photos16.htm Another set of pictures of Fukushima Daiichi, including "Flood in Electric Equipment Room of Unit 6 (pictured on March 17, 2011)" and "Setting work of submersible pump (pictured on March 17, 2011)" in front of unit 5, which I had not seen before.

http://www.tepco.co.jp/cc/press/betu11_j/images/110812d.pdf Report in Japanese about worker exposure at Fukushima Daiichi. The positions (green, orange, pink circles) and movements (red and black arrows) of the workers inside control rooms and the direction of the wind (blue arrows) are shown on the maps on page 53 (units 1&2 control room) and 54 (units 3&4 control room) (pdf page numbers). Pages 35, 36, 52, 53 provide detailed timelines of the tasks performed by four workers named "C", "D", "E", and "F". The table on page 40 describes the exposure circumstances for twenty workers (A ~ F, ア ~　セ). The column 1 on the left is their internal contamination in mSv, column 2 says if they wore a mask, column 3 if they ate or drank(有=yes, 無=no), column 4 if they wore glasses (temples may create an interstice through which contaminated air can leak), column 5 if they worked near the door. The table page 42 provides the radiations in cpm measured at units 3&4 control room on 13 March from 10:00 AM to 01:30 PM. Column 1 (on the left) at the front door, column 2 at the emergency door, column 3 at the desk unit 3 side, column 4 at the desk unit 4 side.

http://www.tepco.co.jp/cc/press/betu11_j/images/110812b.pdf Very big (767 pages / 55.1 MB) report to NISA, in Japanese, about the impact of the 11 March earthquake on Fukushima Daini.

http://mainichi.jp/select/jiken/news/20110607ddm003040107000c.html :

It was discovered that the 13 km long Yunotake fault which runs in Iwaki city 40 km south of Fukushima Daini was activated by aftershocks of the 11 March earthquake. The problem is that this fault had been overlooked in past earthquake safety designs. NISA instructs all NPP operators to review their earthquake safety assessments to ensure similar faults elsewhere are not being overlooked.

http://www.asahi.com/national/update/0812/OSK201108120219.html Tsuruga nuclear power plant units 1 & 2 are located on a crush zone thought to be a normal fault resulting from horizontal stretching forces, which differs from the reverse faults, which result from compressive forces. The probablility that a normal fault causes an earthquake or is activated by an earthquake was thought to be low. However, the magnitude 7 aftershock of 11 April 2011 activated such a normal fault, the Idosawa fault. The head of the Geographical Survey Institute's Kanto regional survey department, Mr Hiroshi Une, who is also a member of the NISA's working committee for the re-examination of Tsuruga power plant's earthquake safety, says that although Japan's normal faults are not supposed to move, since the 11 March earthquake the Earth's crust is subject to forces which are different from those observed in the past, and although there are crush zones all over Japan, the Tsuruga power plant is a case needing special attention because an active fault (the Urasoko fault, last activated 4000 years ago) is running inside the plant premises. Japco will announce its conclusions by the end of August. (The orange lines on the map are the active faults. The grey lines are the crush zones).

http://www.chunichi.co.jp/s/article/2011081290085007.html It has been found that just below Tsuruga nuclear power plant's reactors, faults called "crush zones" could move under the influence of the Urasoko active fault. Crush zones were previously thought as having "no activity", and they were not taken into account in the nuclear plant's earthquake safety design, but it was discovered that in the Great Eastern Japan Earthquake, this kind of fault had moved. Thinking the consequences for the plant again, Japco will disclose its opinion on the matter by the end of August. Hiroshi Une said: "the commonly held opinion that normal faults don't move has collapsed". Fast breeding reactor Monju is close to the Shiraki-Nyu active fault, and crush zones of the normal fault type were confirmed below the reactor. Tectonic geomorphology professor Mitsuhisa Watanabe of Toyo University says : "however robust a reactor is made, if the ground tilts, it will get broken. Keeping normal faults out of one's thought was a mistake and that must be revised".

 

[zapperzero]

http://www.tepco.co.jp/en/nu/fukushima-np/images/handouts_110814_02-e.pdf A magnitude 6 earthquake occurred at 3:22 AM on 12 August off the Fukushima prefecture shore. This press release explains the consequences on the Fukushima Daiichi plant: boiler stops at the desalination facility, injection rate into unit 1 reactor declined to 3.2 m³/hour, and one air-control compressor breaks down at unit 1 at 5:06 AM. The small leak at SFP4 cooling system was found at 5:27 AM.

http://www.chunichi.co.jp/s/article/2011081290085007.html It has been found that just below Tsuruga nuclear power plant's reactors, faults called "crush zones" could move under the influence of the Urasoko active fault. Crush zones were previously thought as having "no activity", and they were not taken into account in the nuclear plant's earthquake safety design, but it was discovered that in the Great Eastern Japan Earthquake, this kind of fault had moved. Thinking the consequences for the plant again, Japco will disclose its opinion on the matter by the end of August. Hiroshi Une said: "the commonly held opinion that normal faults don't move has collapsed". Fast breeding reactor Monju is close to the Shiraki-Nyu active fault, and crush zones of the normal fault type were confirmed below the reactor. Tectonic geomorphology professor Mitsuhisa Watanabe of Toyo University says : "however robust a reactor is made, if the ground tilts, it will get broken. Keeping normal faults out of one's thought was a mistake and that must be revised".

