Japan Earthquake: nuclear plants Fukushima part 2

[Hiddencamper]

I know what you are saying. I'm just saying what I heard yesterday from sandia labs and TEPCO. I didn't ask why they think that.

The discussion was that the models we use assume you have a big penetration failure and the fuel comes out rapidly and in a mostly molten liquid. Apparently the observations don't line up with that model.

 

[Hiddencamper]

http://ndf-forum.com/ref/d2_mizokami_en.pdf

A pretty detailed report about the status of units 1-3

 

[Charles Smalls]

Something I wrote in my spalled concrete post back in March:

This is the part that doesn't add up. On the video it looks like they basically lower the camera all the way down until it hits the surface of the sediment. If it was a loose amalgamation of materials as spalled concrete normally is, I would have expected this impact to have thrown some of the sediment into suspension and clouded the water. It could be that the material is too heavy to be dislodged by such a light impact or maybe it formed a depression which the camera didn't pick up. I would be very interested to learn more on the make up of the sand/sediment i.e. particle sizes and whether it is loose material or a solidified mass.

Page 19 of Hiddencampers new July Tepco report:

I'm glad to see that mankinds instinctive investigative technique of: "hit it and see what happens" is still contributing to science lol. Kidding aside though, this looks like a great document with tons of information in it. Thanks for sharing HC

 

[Sotan]

That was an awesome find Hiddencamper - thank you!

Here are more links to various presentations made by participants to the 2nd International Forum on the Decommissioning of the Fukushima Daiichi Nuclear Power Station that took place on 2-3 July 2017:
http://www.ndf-forum.com/program_en/day1.html
http://www.ndf-forum.com/program_en/day2.html

I can;t see those on my phone now but definitely plan to download them when I get home.

 

[jim hardy]

http://www.ndf-forum.com/program_en/day1.html

that one gives me "error 404 not found" .

Maybe its not accessible from US ?

 

[Sotan]

I'm sorry, my bad. The first link should have been simply http://www.ndf-forum.com/program_en.html

 

[Hiddencamper]

Another few interesting tidbits.

Sandia labs did a spent fuel pool fire study that concluded days before Fukushima.

In this test, they had 5 fuel bundles in high density fuel racks. No actual fuel, just the rods. The center rod had electric heaters set to simulate decay heat, the other four had no heat. The center bundle started a zirconium reaction and ignited, the bundles adjacent with zero decay heat then ignited as well. A single hot bundle can ignite the entire fuel pool. This was provided to the NRC and is what led to heightened emotions about unit 4. Sandia told us that even when they tried to inert the test chamber with nitrogen or argon it didn't stop the fire. The zirconium was reacting with the nitrogen. The argon couldn't purge it enough. They tried putting sand on it, and it melted then the zirconium sucked the oxygen molecules out of the sand to keep burning. With no water injection It took several days before the Reaction rate lowered so that they could open the test chamber back up.

This is in NUREG-2161 (might be CR/2161)

Another interesting thing I learned: if fuel overheats, its recommended to not try to quench unless you have a lot of injection. If injection is too low you just feed the zirconium reaction and can't cool enough to beat it. Emergency procedures are going to be revised to try and anticipate blowing down the reactor early to prevent the fuel from being uncovered if you don't have sufficient quenching capability, rather than waiting until the fuel is uncovered and at 1800 degF.

 

[Astronuc]

A single hot bundle can ignite the entire fuel pool. This was provided to the NRC and is what led to heightened emotions about unit 4. Sandia told us that even when they tried to inert the test chamber with nitrogen or argon it didn't stop the fire. The zirconium was reacting with the nitrogen. The argon couldn't purge it enough. They tried putting sand on it, and it melted then the zirconium sucked the oxygen molecules out of the sand to keep burning. With no water injection It took several days before the Reaction rate lowered so that they could open the test chamber back up.

I wonder if they looked at the Ellingham diagrams and thermodynamic stability, of which Zr to ZrN and Zr to ZrO₂ are among the most energetic in their class, with the exception of Y/Yb to Y/Yb₂O₃ (and other lanthanides/actinides), if the passive oxide/nitride breaks down. Pyrophoricity is a tricky subject to those who don't understand the physics.

