No strong conclusions can yet be drawn on the Fukushima Nuclear Crisis, because so much detail and hard data remains unclear or unavailable. Indeed, it will probably take years to piece the whole of this story together (as has now been done for accidents like TMI and Chernobyl [read this and this from Prof. Bernard Cohen for an absolutely terrific overview]). Still, it will definitely be worth doing this post-event diagnostic, because of the valuable lessons it can teach us. In this spirit, below an associate of mine from the Science Council for Global Initiatives discusses what lessons we’ve learned so far. This is obviously a huge and evolving topic that I look forward to revisiting many times in the coming months…
Guest Post by Dr. William Hannum. Bill worked for more than 40 years in nuclear power development, stretching from design and analysis of the Shippingport reactor to the Integral Fast Reactor. He earned his BA in physics at Princeton and his MS and PhD in nuclear physics at Yale. He has held key management positions with the U. S. Department of Energy (DOE), in reactor physics , reactor safety, and as Deputy Manager of the Idaho Operations Office.
He served as Deputy Director General of the OECD Nuclear Energy Agency, Paris, France; Chairman of the TVA Nuclear Safety Review Boards, and Director of the West Valley (high level nuclear waste processing and D&D) Demonstration Project. Dr. Hannum is a fellow of the American Nuclear Society, and has served as a consultant to the National Academy of Engineering on nuclear proliferation issues. He wrote a popular article for Scientific American on smarter use of nuclear waste, which you can download as a PDF here.
On 11 March 2011, a massive earthquake hit Japan. The six reactors at Fukushima-Dai-ichi suffered ground accelerations somewhat in excess of design specification. It appears that all of the critical plant equipment survived the earthquake without serious damage, and safety systems performed as designed. The following tsunami, however, carried the fuel tanks for the emergency diesels out to sea, and compromised the battery backup systems. All off-site power was lost, and power sufficient operate the pumps that provide cooling of the reactors and the used-fuel pools remained unavailable for over a week. Heroic efforts by the TEPCo operators limited the radiological release. A massive recovery operation will begin as soon as they succeed in restoring the shutdown cooling systems.
It is important to put the consequences of this event in context. This was not a disaster (the earthquake and tsunami were disasters). This was not an accident; the plant experienced a natural event (“Act of God” in insurance parlance) far beyond what it was designed for. Based on the evidence available today, it can be stated with confidence that no one will have suffered any identifiable radiation-related heath effects from this event. A few of the operators may have received a high enough dose of radiation to have a slight statistical increase in their long term risk of developing cancer, but I would place the number at no more than 10 to 50. None of the reports suggest that any person will have received a dose approaching one Sievert, which would imply immediate health effects.
Even ignoring the possibility of hormetic effects, this is approaching the trivial when compared with the impacts of the earthquake and tsunami, where deaths will likely come to well over 20,000. Health impacts from industrial contamination, refinery fires, lack of sanitation, etc., etc. may reasonably be supposed to be in the millions. Even the “psychological” impacts of the Fukushima problems must be seen to pale in contrast to those from the earthquake and tsunami.
The radiological impact on workers is also small relative to the non-radiological injuries suffered by them. One TEPCO crane operator died from injuries sustained during the earthquake. Two TEPCO workers who had been in the turbine building of Unit 4, are missing. At least eleven TEPCO workers were take to hospital because of earthquake-related physical injuries.
TEPCO has suffered a major loss of capital equipment, the value of which is non-trivial even in the context of the earthquake and tsunami devastation. They also face a substantial cost for cleanup of the contamination which has been released from the plants. These are financial costs, not human health and well being matters.
The Sequence of Events
Following the tsunami, the operators had no power for the pumps that circulate the primary coolant to the heat exchangers. The only way to remove the decay heat was to boil the water in the core. After the normal feed water supplies were exhausted, they activated the system to supply sea water to the core, knowing this would render the plant unfit to return to operation. In this way, the reactors were maintained in a relatively stable condition, allowing the water to boil, and releasing the resulting steam to the containment building. Since this is a Boiling Water Reactor (BWR), it is good at boiling water. Operating with the water level 1.7 to 2 meters below the top of the core, they mimicked power operation; the core normally operates at power with the water level well below the top of the core, the top part being cooled by steam. Cold water in, steam out, is a crude but effective means of cooling.
