Fukushima Management and Government Performance

[nikkkom]

I still feel the hope that EDGs will always save the day is stupid and dangerous.

I think that total SBO should not be treated as unthinkable event; instead, operators need to know exactly what to do.

Can people who have the first-hand knowledge of current operators' accident training tell me whether operators are trained for full SBO? Will they know where to go and which valves to open or close (manually, or with portable energy sources), etc?

 

[etudiant]

1e-2 failure probability per EDG is usually considered acceptable. In older 2 x 100 % plants this means 1e-4 failure probability of both diesels due to independent single failures. At newer, 4 x 50%, 3 x 100 % or 4 x 100 % plants, common cause failures dominate the EDG loss chains.

If memory serves, the Shoreham site had 3 large EDGs, all of which failed, so presumably there was some sort of common problem.
That would seem to undermine the expectation that the EDG failures can be taken independently, highlighting the common cause issue rmattila has raised.

 

[tsutsuji]

3.5. Evaluation of reliability against SBO etc. [http://www.nsc.go.jp/info/20110713_dis.pdf 23/96]

(1) Reliability against SBO

There has been no SBO precedent in Japan until now. However, SBO caused core damage PSA results are provided in (3) below.

(2) Reliability against external power, EDG, etc.

PSA of Japanese representative plants based on the loss of offsite power precedents and EDG failure precedents mentioned in "3.4. failure etc. precedents in Japan" provide the following estimates of the reliability of external power, EDG, etc.

a) External power reliability

① Loss of offsite power frequency

There are 3 precedents of BWR loss of offsite power and 1 precedent of PWR loss of offiste power. Among these, one of the BWR cases was generated by a design characteristic of the concerned plant, and later design change measures were taken, so that it can be explained that a similar event cannot occur again in Japan in the future, so we removed it from the survey for calculating the present loss of offsite power frequency. Thus, based on the 2 BWR cases and the one PWR case, loss of offsite power frequencies are determined as follows:

i)BWR: 2 cases generated in 153.8 Reactor*Year, about 1.4 10^-2 /Reactor*Year.
Under the hypothesis of an error factor of 3, the median is about 1.1 10^-2 /Reactor*Year, the 95% upper limit value is about 3.3 10^-2 /Reactor*Year, and the 5% lower limit value is about 3.7 10^-3 /Reactor*Year.

ii)PWR: 1 case generated in 136.7 Reactor*Year, about 7.3 10^-3 /Reactor*Year.
Under the hypothesis of an error factor of 3, the median is about 5.8 10^-3 /Reactor*Year, the 95% upper limit value is about 1.7 10^-2 /Reactor*Year, and the 5% lower limit value is about 1.9 10^-3 /Reactor*Year.

② Loss of offsite power restoration capacity

As shown in "3.4", loss of offsite power restoration capacity according to national records, using the two-line accident data since April 1962, the probability of restoration failure for 30 minute long accidents is about 0.05, and the one for 8 hour long accidents is about 0.001, which is a good restoration capacity in comparison with American values. In other words, in the United States, in NUREG-1032, the relation between loss of offsite power frequency and duration is evaluated by categorizing (categorization by cluster) in function of offsite power system design, power transmission system characteristics, severe weather and extremely severe weather. Just for reference, when these results are compared with the restoration capacity in real evaluations of Japanese plants, even in the plants belonging to the cluster with the most reliable offsite power, for a 30 minute accident duration, the restoration failure probability is about 0.5, and for an 8 hour long accident, the restoration failure probability is about 10^-2, so theses results are about 10 times worse than the results found in realistic evaluations of Japanese plants' restoration capacity. In the PSA of representative Japanese plants, these results are estimated conservatively, and we use data amounting to a restoration failure probability of 10^-2 for an 8 hour accident duration.

b) EDG reliability

① EDG start failure probability
As mentioned in 3.5.(3), the EDG start failure probability used in PSA is calculated from the start records from April 1970 to March 1983, as follows:

* number of starts : 14,878
* number of start failures : 18
* start failure probability : 18/14,878 = 1.2 10-3/demand (henceforth referred to as "d")

Under the hypothesis of an error factor of 3, the median is about 9.6 10^-4/d, the 95% upper limit value is about 2.9 10^-3/d, and the 5% lower limit value is about 3.2 10^-4/d.