This is in NUREG-2161 (might be CR/2161)

It is NUREG-2161 (internal to NRC, i.e., not a contractor report (CR))
https://www.nrc.gov/docs/ML1425/ML14255A365.pdf

 

[MadderDoc]

This was provided to the NRC and is what led to heightened emotions about unit 4.

One can understand how it could have ignited thoughts of a nightmare scenario, when reports ran in about a hydrogen explosion and fires on and off in the unit 4 building, while the spent fuel pool just happened to have heated up sufficiently to start visibly steaming at the same time. And that,right in the swift succession of the shocking events in the other three units.

 

[turi]

There's a whole bunch of new press releases from Tepco. I won't find time to read them in detail at the moment. Interesting documents:
http://www.tepco.co.jp/nu/fukushima-np/handouts/2017/images2/handouts_170830_02-j.pdfAbout water in the reactor buildings.
http://www.tepco.co.jp/nu/fukushima-np/handouts/2017/images2/handouts_170830_03-j.pdfCurrent status of groundwater issues.
http://www.tepco.co.jp/nu/fukushima-np/handouts/2017/images2/handouts_170830_05-j.pdfEarthquake/tsunami countermeasures. They plan on removing the exhaust stacks among other things.
http://www.tepco.co.jp/nu/fukushima-np/handouts/2017/images2/handouts_170830_06-j.pdf
http://www.tepco.co.jp/nu/fukushima-np/handouts/2017/images2/handouts_170830_07-j.pdf
http://www.tepco.co.jp/nu/fukushima-np/handouts/2017/images2/handouts_170830_08-j.pdf

 

[MadderDoc]

Another interesting thing I learned: if fuel overheats, its recommended to not try to quench unless you have a lot of injection. If injection is too low you just feed the zirconium reaction and can't cool enough to beat it.

According to the recent report which was posted here, Tepco think that a zirconium water reaction occurred in the core of unit 3, with a heat output that maxed at 177 MW, during the morning of March 13 in 2011. BOE, that is an impressive burn rate, consuming about 1.6 tons of zirconium/minute. Makes me wonder how much Zr was present in total in the core.

 

[jim hardy]

According to the recent report which was posted here, Tepco think that a zirconium water reaction occurred in the core of unit 3,with a heat output that maxed at 177 MW,

That's 604 million BTU/hour , quite a fire .
Help me find that report ? I don't remember seeing it - but i miss a lot. Ones immediately above seem to be printed in Japanese.

 

[Hiddencamper]

That's 604 million BTU/hour , quite a fire .
Help me find that report ? I don't remember seeing it - but i miss a lot. Ones immediately above seem to be printed in Japanese.

According to the recent report which was posted here, Tepco think that a zirconium water reaction occurred in the core of unit 3, with a heat output that maxed at 177 MW, during the morning of March 13 in 2011. BOE, that is an impressive burn rate, consuming about 1.6 tons of zirconium/minute. Makes me wonder how much Zr was present in total in the core.

I can't tell you exact power levels. But I do know that we learned if you try to add water to fuel that's already deep in the zirconium reaction you just further fuel the reaction. Reactor and containment Pressure increased at unit 2 as water was added. We know that it gets worse before it gets better, but there is a minimum injection rate needed for quenching. The exact number is unknown but a good rule of thumb is 10% power worth of injection for a quick quench without excessively fueling the hydrogen reaction, and not much more than about 1 ECCS pumps worth if the fuel is molten (5500 gpm) to avoid adding too much water. Of course the core is still mostly or all in place you want to inject with whatever you have but if it's not you want to avoid a big pressure spike.

The BWR severe accident guidelines and EOPs are being revised to account for this info.

 

[MadderDoc]

That's 604 million BTU/hour , quite a fire .
Help me find that report ?