Before using sea water, according to reports, water levels are thought to have dropped far enough to allow the fuel to overheat, damaging some of the fuel cladding. When overheated, the cladding (Zirconium) reacts, claiming oxygen from the water. Water, less oxygen, is hydrogen. When vented to the containment and then to the outer building, the hydrogen built up, and eventually exploded, destroying the enclosing building. With compromised fuel, the steam being vented contains radioactive fission products. The design of BWRs is such that this venting goes through a water bath (in the Torus), that filters out all but the most volatile fission products.
With time, the heat generated in used fuel (both in the core and in the fuel pool) decreases. From an initial power of about 2% of full power an hour after shutdown (when the coolant pumps lost power) to about 0.2% a week later, the amount of steam venting decreases, and releases can be controlled and planned for favorable weather conditions.
A second concern arose because of the inability to provide cooling for the used-fuel pool in Unit 4, and later Unit 3. The Unit 4 pool was of concern because, for maintenance, the entire core had been off-loaded into the pool in November (it is believed that two older core loadings were also in this pool, awaiting transfer to the central storage pool). With only a few months cooling, the residual heat is sufficient to raise the temperature of the water in the pool to boiling within several days or weeks. There is also some suggestion that the earthquake may have sloshed some water out of the pool. In any case, the fuel pools for Units 3 and 4 eventually were thought to be losing enough water such that the fuel would no longer be adequately cooled. Since the fuel pools are outside the primary containment, leakage from these pools can spread contamination more readily than that from the reactor core. High-power water hoses have been used to maintain water in the fuel pools.
While many areas within the plant complex itself, and localized areas as far away as 20 Km may require cleanup of the contamination released from the reactors and from the fuel pools, there is no indication that there are any areas that will require long term isolation or exclusion.
It is not the purpose of this paper to anticipate the lessons to be learned from this event, but a few items may be noted. One lesson will dominate all others:
Prolonged lack of electrical power must be precluded.
While the designers believed their design included sufficient redundancies (diesels, batteries, redundant connections to the electrical grid), the simultaneous extended loss of all sources of power left the operators dependant on creative responses. This lesson is applicable both to the reactor and to fuel pools.
All nuclear installations will probably be required to do a complete review of the security of their access to electrical power. It may be noted that this lesson is applicable to many more activities than just nuclear power. Extended loss of electrical power in any major metropolitan area would generate a monstrous crisis. The loss of power was irrelevant to other activities in the region near the Fukushima plant because they were destroyed by the tsunami.
Other lessons that will be learned that may be expected to impact existing plants include:
Better means of control of hydrogen buildup in the case of fuel damage may be required.
In addition, detailed examinations of the Fukushimi plants will provide evidence of the margins available in seismic protection. Detailed reconstruction of the event will give very helpful insights into the manner that fission product can release from damaged fuel, and their transport.
Applicability of Fukushima Information to MOX-fueled Reactors:
The core of Unit 3 was fueled with plutonium recycled from earlier used reactor fuel. Preliminary information suggests that the release of hazardous radioactive material, for this type of event, is not significantly different than that non-recycle fuel. More detailed examinations after the damaged cores are recovered, and models developed to reconstruct the events, will be necessary to verify and quantify this conclusion.
Applicability of Fukushima Information to Gen-III Reactors:
In the period since the Fukushima plants were designed, advanced designs for BWRs (and other reactor types) have been developed to further enhance passive safety (systems feedback characteristics that compensate for abnormal events, without reliance on operator actions or on engineered safety systems), simplify designs, and reduce costs. The results of these design efforts (referred to as Gen-III) are the ones now under construction in Japan, China and elsewhere, and proposed for construction in the U.S.
One of the most evident features of the Gen-III systems is that they are equipped with large gravity-feed water reservoirs that would flood the core in case of major disruption. This will buy additional time in the event of a Fukushima type situation, but the plants will ultimately rely of restoration of power at some point in time.
The applicability of the other lessons (hydrogen control, fuel pool) will need to be evaluated, but there are no immediately evident lessons beyond these that will affect these designs in a major way.
Applicability of Fukushima Information to Recycling Reactors:
As noted above, Unit-III was fueled with recycled plutonium, and there are no preliminary indications that this had any bearing on the performance of this plant during this event.
Advanced recycling, where essentially all of the recyclable material is recovered and used (as opposed to recovery and recycle of plutonium) presents a different picture. Full recycling is effective only with a fast reactor. A metal fuel, clad in stainless steel, allows a design of a sodium-cooled fast reactor with astonishing passive safety characteristics. Because the sodium operates far from its boiling point in an essentially unpressurized system, catastrophic events caused by leakage or pipe failures cannot occur. The metal fuel gives the system very favorable feedback characteristics, so that even the most extreme disruptions are passively accommodated. A complete loss of cooling, such as at Fukushima, leads to only a modest temperature rise. Even if the control system were rendered inoperable, and the system lost cooling but remained at full power (this is a far more serious scenario than Fukushima, where the automatic shutdown system operated as designed) the system would self-stabilize at low power, and be cooled by natural convection to the atmosphere. Should the metal fuel fail for any reason, internal fission product gases would cause the fuel to foam and disperse, providing the most powerful of all shutdown mechanisms.