As indicated above in 3.4, in the more recent records values are smaller, and according to the operation records from April 1970 to March 1990:

* start failure probability : 30/28,012 = 1.07 10^-3/d

Under the hypothesis of an error factor of 3, the median is about 8.6 10^-4/d, the 95% upper limit value is about 2.6 10^-3/d, and the 5% lower limit value is about 2.9 10^-4/d.

and according to the records from April 1980 to March 1990, it improves to about 5.5 10-4/d.

Under the hypothesis of an error factor of 3, the median is about 4.4 10^-4/d, the 95% upper limit value is about 1.3 10^-3/d, and the 5% lower limit value is about 1.5 10^-4/d.

② EDG failure probability in continuous operation

As no data were prepared in Japan concerning EDG continuous operation failure rates, for PSA values estimated on the basis of the the US data to which we apply the ratio of start failure probabilities in Japan and in the United States as a corrective.

In the future, it is necessary to carry out the preparation of national data on EDG continuous operation failure rates.

c) Reliability of DC sources such as emergency batteries

As mentioned above, there is no failure precedent concerning DC sources such as emergency batteries, and although its reliability is thought to be high, in the PSA evaluation we used US data as follows.

(3) SBO as seen in probabilistic safety assessment

We shall examine SBO as seen in the results of PSA performed in national representative plants.

Here, our considerations are based on PSA performed by the industry. In this PSA, transient occurrence frequencies use operation records in Japanese plants, but equipment failure data, common factor failure data are based on US data. However, as EDG failure rate data have been prepared in Japan, they can be used, so they are used. The results of PSA performed for representative Japanese plants are low, as total core damage frequencies are below the 10^-5/Reactor*Year safety goal set by the IAEA in its basic safety principles for new reactor design.

Concerning representative Japanese BWR-3, BWR-4 and BWR-5 plants, each accident sequence's contribution rate to total core damage frequency is indicated in Figure 3-16 [http://www.nsc.go.jp/info/20110713_dis.pdf 94/96].

In representative BWR-4/BWR-5 plants, evaluations including shared EDGs conclude that the SBO contribution rate is higher than that in BWR-3 plants. As BWR-3 plants possesses 2 IC systems , their design is comparatively stronger against loss of offsite power, and SBO is not dominant. In all plants, themselves, core damage frequencies generated by the SBO sequence are not high. (The SBO (TB sequence) generated core damage frequency and the contribution rate of SBO to total core damage frequency are respectively about 1.6 10^-8/Reactor*Year and 2% in BWR-3 plants, about 1.9 10^-7/Reactor*Year and 24% in BWR-4 plants, and about 7.2 10^-8/Reactor*Year and 22% in BWR-5 plants).

Also, in the Japan Institute of Nuclear Safety assessment results, the core damage frequency generated by the SBO sequence is small like in the industry's assessment. (As a result of a PSA performed for a representative 110,000 kW class Japanese BWR, the SBO (TB sequence) generated core damage frequency and the contribution rate of SBO to total core damage frequency are respectively about 2.4 10^-9/Reactor*Year and 1%).

Concerning representative Japanese PWR plants (dry type 4 loop plant, ice condenser type 4 loop plant), the the breakdown by generating factor of contribution rates to total core damage frequency is indicated in Figure 3-17 [http://www.nsc.go.jp/info/20110713_dis.pdf 95/96].

The contribution of loss of offsite power caused sequences is low in both the dry type and the ice condenser type. The contribution of loss of offsite power caused sequences is even smaller in ice condenser type 4 loop plants because they are equipped with 2 turbine driven auxiliary feedwater pumps (1 pump in dry type 4 loop plants). (The SBO (TB sequence) generated core damage frequency and the contribution rate of SBO to total core damage frequency are respectively about 1.1 10^-9/Reactor*Year and 0.2% in dry type 4 loop plants, and about 2.1 10^-10/Reactor*Year and 0.01% in ice condenser type 4 loop plants).

