Japan Earthquake: Nuclear Plants at Fukushima Daiichi

[elektrownik]

New temperature data for unit 2, it is interesting that CRD sensor value is increasing and that at some points it was close to second sensor value (before >200C readings):

 

[Shinjukusam]

They announced that three thermocouples were broken. Would that be one of them?

 

[zapperzero]

New temperature data for unit 2, it is interesting that CRD sensor value is increasing and that at some points it was close to second sensor value (before >200C readings):

It's maddening that they track up to a point. Makes it seem like something is happening and they're failing because of it. Simplest thing I can think of is corrosion due to water. But I don't know if failure due to corrosion is as fast as all that.

Another thing that's clouding the issue is that some of these are actually normalized readings - there has been some corrosion since day one and TEPCO said they are adjusting for this in their reporting - up to a point.

 

[elektrownik]

They announced that three thermocouples were broken. Would that be one of them?

The problem is that they are close to each other so this suggests that something is going on there
here is plot:

 

[jim hardy]

""Another thing that's clouding the issue is that some of these are actually normalized readings - there has been some corrosion since day one and TEPCO said they are adjusting for this in their reporting - up to a point.""

sigh - it is so hard to troubleshoot through somebody's interpretation of facts.
If only they'd post the raw millivolt readings..

do i recall those are iron-constantan thermocouples? Type J? I'm not at all sure.
In my plant we preferred chromel-constantan type E because the iron wire rusts but chromel doesn't. The tc's inside our reactor were chromel-alumel type K.

here's a link to some millivolt-temperature tables.
http://www.temperatures.com/tctables.html

 

[Rive]

My impression is that this happened because this thread threatened to become blasphemic in the way that it could show that recriticality issues are in fact possible.

IIRC what made that topic closed is flaming about recriticality in U3 fuel pool, without any scientific backup but with many emotions involved.

The possibility of recriticality is 'on the desk' here, from start as I recall: and as you can see, it's already discussed in depth by the nuclear scientific community (some links already here).

 

[zapperzero]

do i recall those are iron-constantan thermocouples? Type J? I'm not at all sure.
In my plant we preferred chromel-constantan type E because the iron wire rusts but chromel doesn't. The tc's inside our reactor were chromel-alumel type K.
[/url]

copper-constantan, according to ex-skf, who in turn quotes Matsumoto:

http://ex-skf.blogspot.com/2012/02/fukushima-i-nuke-plant-additional-info.html

 

[tsutsuji]

I translated two pages of Tepco's Report on “The Concept of Securing the Mid-Term Safety” (part 1, revision 2, December 2011) about unit 2's thermometers:

http://www.tepco.co.jp/cc/press/betu11_j/images/111207a.pdf pages 1-81 - 1-82 (87/131 - 88/131).

 

[tsutsuji]

http://www.tepco.co.jp/cc/press/betu11_j/images/111207a.pdf Figure 4-4 page 1-63 (69/131).

Va = (Ra/Rb) * (Vin-Vb) + Vin

 

[jim hardy]

Remarkable find tsujitsu !

THANK YOU!

copper constantan is a good choice because niether wire rusts.

And your find confirms they know about insulation damage and have localized it with TDR as one would expect of top notch folks.
Recall i said they have better technicians than me over there...
they know their plant like their own living room so know the likelihood of water where that damaged insulation is.

here's why i look for corroboration on those temperature measurements:

Galvanic Action
Some thermocouple insulating materials contain dyes that form an electrolyte in the presence of water. The electrolyte generates a galvanic voltage between the leads, which in turn, produces output signals hundreds of times greater than the net open circuit voltage. Thus, good installation practice calls for shielding the thermocouple wires from high humidity and all liquids to avoid such problems.

http://www.sensorland.com/HowPage017.html


Constantan is a copper-nickel alloy 55/45 i think. A us nickel is 75/25...
i just put a US nickel and a US penny in a coffee cup of salted water and read ~170 millivolts between them.

Thank you!

ps - edit:
And they've used circuit analysis to estimate thermocouple voltage in presence of interfering Vb , as from galvanic action...
nice.
now i need to repeat that penny-nickel experiment at various temperatures...

thanks, Tsujitsu

old jim

 

[Yamanote]

Thanks to you as well for explaining Jim!