Here is the link to the report http://ndf-forum.com/ref/d2_mizokami_en.pdf

I am looking at the figure on page 43 of the report. The figure assumes the Zr-water reaction to have occurred mainly at about 9:00, and to a smaller extent at 12:00, in the morning of March 13 2011. The heat output of the reaction is shown as the purple double peak. Eye-balling the area of the peaks, to produce this added 'chemical reactivity' into the core during this period would have needed something like 60 tonnes of Zr to have been transformed by the reaction into about 80 tonnes of ZrO2, along with a production of about 2.5 tonnes of hydrogen. Seeing that it is the only purple peaks drawn into the figure, that would seem to suggest basically, that Tepco think the complete Zr inventory of the core was oxidated to ZrO2 and all H2 produced by the reaction had ben produced already by noon on March 13. (That would be ~ 23 hours before this produced hydrogen supposedly fueled an explosion inside unit 3, and ~ 40 hours before it would fuel another explosion inside unit 4.)

 

[jim hardy]

Thanks Doc
i had seen that graph but was suffering information overload... Will look further now .

old jim

 

[MadderDoc]

Of course the core is still mostly or all in place you want to inject with whatever you have but if it's not you want to avoid a big pressure spike.

I can understand that. You'd be risking injection of water into a system in which water(steam) availability had been rate limiting for the exothermic Zr-water reaction, and seeing that reaction produces hydrogen, pressure could rise dramatically.

 

[jim hardy]

i've only added two lines to your picture to help me visualize the time frame

wetwell vent(purple) and explosion(red)

Stolfi's plot of pressure from data available at the time( i think they were manually logging it from gages, recall plant conditions then)
snipped from here http://www.ic.unicamp.br/~stolfi/EXPORT/projects/fukushima/plots/cur/out/plot-un3-t-I-full.png
shows quite a pressure rise starting shortly after midnight on the 13th , about the time injection stopped.
I retrieved it because it shows wider range.
Caution log scale takes some getting used to but is great... He's a professor after all.

indeed it shows (red dots) pressure rose a lot

from the June 2011 gov't report to IAEA, for times:

the HPCI stopped at 2:42 on March 13. The reason for that appears to be a drop of pressure in
the reactor. The other probable cause could be water-vapor outflow from the HPCI system.
∙ (Status of the reactor core) The operation for injection of water containing boric acid
commenced using a fire extinguishing line at around 9:25 on March 13. However, the water
could not be injected sufficiently due to the high pressure in the reactor, and the water level in
the reactor lowered. As a result, water injection was halted at least for 6 hours and 43 minutes
after the HPCI stopped at 02:42 on March 13 until water injection using the fire extinguishing
line started at 09:25 on the same day.

It's ambiguous how much water the fire engines pumped in. Probably none until pressure got below ~300 psi , ~2000 kpa.
I don't know what made pressure drop so sharply around 9AM the morning of 13th but it does look to be same time as injection started via fire truck.

Could be that's when reactor vessel opened up to drywell - their pressures started tracking about 9:10
here's a snip of Stolfi's data from http://www.ic.unicamp.br/~stolfi/EXPORT/projects/fukushima/plots/cur/data/pres-un3-t.txt

and that great graph he made showing that they indeed coupled
http://www.ic.unicamp.br/~stolfi/EXPORT/projects/fukushima/plots/cur/out/pcor-PCA-PD-un3-full.png

.

Hard to believe it was seven years ago. I have forgotten a lot of what we analyzed back then.

BUt what's curious to me is -

IF there was 177 megawatts of heat being produced in that vessel,
and presumably tons of hydrogen gas as well...
Adding energy to gas usually raises its pressure...
THEN the time of heat production (purple in the NDF chart, first image in this post ) should correlate with the pressure rise around 3:30 AM, beginning of "without water injection" interval, not the pressure drop around 9 am at end of that interval ?

just one of those little physics things my pea brain has to resolve.

old jim

 

[MadderDoc]

Adding energy to gas usually raises its pressure...
THEN the time of heat production (purple in the NDF chart, first image in this post ) should correlate with the pressure rise around 3:30 AM, beginning of "without water injection" interval, not the pressure drop around 9 am at end of that interval ?

The RPV pressure rise in the early hours past midnight on March 13 correlates with the loss of water injection by HPCI. The pressure in the RPV then went up to be steady at the level where SRVs starts cycling, such as to release steam to the suppression pool, while consuming the water inventory of the RPV. That would seem to be the expected behaviour.