In the Japan Institute of Nuclear Safety assessment results too, the core damage frequency generated by the SBO sequence is small like in the industry's assessment. (As a result of a PSA performed for a representative 110,000 kW class Japanese PWR, the loss of offsite power generated core damage frequency is about 6.6 10^-9/Reactor*Year and the contribution rate to total core damage frequency is about 4%).

As the PSA performed in Japan and the PSA performed in foreign countries are not using a unified way of thinking as regards the assessment method's details or the prerequisites of data, etc., an indiscriminate comparison is not suitable, but we want to try to use the NUREG-1150 assessment results as reference. Those results are presented in Figure 3-18 [http://www.nsc.go.jp/info/20110713_dis.pdf 96/96]. In the NUREG-1150 assessment, SBO is a protruding accident sequence at the Surry and Grand Gulf reactors.

4. Assessment of guidelines and safety securing countermeasures against SBOs

 

[tsutsuji]

In the course of checking my translations of the acronyms (TQUX etc.) I found the following document written in English:

M.Kajimoto, M.Sugawara, S.Sumita, K.Funayama, F.Kasahara, N.Tanaka and M.Hirano, "Evaluation of Technological Appropriateness of the Implemented Accident Management Measures for BWR by Level 1 and Level 2 PSA Methods", OECD/NEA/CSNI Workshop on "Implementation of Severe Accident Management", at Paul Scherer Institute, September 10-13 in 2001: http://sacre.web.psi.ch/ISAMM2009/oecd-sami2001/Presentations/p6-M.Kajimoto-BWR/SAMBWR.pdf

 

[tsutsuji]

And now, the last installment of a nearly complete translation of the NRC's SBO group's final report (11 June 1993) entitled "Full loss of AC power events in nuclear power plants". Nearly complete, because I did not translate some of the attachments. This is now part 4., followed by part 5 already translated at https://www.physicsforums.com/showpost.php?p=3943942&postcount=387 .

4. Assessment of guidelines and safety securing countermeasures against SBOs [http://www.nsc.go.jp/info/20110713_dis.pdf 26/96]

(1) About the Regulatory Guide for Reviewing Safety Design

Based on the Regulatory Guide for Reviewing Safety Design, the power supply systems in Japanese nuclear power plants are required among other things to have a high reliability and redundancy and to secure reactor safety against short time full loss of AC power. On the other hand, as the operation records of Japanese nuclear power plants cover over 300 Reactor*Year, as it is useful to evaluate whether power supply equipments reached sufficient reliability during that period, we evaluated the damages generated by power supply equipments and their consequences on plants, and also the reactor resistance capacity during full loss of AC power.


① There has not been any SBO precedent in Japan's nuclear power plants until now. However, as they constitute the main SBO precedents occurring in foreign countries, we investigated the 3 cases that occurred at light water reactors in the USA. Although it is difficult to study by direct comparison because the situation of design and operational management is not necessarily the same as in Japan, the points from those precedents that are to be reflected in Japanese nuclear power plants, as general lessons whose awareness must be renewed, are ① the importance of countermeasures against human errors (training of operators, etc.) and ② adequate inspections during nuclear reactor shutdowns, of the facilities whose purpose is to secure the safety of nuclear reactor facilities including electric supply equipments, and the importance of conservative design.

② In our country's records, the loss of offsite power frequency is low at about 0.01 /Reactor*Year, which is bout 10 times lower than the United States' 0.1/Site*Year. In our country's records, all occurrences are generated by causes located in the transmission lines outside the power plants. The occurrence rate of these transmission line causes is nearly the same as the American one, but in the United States, there are many occurrence precedents generated by onsite causes, which create a low reliability.

③ Considering the small number of data concerning loss of offsite power in Japan, we broadened our scope by not limiting ourselves to nuclear power plant records, and we evaluated the restoration capacity of Japanese two-line power transmission lines, and compared, for reference, with the records of American nuclear power plants. As a result, the restoration capacity in our country is generally good, and for example we can use as indicator the 8 hour restoration failure probability, which is about 10^-3 in Japan, which is even higher than the 10^-2 probability in the highest reliability cluster in the United States. This comparison is a general one, and although making an exact evaluation of the reasons behind this difference is difficult, it is presumed that the good Japanese records are perhaps due to a difference in the structure of transmission lines. However, according to the records of Japanese nuclear power plants, restoration took place within 30 minutes in all cases, while in the United States, the longest precedent is a 19 hour long one (statistics until 1989).