 

[tsutsuji]

http://www3.nhk.or.jp/news/genpatsu-fukushima/20120214/1350_ondokei.html The NISA instructed Tepco to write a report by 15 February, on such things as the cause of the high thermometer values and the way to measure temperatures by other methods.

http://www.47news.jp/47topics/e/225687.php On 15 February Junichi Matsumoto said "the report to NISA will be sent late in the night. I will explain its contents on 16 February".

 

[Atomfritz]

Thank you Tsutsuji for your finding.

I feel really amazed that a cheap two-wire thermocouple configuration was used and no four-wire-one with return wires.

With a four-wire thermocouple it would have been possible to narrow down the cause of the problem without having physical access. Possibilities include damage of the thermocouple element itself, wiring damage, shifting/decalibration etc.


Further reading about thermocouple basics for the interested reader "www.omega.com/temperature/z/pdf/z021-032.pdf" .

 

[mscharisma]

http://www.47news.jp/47topics/e/225687.php On 15 February Junichi Matsumoto said "the report to NISA will be sent late in the night. I will explain its contents on 16 February".

And then there was silence?

 

[Shinjukusam]

And then there was silence?

From NHK :

The operator of the crippled Fukushima Daiichi nuclear power plant has attributed abnormally high temperature readings at one of the facility's reactors to a malfunctioning thermometer.

Tokyo Electric Power Company, or TEPCO, reported the analysis of the problem at the Number 2 reactor to the government's nuclear safety agency on Thursday.

The thermometer at the reactor has been showing much higher readings than 2 others.

The utility said it's highly unlikely that temperatures could rise so high unless at least 60 percent of the melted nuclear fuel in the reactor were concentrated near the thermometer.

TEPCO confirmed signs of rising temperatures in experiments it conducted under unusually large electric resistance, as was found in the thermometer.

The thermometer serves as an indicator to assess whether the reactor can stay in a state of cold shutdown.

TEPCO says the thermometer will no longer be monitored. The firm says it will comprehensively examine data, including other thermometer readings and radiation levels in the reactor's containment vessel, to determine whether a state of cold shutdown is achieved.

The utility plans to reduce water injections into the reactor to a level at which such injections were done before the thermometer readings began rising, if the nuclear safety agency says that doing so is reasonable. Thursday, February 16, 2012 17:28 +0900 (JST)

 

[elektrownik]

And this sensor is also damaged ?

 

[zapperzero]

And this sensor is also damaged ?

Common cause failures have been a recurring theme in this accident. I wouldn't be too surprised.

 

[tsutsuji]

http://www3.nhk.or.jp/news/genpatsu-fukushima/20120216/index.html Tepco is studying new thermometers for unit 2. Already 8 out of 41 or 20% of unit 2's thermometers are broken. As more might break down in the future, Tepco is studying how to install new thermometers. As one cannot easily approach the reactor, the study will take from one to two years.

http://www.tepco.co.jp/cc/press/betu12_j/images/120216a.pdf Report to NISA about unit 2's thermometer (48 pages, Japanese)

 

[jim hardy]

Common cause failures have been a recurring theme in this accident. I wouldn't be too surprised.

Tsujitsu's post 12383 mentions in the third thumbnail a "multiconductor cable" for some TC's around bellows seal.
Assuming multiconductor cable was also used around RPV bottom, it would not be surprising to find that common mode failure.

These TC's may not have been deemed important for post accident monitoring so not run in seismic grade conduit?
Perhaps a BWR person would know.

Looking at the formula they use for estimating thermocouple voltage,
again from Tsujitsu's thumbnails this time the one in post 12384,
Va = (Ra/Rb) * (Vin-Vb) + Vin
When everything is dry all three voltages Va Vb and Vin are same order of magnitude.
When the low resistance location gets wet Rb becomes a small denominator and Vb grows to a substantial multiple of Va.
a good math guy would run sensitivities...