The RPV pressure drop around 9 am Tepco have thought might have been caused by the Automatic Decompression System's being (unintendedly) triggered. IOW, the abrupt RPV pressure drop at 9 am would then be explained by the ADS forcing the opening of several (6-8) SRVs, adding further to the loss of RPV water inventory at that time.

Timewise, the proposed Zr water reaction is put right up in the tail of these events; while the core was getting uncovered, its temperature would increase due to decay heat until the the temperature threshold to set the Zr-water reaction going was reached. Tepco think that happened in a major burst from just after 9 am, in very close timely connection to the RPV pressure drop, and that another portion of Zr went off again around 12:00 on the same day, each time adding heat and pressure to the system. The operators had recorders running during this period, and they did record three PCV/RPV pressure increases/spikes, at 9 am (major), at 10 am (relatively small), and at 12 am (minor). The recorded values are shown on page 47 in the report. As you note, the RPV and PCV pressures were coupled ever since whatever it was that happened inside the reactor at about 9 am that morning.

 

[MadderDoc]

i've only added two lines to your picture to help me visualize the time frame
wetwell vent(purple) and explosion(red)

Yeah, that's nice. That serves as orientation points in time on March 14. There were vents or such also earlier, on March 13, at about 10:00, 13:00, and 14:00. Steam fans from the stack from these vents can be seen in the hourly webcam images, and was captured also by a satelite photo shot at 9:50. If the suggested timing of the zirconium-water reaction in the core is correct, these vents would have drawn down the hydrogen inventory that had been produced by it. It would have reduced the amount of hydrogen available to the later explosions, to the extent the vents managed to get it out through the stack.

 

[jim hardy]

The RPV pressure drop around 9 am Tepco have thought might have been caused by the Automatic Decompression System's being (unintendedly) triggered. IOW, the abrupt RPV pressure drop at 9 am would then be explained by the ADS forcing the opening of several (6-8) SRVs, adding further to the loss of RPV water inventory at that time.

Thanks !
My simplistic thinking was depressurization was likely incore thimbles melting creating a vent path .

Your ADS seems a better fit with Occam.
old jim

 

[MadderDoc]

Your ADS seems a better fit with Occam.

It was Tepco's idea, and one they came up with it rather early. The implication of the ADS explanation is the negative, that the abrupt pressure drop at 9am on March 13 was _not_ caused by the RPV's failing in any major way. However, the recent investigations in the pedestal area could call that into question. At the least one is allowed to ask: if the RPV didn't fail in a major way at 9 am on March 13 -- such as to produce the mess that can be observed there now -- when did it?

 

[jim hardy]

At the least one is allowed to ask: if the RPV didn't fail in a major way at 9 am on March 13 -- such as to produce the mess that can be observed there now -- when did it?

Fair question.

Said mess, from Sotan's link a few pages back

I see lots of 'stuff' that appears to have run down from above. No clue whether it came through vessel bottom or down outside of vessel from above, running along underneath the insulation.
No clue what it is.
Looks a lot like sea salt , maybe some boric acid too, but that's just a visual interpretation of course prejudiced because i know they injected both.

To those of you who've been under a BWR
are these broken tubes of some sort ?

.
2X snip from Sotan's picture , http://photo.tepco.co.jp/library/170330_01/170330_07.jpg

 

[MadderDoc]

are these broken tubes of some sort ?

.
2X snip from Sotan's picture , http://photo.tepco.co.jp/library/170330_01/170330_07.jpg

We are in unit 2. The right arrow, that would be very likely a radiation artefact. Left arrow points, little doubt, to the end of a broken or cut tube.

Below is a blow up from an unmasked-enhanced image, in which the end of the same broken tube can be located close to bottom right of center. It is hanging down from above, bordering to a section of the overhanging structure (bottom left corner of image), inside which Tepco says several tubes to the control rod position and the local power range readouts, tubes that should be visible, are not.

full view of the enhanced image http://gyldengrisgaard.dk/fuku_docs/170330_04.jpg, original at Tepco

 

[MadderDoc]

I see lots of 'stuff' that appears to have run down from above. No clue whether it came through vessel bottom or down outside of vessel from above, running along underneath the insulation.
No clue what it is.
Looks a lot like sea salt , maybe some boric acid too, but that's just a visual interpretation of course prejudiced because i know they injected both.