Thanks, again, for your excellent and ongoing effort to bring out the news.

Is it possible that the geophysicists were right? Maybe crush zones did not move in the past, but are moving now?

 

[zapperzero]

Speaking of news. TEPCO says tritium was detected in intake water canal and sub-drain of unit 2, on June 13.

http://www.tepco.co.jp/en/press/corp-com/release/betu11_e/images/110624e10.pdf
http://www.tepco.co.jp/en/press/corp-com/release/betu11_e/images/110624e7.pdf

H/T the tireless ex-skf

 

[NUCENG]

Speaking of news. TEPCO says tritium was detected in intake water canal and sub-drain of unit 2, on June 13.

http://www.tepco.co.jp/en/press/corp-com/release/betu11_e/images/110624e10.pdf
http://www.tepco.co.jp/en/press/corp-com/release/betu11_e/images/110624e7.pdf

H/T the tireless ex-skf

I have not seen any recent data that showed tritium in the sub-drains. The data you referenced was 7 weeks old and was less than 1% of the regulatory limit for tritium then. As I understand the samples of the sub-drains are made in water that is being contained for processing so there is no significant release ongoing. What point are you trying to make?

 

[joewein]

H3 is something that's released from the reactors even during normal operations as it's very difficult to separate from regular hydrogen in cooling water at low concentrations. It's produced from boron in the reactor and from neutron activation of deuterium in the coolant. Not sure how the quantities compare here though.

See:
http://www.nrc.gov/reactors/operating/ops-experience/tritium/faqs.html#normal

 

[joewein]

Moving the discussion from https://www.physicsforums.com/showthread.php?p=3452884#post3452884" to the main thread:

My recollection from the documentation published by TEPCO is that there would have been enough water on the secondary side to boil off the decay heat for 90 minutes, which does sound rather a short time window. Still, a lot can be accomplished in that time, if there are practised EOPs in place. It would be really interesting to hear the details why they failed to restart the IC after the tsunami, as the IC is something many of the the new BWR designs rely heavily upon in cases of emergency.

It appears the initial shutdown of the IC was prompted by the rapid fall in temperature after the IC was activated. There was a spec for the maximum drop in temperature per hour. It sounds there was concern about stressing the steel of the RPV:

"It is possible that a worker may have manually closed the valve (of the isolation condenser) to prevent a rapid decrease in temperature, as is stipulated by a reactor operating guideline," Tepco spokesman Hajime Motojuku told The Japan Times.

A worker may have stopped the condenser to keep cold water from coming into contact with the hot steel of the reactor to prevent it from being damaged.

However, nuclear reactors are designed to withstand this procedure in case of an emergency, said Hiromi Ogawa, a former nuclear plant engineer at Toshiba Corp.

According to Tepco, the isolation condenser's valve was confirmed open at 6:10 p.m. March 11 but it is unknown whether it was open between 3 p.m. and 6:10 p.m.

The valve was confirmed closed at 6:25 p.m. and confirmed open again at 9:30 p.m. Finally, the condenser was shut down due to a pump malfunction at 1:48 a.m. March 12, roughly eight hours after the tsunami, matching the battery life of the isolation condenser.

(http://search.japantimes.co.jp/cgi-bin/nn20110517x1.html" )

There was concern about the steel getting brittle from neutron bombardment over the years and unit 1 was the oldest at 40 years and rapid changes in temperatures is something brittle materials don't handle well (think of glass).

I reckon you'll be interested in http://www.tepco.co.jp/en/press/corp-com/release/betu11_e/images/110618e15.pdf", on the logs and testimony of operator response during the first days after the earthquake.

From the available evidence and testimony, the operators do in fact appear to have opted for the use of just one of the IC systems (the 'A' system) for the control of reactor pressure, judging it to be sufficient to keep the vessel at 6-7MPa, while they initially relied on the HPCI system for the control of the reactor water level.

In that Tepco document it also says they understood the water level of unit 1 at 21:19, 11 minutes before IC was confirmed active again.

Earlier on, after they had lost EDG power it says "Operaters judged HPCI was not operable because indicators on the control panel were gradually faded." I don't understand why that would be, as HPCI should work on steam pressure and batteries.

Another concern I have is how the IC would behave once the core gets uncovered and the zirconium cladding starts to burn, since then hydrogen would start flowing into it if the inlet valve is open, but that hydrogen could not be condensed. Somebody had explained that the (battery operated) valves at the bottom of the IC are usually closed until enough steam had been condensed to fill it. Then valve are opened, the water is let drain back to the RPV and after that the valves are closed again.