That they did so well with it so long really impresses me
and makes me think that when they changed injection flow it washed down the area.
That's my two cents from 6536 miles away.
What do you folks make of those thermocouple symptoms?

old jim

 

[zapperzero]

makes me think that when they changed injection flow it washed down the area.
That's my two cents from 6536 miles away.
What do you folks make of those thermocouple symptoms?

old jim

My head hurts and I'm 8500 km away, give or take. What I can't understand is why the indicators first go way up, and only then way down. As far as I understand it, the actual measuring device is a voltmeter, right?

The way up part I can understand - the "electromotive force" as the Japanese so quaintly call it increases, because the water has salts in it, there's a (bigger) battery in there all of a sudden, so indicated temp rises.

But then, why the sudden crash to small but positive values? Does the wire get corroded through and through, so there is a huge increase in resistivity?

Then again, maybe I didn't understand something about that circuit correctly.

 

[jim hardy]

Wet thermocouples are not well behaved at all.

Their diagram is okay so far as it goes.

BUt think what they probably have in there:
thermocouple extension wire with insulation scraped or melted, adjacent structural material perhaps conduit wall or insulation?
so you have the possibility not only the 'electromotive force' from the copper-constantan galvanic cell,
but copper to conduit galvanic cell
constantan to conduit galvanic cell
either of above to insulation
any combination of above, and that's too many permutations for me.
and what's the conduit? In my plant the conduit was galvanized (zinc) and reactor vessel instlation was stainless steel

not trying to be difficult here, just explaining why i am not surprised at Tepco's difficulties with pulling a reading out of this mess.

Tsujitsu's link in post 12393
http://www.tepco.co.jp/cc/press/betu12_j/images/120216a.pdf

see page 21, right hand chart looks to me like a wet thermocouple ,
its reading swings between -300 and +400 and it looks like they drew a line through mean..
They showed more patience than i would have.

A wet TC will act like a small battery. You can charge it with your ohm-meter and watch voltage drift back down.
My heart goes out to those guys .
Whole world is watching and seldom is heard an encouraging word..

If any Daichi instrument techs are reading this - I say you have "True Grit."

Sorry i can't be more scientific about it.


old jim

 

[jim hardy]

and indeed ZZ it's a voltage measurement.

 

[zapperzero]

any combination of above, and that's too many permutations for me.

You and me both. Out the window goes my neat little mental model.

Tsujitsu's link in post 12393
http://www.tepco.co.jp/cc/press/betu12_j/images/120216a.pdf

see page 21, right hand chart looks to me like a wet thermocouple ,
its reading swings between -300 and +400 and it looks like they drew a line through mean..

A-ha. I hadn't seen that. It makes sense now. Thanks.

 

[tsutsuji]

Tsujitsu's link in post 12393
http://www.tepco.co.jp/cc/press/betu12_j/images/120216a.pdf

see page 21, right hand chart looks to me like a wet thermocouple ,
its reading swings between -300 and +400 and it looks like they drew a line through mean..

Page 21 is an experiment they made by connecting a variable resistor in series in one of the wires. When they set the resistor at 1.2 kΩ a small oscillation takes place and at 8 kΩ they find a large oscillation. The water's temperature is controlled by the reference thermometer (black curve) : 59°C for the 1.2 kΩ experiment, and 100°C for the 8 kΩ experiment.

They call these variations "hunching" (?) in English.

On pages 22 and 23 they make another experiment, by simulating cable damage, stripping the cable's insulator, leaving only one strand of copper on the copper wire, connecting a 40 kΩ resistor, and using salt water.

The resistance measurement results on unit 2's thermometers are in the table at the bottom of page 8 (copied also in the table page 15) :

30 september 2011 : 175.47 Ω (low insulation)
3 February 2012 : 244.25 Ω (low insulation)
13 February 2012 : from 500 to 523 Ω (rupture)
Average value of regular inspections : 303.37 Ω

They say that as a rule when the ratio of ( measured resistance / average regular inspection value ) is higher than 1.1, it means that the wire is in a rupture trend. It seems also that when this ratio is smaller than 1 they conclude "low insulation" in the table page 8.

At paragraph d) on page 18, which is the last sentence of the part of the report concerning the causes of the problem (following pages are attachments), they tell about their plans for the future : to verify the consistency between the experiments and the real phenomenon and to elucidate how the problem was generated.

The last part of the report, from page 36 to the end is about the different approaches that can be thought to perform alternative temperature measurements. They are listed in the table page 38. The schedule is on page 39.