I reckon one could look at it as a scene from the 'sewer' of unit 2. There has been coming water down and past here for a long time transporting stuff from above, and further on downwards. So, we are looking at something which has been changing over time, it was once spotless clean, then something happened and this drainage path was established, with many processes physical, chemical and who knows, biological, going on since then such as to produce the image that we can look at right now. It will continue to develop over time albeit surely slowly from now on. It is difficult to say, what composition is the deposits we see have built up on the hanging structures, or indeed the 'goo' on the remnants of the platform. I find it immediately interesting, that we don't see blobs of rusty aggregations of material on steel structures, which very much catch the eye in the videos from the corresponding area under the unit 3 RPV. Perhaps it lies in the difference, that the structures in the case of unit 3 have formed while being flooded in 'standing water', while those structures in unit 2 are being formed while continually being wetted in air, by streaming water from above.

 

[Hiddencamper]

How automatic depressurization system works:

Low level 1 (typically 1 foot above the fuel), low level 3 (2 feet below normal operating level as a confirmation signal this also triggers a scram), plus a high drywell signal (1.6-2.0 psig-ish), plus a detected high discharge pressure on an ECCS pump, plus a 105 second timer.

All signals were met except for the ECCS pump permissive. Follow up data reviews identified that wetwell pressure was so high that it raised ECCS pump suction pressure above the ADS set point, tricking the logic into thinking LPCI was operating and initiating an ADS blowdown.

The logic is 1/2 twice. Also some plants remove the high drywell pressure permissive, and other plants put an override timer in the event you have a sustained low water level for more than a few minutes with no high drywell pressure.

 

[MadderDoc]

How automatic depressurization system works

Thank you for explaining this, Hiddencamper. Do you know, once an ADS blow-down has been initiated, how or when is it triggered to end, meaning, is there some automatic to release the SRVs to come back closed again dependant on changing signals of water level/pressure? They would by design, I believe, come back closed passively, when/if power has run out to open/keep them actuated in the open position -- but besides that, assuming there is enough power, I wonder if there is some logic to actively trigger them to come back closed.

 

[Hiddencamper]

Thank you for explaining this, Hiddencamper. Do you know, once an ADS blow-down has been initiated, how or when is it triggered to end, meaning, is there some automatic to release the SRVs to come back closed again dependant on changing signals of water level/pressure? They would by design, I believe, come back closed passively, when/if power has run out to open/keep them actuated in the open position -- but besides that, assuming there is enough power, I wonder if there is some logic to actively trigger them to come back closed.

Once ADS initiates the safety function is to hold the valves open. If they went shut in the middle of the transient this can result in a loss of steam cooling to the core which is likely uncovered prior to core spray or LPCI cutting in. The ADS logic seals in once it activates and remains sealed in. It can be manually reset (which also resets the 105 second delay timer), and if the initiating signal doesn't clear you have to reset it every 105 seconds to keep the delay timer from activating the logic again.

Many plants have an ADS inhibit switch to block the logic. It's only used in an ATWS (scram failure) and many plants use it in Level control contingency 1 when they are planning on steam cooling the reactor or are trying to take advantage of any additional time below Level 1 to try and restore a high pressure injection system. The NRC doesn't like that though and a number of plants stopped inhibiting ADS outside of ATWS situations where water level is intentionally lowered to reduce natural circulation and power generation and a spurious blowdown can significiantlg challenge the safety of the core.

After a blowdown you want to keep the core depressurized. The steam gets dumped into the suppression pool which is likely to be at or over the HCTL (heat capacity temperature limit), and you may have a number of other "EMERGENCY DEPRESSURIZATION IS REQUIRED" actions in the EOPs keeping you there. At a minimum you want the minimum number of SRVs for decay heat removal open.