④ The EDG failure rate based on the records of the past 10 years is about 5.5 10^-4/demand in Japan, which indicates a higher reliability than the United States' 2 10^-2/demand, and we evaluate that this is the result of a variety of reliability improving countermeasures. In the future, it is desirable to split, collect and arrange EDG start reliability data and EDG continuous operation after start data, and to reflect this in failure analysis and PSA.

⑤ The emergency DC power sources (emergency batteries, etc) are important in the hypothesis of a SBO, and in Japanese nuclear power plants, the emergency batteries' capacity is 5 hours or more (on the basis where some of the loads are disconnected). Concerning the reliability of emergency DC power equipments, in Japan there is no loss of function precedent, including degradation of charging capacity. For that reason, it is thought that the reliability of emergency DC power equipments is maintained at a high level, but as a continuation, it is desirable that efforts are paid to secure reliability based on collecting and arranging foreign countries' precedents and learning the lessons from them. One must note that in the United States failure precedents have been reported concerning the emergency batteries etc. of the emergency DC power supply system. Also, emergency battery capacity, for example at Surry, in the case where some of the loads are disconnected, is estimated to be 4 hours.

⑥ Thus, the reliability of Japanese offsite power systems, EDGs and emergency DC power equipments is good, but we further evaluated the reactors' endurance capacity under the hypothesis of a SBO. In other words, due to the response operations that have already been integrated in procedures, such as disconnecting part of the battery loads, reactor endurance capacity was estimated to be 5 hours. If we tentatively estimate the appropriateness of Japanese plants against the new American regulations based on RG.155, considering the characteristics of EDG reliability and weather conditions surrounding the plants in Japan, endurance capacity duration becomes 4 hours in all the plants, and in response to this, by way of response operations already integrated into procedures such as disconnecting part of the battery loads, endurance capacity against SBO in representative Japanese plants, was estimated to be about 5 hours or more, which means that the SBO regulations set by the American NRC are satisfied. Based on this, one can say that our country's endurance capacity against SBO is good.

⑦ According to results of PSA performed in representative Japanese plants (causal events: internal events only), the total core damage frequency is small, and the SBO-caused core damage frequency itself is also small. Also, in PWR and BWR-3, the SBO is not the main contributor to core damage, and, on the other hand, in BWR-4 and BWR-5 while the contribution rate is higher than that of PWR and BWR-3, the SBO-caused cored damage frequency is not high by itself.

⑧ In the main foreign countries, regulatory requirements nearly similar as the Japanese ones are set concerning offsite power and emergency onsite power source, etc.. Also, as regards regulatory requirements against SBO in the main foreign countries, the United States and France are implementing regulatory requirements against SBO (including prolonged SBO). The United Kingdom and Germany have nearly similar regulatory requirements as Japan.

(2) About the Regulatory Guide for Reviewing Safety Design

As mentioned above in 4.(1), today, the reliability of Japanese nuclear power plants' electric supply systems is high, and efforts are being engaged to maintain and increase that reliability. The occurrence probability of SBO is small. Also, should a SBO occur, as the restoration of offsite power, etc. can be expected, it is thought that the probability that it leads to a serious situation is low.

(3) About bringing safety one step further

① Today, the safety of Japanese nuclear plants against SBO is based on a good management of operations and maintenance, and it is necessary to pay efforts for continuation. Furthermore, in order to bring safety one step further, it is needless to say that operators have to master the manual, and if new knowledge is obtained in the future, it is necessary to pay efforts to appropriately reflect that new knowledge in design, operation, maintenance, and manual, etc..

② According to the results of PSA in representative Japanese plants, SBO-caused core damage frequencies are not especially high, but as the Nuclear Safety Commission has just decided to to encourage accident management as a countermeasure against severe accidents, while conducting studies, etc. of the SBO core damage frequency by PSA of individual plants, it is important to pay efforts to conduct the studies that will pave the way toward a higher efficiency level of preparatory measures such as accident management.