 

[jim hardy]

Thanks Tsujitsu

I guess i should have kept quiet - all i could see was the pictures..
I hope i didnt mis-lead severely, thought that was plant data. I apologize.

Page 21 is an experiment they made by connecting a variable resistor in series in one of the wires.
When they set the resistor at 1.2 kΩ a small oscillation takes place and at 8 kΩ they find a large oscillation.

Unbalance in the leads can upset low level measurement systems.

They are becoming the world experts at thermocouple failure interpretation...

Thank you again for your hard work at finding and posting this information.



old jim

 

[rmattila]

Tsutsuji-san is truly doing a remarkable effort in helping many people to get an ever-improving picture of the events that took place and the current status of the plants. I have no doubt in my mind that some of this information will eventually lead into actual improvement of nuclear safety in many existing plants, as well as in those built in the future. With persistence and devotion, one man can really make a difference.

Thank you for not giving up!

 

[NUCENG]

Tsujitsu's post 12383 mentions in the third thumbnail a "multiconductor cable" for some TC's around bellows seal.
Assuming multiconductor cable was also used around RPV bottom, it would not be surprising to find that common mode failure.

These TC's may not have been deemed important for post accident monitoring so not run in seismic grade conduit?
Perhaps a BWR person would know.

Looking at the formula they use for estimating thermocouple voltage,
again from Tsujitsu's thumbnails this time the one in post 12384,
Va = (Ra/Rb) * (Vin-Vb) + Vin
When everything is dry all three voltages Va Vb and Vin are same order of magnitude.
When the low resistance location gets wet Rb becomes a small denominator and Vb grows to a substantial multiple of Va.
a good math guy would run sensitivities...

That they did so well with it so long really impresses me
and makes me think that when they changed injection flow it washed down the area.
That's my two cents from 6536 miles away.
What do you folks make of those thermocouple symptoms?

old jim

An additional consideration is the effect of radiation (beta and gamma) on the instrumentation cables insulation and cable jackets. Over time I would expect more of the thermocouple cables to have insulation degradation and develop problems. These instruments were not required to be qualified as BWR post-accident monitoring instruments in US plants under RG 1.97. That may be changing.

 

[tsutsuji]

http://www3.nhk.or.jp/news/genpatsu-fukushima/20120217/index.html Tepco will reduce unit 2's injection rate by 4 tons/hour on 18 February.

http://www3.nhk.or.jp/news/genpatsu-fukushima/20120218/index.html The cause of the collapse of the Yonomori line tower No. 27 was underground water. The tower collapsed 2 minutes 30 seconds after the start of the earthquake. Underground water flows under the embankment. The ground became weaker as a consequence of shaking for a long time. The embankment was constructed during the 1965 - 1975 period to fill up a marsh, so there is much water. Tepco checked 530 towers concerning the Fukushima Daini and Kashiwazaki Kariwa plants, etc. but found no similar case.

http://www.tepco.co.jp/cc/press/betu12_j/images/120217c.pdf "causes of damage situation of power facilities inside and outside of Fukushima Daiichi Nuclear Power" (2nd release) (Japanese, 15 pages)

http://www.tepco.co.jp/cc/press/betu12_j/images/120217b.pdf "safety evaluation of transmission tower foundations regarding the offsite power supply to nuclear power stations and reprocessing plants" (Japanese, 17 pages)

 

[tsutsuji]

http://www.yomiuri.co.jp/e-japan/chiba/news/20120221-OYT8T00073.htm Robots Quince No.2 and No. 3　were dispatched from Chiba Institute of Technology to Fukushima Daiichi on 20 February. They are equipped with a cutter so that they can cut their cable to free themselves in case the cable is entangled. Quince No.2 has a camera arm long enough to look over the fuel pool fence, which Quince No.1 could not do.

 

[LabratSR]

Nuclear Regulatory Commission releases audio of Fukushima disaster

http://www.marketplace.org/topics/w...-commission-releases-audio-fukushima-disaster



About 3000 pages of transcripts

http://pbadupws.nrc.gov/docs/ML1205/ML120520264.html

Audio
http://youtu.be/ciRRsCIAy6A