What would happen if uncondensible hydrogen prevented the IC from ever filling to the top, as the hydrogen displaced condensible steam, preventing it from coming into contact with cool IC water?

Tepco estimates that fuel was uncovered 5 hours after the quake, i.e. before 20:00 on 3/11. The valve would have been closed at the time. If it was opened at 21:30, that would have been 90 minutes into the exposed core overheating.

 

[rmattila]

It would be really interesting to hear the details why they failed to restart the IC after the tsunami,

It appears the initial shutdown of the IC was prompted by the rapid fall in temperature after the IC was activated. There was a spec for the maximum drop in temperature per hour. It sounds there was concern about stressing the steel of the RPV:

Controlling the cooldown rate of the RPV is a standard procedure when running the plant to cold shutdown, and it appears that the IC in Fukushima Dai-ichi unit 1 was so powerful that it needed shutting down in order to stay within TechSpecs cooldown rate. This all seems to have been nice and well after the earthquake, when the plant had experienced what then seemed an anticipated operational occurrence.

Things changed when the tsunami hit and knocked out the AC power supply. At this time, the plant situation degraded from an AOO to an accident, and I would imagine re-activating the IC would be among the first tasks instructed by the EOP at station blackout. However, the details of what actually happened at that time are somewhat unclear to me: did they attempt to restart the IC and if they did, why did it not prevent the core uncovery?

 

[joewein]

I would imagine re-activating the IC would be among the first tasks instructed by the EOP at station blackout. However, the details of what actually happened at that time are somewhat unclear to me: did they attempt to restart the IC and if they did, why did it not prevent the core uncovery?

The report says the valve status of the isolation return valves was not indicated on the control panel. At 15:50 they lost power to the instruments and no longer knew the reactor water level either. If they know what was done to control the isolation condenser for the rest of the afternoon they're not telling us. However, the statement about not knowing the water level or the valve status could be interpreted as a way to excuse why perhaps the correct action was not taken.

As soon as HPCI was no longer available, the IC was the only thing left to prevent the core of unit 1 from boiling dry.

If there was enough water in the IC to last for only 90 minutes (can you still find the source for that), refilling the IC tank should have been a high priority.


This also reminds me again of the issue of running out of fresh water on site. When they later started pumping highly radioactive water from the flooded basement into the condenser tank (1600 m3) at unit 1, they then wanted to pump water from the condenser to the condenser storage tank (1900 m3), but found it to be full and had to empty that water into the suppression pool surge tanks first.

They also had 10,000 m3 of water previously used in primary cooling system circuits available at the Centralized Radioactive Waste Disposal Facility, which Tepco later dumped into the ocean to make space for highly radioactive water from the basements, saying that the radioactivity of the 10,000 tons equaled that of 10 liters of unit 2 basement water.

It sounds to me like perhaps there was water available that wasn't considered. What kind of a plan did they have?

 

[rmattila]

If there was enough water in the IC to last for only 90 minutes (can you still find the source for that),

http://pbadupws.nrc.gov/docs/ML0230/ML023010606.pdf, page 39/92:

Following a reactor isolation and scram, the energy added to the coolant will cause reactor pressure to increase and may initiate the isolation condenser. The capacity of this system is equivalent to the decay heat rate generation 5 minutes following the scram and isolation. With no makeup water, the volume of water stored in the isolation condenser will be depleted in 1 hour and 30 minutes. This allows sufficient time to initiate makeup water flow to the shell side of the condenser.

That appears to be a generic description of GE BWRs, so I don't know how it accurately it describes the 1F1 plant. 1F1 has rather small thermal power, so I guess the grace period could be somewhat longer.

 

[tsutsuji]

Although the NHK website was quite quiet in the past few days (because of Obon holiday?), we are having Fukushima-NPP-related NHK news again today:

http://www3.nhk.or.jp/news/genpatsu-fukushima/20110816/index.html SARRY has been started for a test run today. If everything is OK, this test will be performed until 17 August night, after which normal operation will ensue.

http://www3.nhk.or.jp/news/genpatsu-fukushima/20110816/1300_jisshi.html Tepco plans to install desalination systems and decontamination systems with zeolite at units 2,3,4 spent fuel pools [How about unit 1?]. The desalination system for SFP4, located on 5 truck platforms, using a special membrane and electricity [that must be reverse osmosis], will be installed by the end of this week and is expected to bring salt concentration to 1/25th in two month's time.

Tomari NPP unit 3, in Hokkaido, is going to resume commercial operation:

Nuclear reactors suspended for regular checkups need to undergo "stress tests" before resuming operations, but the central government has said that the case of the Tomari reactor is not a restart because the reactor is already activated
http://mdn.mainichi.jp/mdnnews/news/20110813p2g00m0dm010000c.html