③ Since, in recent years, precedents of loss of emergency onsite AC power source during reactor shutdown took place in in foreign countries' plants and loss of redundancy of safety systems and equipments due to overhaul during regular inspections is possible, it is vital to pay sufficient attention to operation management etc. during reactor shutdown. Also, it is desirable to study PSA during reactor shutdown.

 

[tsutsuji]

http://www3.nhk.or.jp/news/genpatsu-fukushima/20120618/2050_map.html It has been found that in March last year the ministry of education and science did not publicly release a map prepared by the US government that showed that the radiations were spreading in the north-west direction. The map was created by aircraft from 17 to 19 March. For example, it showed the locations above 125 microsievert/hour in red and one could understand at the first glance how radiations were distributed. The map was offered to the Japanese ministry of foreign affairs by the US department of energy on 20 March and it was immediately transmitted to the ministry of education and science and to the NISA. However, the ministry of education of science and the NISA did not release it publicly, nor did they transmit the informations to the Prime Minister office or to the concerned government agencies. The data were publicly released by the US government on its website 3 days later on 23 March.

The vice head of the ministry of education and science's Science and Technology Policy Bureau, Mr Watanabe, says for example: "We thought the public release had to be done by the US government. At that time, we performed a radiation survey at 180 locations and released the results", and believes that there was no problem in their response. The NISA's Tetsuya Yamamoto says: "At that time, an e-mail from the foreign ministry came to our foreign relation room. It was transmitted to the radiation group, but we are presently investigating to understand why it was not publicly released. Thinking about it now, I think it should have been released. Not appropriately offer information is truly regrettable, and we want to pay efforts to respond to the results of the cabinet investigation's results etc.".

http://ajw.asahi.com/article/0311disaster/fukushima/AJ201206180048 "The first hint that the science ministry and NISA had obtained the radiation maps from the Energy Department came in late October"
http://ajw.asahi.com/article/0311disaster/fukushima/AJ201206190064 "It remains unclear if the members were even aware that a radiation map with U.S. data was posted on a whiteboard in the room."

 

[zapperzero]

http://ajw.asahi.com/article/0311disaster/fukushima/AJ201206190064 "It remains unclear if the members were even aware that a radiation map with U.S. data was posted on a whiteboard in the room."

This is incredible. Why such an urge to manufacture excuses?

 

[tsutsuji]

Probably they were afraid to be blamed for creating a panic, like Koichiro Nakamura:

Oh by the way, speaking of the core meltdown, NISA spokesman Koichiro Nakamura was replaced after he spoke of the possibility of the core meltdown in Reactor 1 in his press conference on March 12, at the express and angry demand from the Prime Minister himself. (News Post Seven, in Japanese)
http://ex-skf.blogspot.com/2011/04/fukushima-i-nuke-plant-pm-assistant.html

Ex-skf's source, http://www.news-postseven.com/archives/20110320_15548.html has the following anonymous quote: 「菅首相と枝野官房長官は、中村審議官が国民に不安を与えたと問題視し、もう会見させるなといってきた」（経産省幹部）

"Considering that inspector Nakamura had brought a feeling of insecurity to the citizens, Prime Minister Kan and Chief Cabinet Secretary Edano viewed this as a problem, and said 'don't do press conferences any longer'" (a ministry of Economy and Industry executive)

Here is the New York Times' analysis (2012-06-19) http://www.nytimes.com/2012/06/20/w...-us-radiation-data.html?_r=1&ref=atomicenergy : "The failure is being seen by critics in Japan as another example of the government’s early attempts to play down the severity of the accident by withholding damaging information."

Not telling the citizens is one thing. But why not transmit the data to the Prime Minister's office, then ? Some kind of burying one's head in the sand like an ostrich ? Or were they afraid to rely on US data, out of fear that they might be inaccurate ?

Answers to these questions might be provided in the forthcoming final version of the cabinet investigation report.

 

[zapperzero]

I think this goes here:

http://ajw.asahi.com/article/0311disaster/fukushima/AJ201208060093

TEPCO subcontractors played all sorts of games with their employees' dosimeters.

 

[zapperzero]

NRC covered up aspects of the Fukushima crisis

http://www.nbcnews.com/storyline/fu...oncerns-hailed-safety-record-fukushima-n48561

At the end of that long first weekend of the crisis three years ago, NRC Public Affairs Director Eliot Brenner thanked his staff for sticking to the talking points that the team had been distributing to senior officials and the public.