The only situation that could generate energy to disperse material from the reactor is the possibility of s sodium-water reaction. By using an intermediate sodium system (reactor sodium passes its energy to a non-radioactive sodium system, which then passes its energy to water to generate steam to turn the electrical generator), the possibility of a sodium-water reaction spreading radioactive materials is precluded.
These reactors must accommodate seismic challenges, just as any other reactor type. While there are many such design features in common with other reactor designs, the problem is simpler for the fast reactor because of the low pressure, and the fact that this type of reactor does not need elaborate water injection systems.
In light of the Fukushima event, one must consider the potential consequences of a massive tsunami accompanying a major challenge to the reactor. Since it may be difficult to ensure that the sodium systems remain intact under the worst imaginable circumstances, it may be prudent to conclude that a tsunami-prone location may not be the best place to build a sodium facility (whether a nuclear power plant or something else).
The major lesson to be learned is that for any water-cooled reactor there must be an absolutely secure supply of power sufficient to operate cooling pumps. Many other lessons are likely to be learned. At this early point, it appears that design criteria for fuel storage pools may need to be revised, and hydrogen control assessed.
Given the severity of the challenge faced by the operators at Fukushima, and their ability to manage the situation in such a way as to preclude any significant radiation related health consequences for workers or the public, this event should be a reassurance that properly designed and regulated nuclear power does not pose a catastrophic risk to the public—that, overall, nuclear power remains a safe and clean energy sources.
Given the financial impact this event will have on the utility (loss of four major power plants, massive cleanup responsibilities), it will be worthwhile for the designers, constructors, operators, and licensing authorities to support a thorough analysis of what actually transpired during this event.
167 replies on “Preliminary lessons from Fukushima for future nuclear power plants”
question for Moderator
Can we have direction on where updates, questions on updates and related technical analysis should be posted?
At the moment this stuff is scattered between 3 threads (open thread, preliminary lessons and 10+ days)
I would suggest updates, technical analysis of updates and questions on new developments should all go to 10+ days
I have passed this suggestion on to Prof Brook for his attention. It would be a full-time job to meticulously read each comment and make decisions as to the best possible thread and I am one part-time volunteer. Sometimes topics overlap threads anyway – a subjective judgement really.
Tsunami Risk Well Known to Nuclear Engineers, Regulators Who Failed to Act
Japan’s nuclear regulators and the operator of the crippled Fukushima reactors were warned that a tsunami could overwhelm the plant’s defenses and failed to recognize the threat.
The Trade Ministry dismissed evidence two years ago from geologists that the power station’s stretch of coast was overdue for a giant wave. Tokyo Electric Power Co. engineers also did not implement lessons from the 2004 tsunami off Indonesia that swamped a reactor 2,000 kilometers (1,200 miles) away in India, even as they advised the global nuclear industry on how to cope with the dangers.
“Where did the hydrogen come from”
Do we know where the atmosphere above the lead/acid batteries was vented?
All lead/acid batteries vent hydrogen when discharged and the plants had an awful lot of them.
Apparently the workers burnt were not wearing industrial boots:
‘The two hospitalised men, employees of a Tepco affiliate, were part of a team of six workers attempting to connect a water pump to the power supply and restart the supply of fresh water in an attempt to cool the reactor. Workers in the first and basement floors of the No 3 reactor’s turbine building were ordered to evacuate the area after the accident.
Japan’s nuclear safety agency said water had probably seeped through their protective clothing, allowing radioactive materials to stick to their skin, as they stood in a 15cm-deep puddle. The two injured men were wearing shoes, while the third had boots on and so escaped serious injury.
Radiation levels on the surface of the puddle were later measured at 400 mSv per hour, while the level in the air reached 200 mSv per hour.
The source of the water was not immediately clear. Tepco said no puddle had been spotted in the turbine building the previous day. Fire trucks have been dousing the reactor in recent days in an attempt to cool a storage pool for spent fuel rods.
The accident cast doubt on Tepco’s ability to properly monitor radioactivity at the site. “This kind of exposure, from water, was unforeseen,” the government’s chief spokesman, Yukio Edano, told reporters.