Some exceptions of course are during situations where you only have steam powered injection systems (RCIC/HPCI) and you need to prolong their operating time however you had a leak or containment challenge that requires depressurization. In this case you initiate ADS and lower pressure to the minimum acceptable RCIC/HPCI operating pressure then reset and inhibit the system using SRVs to maintain pressure in the RCIC/HPCI control band. Future revisions of the EOPs are also going to have you partially blowdown in scram failure ATWS situations where adequate core cooling can be maintained without a full depressurization, as we are learning that the reactor can exhibit chaotic behavior at low pressure / high power conditions.

 

[MadderDoc]

The ADS logic seals in once it activates and remains sealed in.

Thank you. That is interesting. So, if the ADS of unit 3 triggered in the morning of March 13th, opening 6-8 release valves to pass steam from the main steam line outlet of the RPV to the suppression pool, those release valves might have stayed open throughout the events that followed, and indeed ever since? (In that case no wonder the RPV and the PCV pressures henceforth appeared to be coupled.)

 

[Hiddencamper]

They would eventually fail closed if pneumatic air supply wasn't restored, or if drywell temperature caused failure. ADS for my unit using similar SRVs is only qualified to 340 degF. But for quite a while they would appear somewhat coupled until the eventual vessel breach.

 

[MadderDoc]

Now I wonder when Tepco think the vessel eventually was breached.

 

[Hiddencamper]

Now I wonder when Tepco think the vessel eventually was breached.

ADS consumes about 80-100 inches of vessel inventory or so when starting from high level. It likely completely uncovered the fuel. Time from fuel uncovery to vessel breach is likely a couple hours at that point. Very dependent on decay heat and injection.

 

[MadderDoc]

Tepco would think, it seems, that the vessel was breached already by about noon on the same day, not long after the temperature in the core had risen to set off the Zr-water reaction, such as to add even more heat. Past noon they started injection of sea water, it is not clear really how much, that went on until the morning on March 14, the next day, when the backwash pool ran out of seawater.

There was a hiatus of water injection then again, for several hours, until some time into the afternoon. In the meantime, there had been a hydrogen explosion in the building, apparently accompanied by a grand hiccup of steam from the PCV. The next several days, steam was seen coming out vigorously from the top of the building; there had apparently by then been achieved good contact between fuel and injected water..

Tepco writes somewhere, that hydrogen from a molten fuel-concrete interaction could possibly have contributed hydrogen to the explosions in unit 3 on March 14, and in unit 4 the next day, although Tepco also notes, that there may have been accumulated water in the basement of the PCV, already from a drywell spraying operation in the morning on March 13 ( which might have quenched the interaction). That is a bit confusing -- if Tepco thinks the vessel was breached by noon, injected sea water from that point in time on would also accumulate in the PCV basement.

 

[jim hardy]

Sorry - these detailed pictures i haven't yet learned which goes with which unit. So i'll make mistakes - and promptly admit them.

Tepco would think, it seems, that the vessel was breached already by about noon

Lest the phrase "vessel... breached" spark China Syndrome speculation,
"Breach" can be just incore instrument tubes failing up high in the core as it uncovers . That'd make many small leaks of whatever gas mixture is above the water level , presumably mostly steam and hydrogen.
That's benign compared to vessel melt through which if it happens at all would be much later.
In my plant we had fifty tubes of approximately 1/4 inch diameter for movable incores, the PWR equivalent of BWR TIP system.. I understand BWR's have a lot more incore instrumentation. old jim

 

[MadderDoc]

Yeah. Sometimes I wonder how they manage to make room for enough water in the vessel between those zillions of tubes of all kinds coming up :-) I guess what I am trying to figure out, is -- now that Tepco have seen the actual state of affairs in the area under the RPV vessel bottom -- when do Tepco think things came to end up to be like that? The report from Tepco with evaluations of the three units, which we've been looking into, cannot have been influenced by the result of the pedestal investigation, seeing that report was finished one month before the investigation was done.

 

[MadderDoc]

i haven't yet learned which goes with which unit.

Speaking of imagery from the upper inside pedestal area, one can get unit 2 and 3 mixed up. Main rule is, If it doesn't look like a complete wreck, it is from unit 2.