"While we know more than these say," Brenner wrote, "we're sticking to this story for now."

 

[QuantumPion]

I still feel the hope that EDGs will always save the day is stupid and dangerous.

I think that total SBO should not be treated as unthinkable event; instead, operators need to know exactly what to do.

Can people who have the first-hand knowledge of current operators' accident training tell me whether operators are trained for full SBO? Will they know where to go and which valves to open or close (manually, or with portable energy sources), etc?

Yes operators are fully trained for SBO. Since Fukushima, additional focus has been added to BDB events. I think it is worth noting that there is a large difference in attitude and culture between US operators and Japanese operators. US Operators receive far more and rigorous training, have more expertise, and take an individual responsibility for protecting the plant. Japanese operators tend to rely on a higher authority to tell them what to do.

 

[QuantumPion]

I find it somewhat strange, how in the case of PWRs, so much emphasis is put on explaining how well the heat removal from the secondary side is secured, while rarely anything is said regarding how the primary inventory - needed to enable heat transfer to the secondary side - is to be maintained. First of all, there's the question of the main coolant pump seal integrity. And even if they all would remain intact, even the allowable normal leak rate might lead to interruption of the heat transfer to the secondary side before the water supply to the steam generators becomes a limiting factor.

Not sure where you got that impression, primary inventory control is basically the central focus for all accident analyses, especially since TMI - "Keep the core cool, keep the core covered". As long as inventory and secondary heat removal is maintained, natural circulation is sufficient to protect the core. LOCA's are mitigated by safety injection pumps and accumulators. Small leaks are handled by the volume control system and charging pumps. In case of a large break LOCA, basically what happens is you just dump in water until the whole reactor cavity fills up.

 

[zapperzero]

US Operators receive far more and rigorous training, have more expertise, and take an individual responsibility for protecting the plant. Japanese operators tend to rely on a higher authority to tell them what to do.

This is "Can't happen here" syndrome in full swing. Sorry to be so blunt.

Also, relying on higher authority is exactly what Yoshida did not do (to his credit, ultimately).

 

[QuantumPion]

This is "Can't happen here" syndrome in full swing. Sorry to be so blunt.

Also, relying on higher authority is exactly what Yoshida did not do (to his credit, ultimately).

I never suggested it could not happen here. Indeed, it (core damage due to human error) has, multiple times. I merely suggested that we are better prepared for such an event.

 

[zapperzero]

My apologies. I should have written "won't happen here".

 

[QuantumPion]

My apologies. I should have written "won't happen here".

I never said it would not happen here either. I think you should reread my original post as you have completely misinterpreted what I said.

 

[zapperzero]

I will welcome a clarification, as it appears you are simply prejudiced against the Japanese and convinced of USian cultural superiority, for reasons you forgot to mention. Or so I am interpreting what you wrote.

 

[QuantumPion]

I will welcome a clarification, as it appears you are simply prejudiced against the Japanese and convinced of USian cultural superiority, for reasons you forgot to mention. Or so I am interpreting what you wrote.

This isn't prejudice, this is observation of factual differences and actual occurrences. Nuclear safety culture simply is not the same between the US/France and Japan/Korea/China. A notable exception is the site superintendent of Fukushima Daini, Naohiro Masuda, whom took active leadership role in protecting his plant (he is now head of the Decommissioning Company).

http://ajw.asahi.com/article/0311disaster/fukushima/AJ201401170010

 

[zapperzero]

This isn't prejudice, this is observation of factual differences and actual occurrences. Nuclear safety culture simply is not the same between the US/France and Japan/Korea/China.

If this were wikipedia, I'd slap a big fat [citaton needed] tag on that.

A notable exception is the site superintendent of Fukushima Daini, Naohiro Masuda, whom took active leadership role in protecting his plant (he is now head of the Decommissioning Company).

http://ajw.asahi.com/article/0311disaster/fukushima/AJ201401170010

Again, how about Yoshida himself? Ignored stupid orders from on high, dealt with three meltdowns and a few explosions, managed to not get anyone killed...