“Atmospheric radiation levels are monitored constantly, but in this case the workers stepped into water. We are trying to find out exactly what happened so we can ensure it doesn’t happen again.”‘
As Red_Blue mentioned, is there a reason for Th, U, Pu, etc, not being in TEPCO’s report? How can you have such extremely high levels of products without even a trace of fuel?
@David Martin, re: NS article.
Omg! There’s either a big mistake with the numbers, or the author can’t do maths!
Oh, man, I wish I’d heard that stuff about the two guys being contaminated last night… we had a safety standdown at work this morning, and it would have been a great example of what not to do!
1) Incorrect PPE; and
2) Ignoring warnings from instruments designed to warn of a hazard.
New Scientist is comparing one day at Fukushima with 10 days at Chernobyl. The daily amounts are close, so the Chernobyl totals are around one magnitude higher.
I don’t see any problems with New Scientist’s numbers.
You are in the right of it. Retracted. Serves me right for not having enough coffee on board before looking at it! ;-)
Ah yeah, I made the same mistake with the NS numbers. So back to my earlier question: how has there been such a large release. Considering to possible graphite fire, and containment?
“Tokyo Electric Power Co. said Friday it has begun injecting freshwater into the No. 1 and No. 3 reactor cores at the crisis-hit Fukushima Daiichi nuclear plant to enhance cooling efficiency”
“although highly radioactive water was found leaking possibly from both reactors as well as the No. 2 reactor.”
(posted with different IP, using a proxy, as internet has been working oddly in China. Cause >
“ocean cables damaged in the Japan earthquake”
[…] E ciò a dispetto del fatto che, come ha notato fin da subito Oscar Giannino (qui e qui) e come altri hanno evidenziato ora che i fatti iniziano a essere chiari, se messo nel contesto del terremoto che […]
@Shelby, on 25 March 2011 at 10:29 PM
Managment seems more risky than technologies involved on this debate.
It seems that studies involving the WWII nuclear fallout in Japan indicates that people exposed to low levels of radiation *live longer* and have *less* cancer than their non-irradiated coutnerparts.
It’s the Hormesis effect: http://en.wikipedia.org/wiki/Hormesis Here’s a copy of the article:
Lawrence Solomon: Japan’s radioactive fallout could have silver lining
Lawrence Solomon Mar 21, 2011 – 10:56 PM ET
“The study’s bottom line: “the low doses of A-bomb radiation increased lifespan of A-bomb survivors.””
I’m inclined not to take at face value such ballpark initial estimates of the volatile Cs-137 and I-131 source terms at Fukushima which are based on measurements taken at CBTBO monitoring stations, some of which are many thousands of miles away from the source, so that time delays, winds, and weather patterns all the way in-between have to be taken into account.
Surely these estimates might be either high or low, based on the particular modelling that was used to extrapolate the point measurements of activity to get the total activity in the plume, and to plot that activity as a function of time as well: but if we take the NS article as accurate and complete, then it seems Wotawa attaches no error range whatever to his estimates.
Altogether, that article surely doesn’t look to me like a scientific discussion of the data.
For estimating the size of the source term, I would have thought it would be far better to rely on measurements of volatiles observed NEAR to Fukushima.
In addition, it’s mentioned in the article that at Chernobyl far more fission products than Cs-137 and I-131 were observed. Yet Wotawa apparently does not report levels of any other products. Does that mean that there were none observed?
Shelby, on 25 March 2011 at 10:29 PM said:
This was one of my first thoughts and I posted that somewhere maybe here (remembering back 2 wks), that a report would turn up underlining the tsunami risks for these reactors. As with all these reports, someone then makes a determination if the likelihood is great enough to spend money on correcting. They bet it wasn’t. Their decision will be analysed to see if this was reasonable. Given the tsunami risk (not large earthquake risk) I would say not.
Since the guest post included some comments on relative safety of the sodium-cooled fast reactor, I’m interested again in what is known regarding the response that is likely from a molten salt reactor. I have read in the past that a very primitive safety system consisting of an electricity-cooled solid plug at the bottom of the reactor, which melts on a loss of power and allows the molten salt/fuel combination to spread in a catch basin. Will this type of design prevent any further damage from decay heat, etc.? Any of the local experts care to comment?
ParetoJ, on 26 March 2011 at 3:41 AM said:
One of the criticisms of the using nuclear fallout survivors in Japan for the Hormesis effect is that they may of been a selected population, i.e. the weaker people exposed to the bomb and immediate effects thereafter died, so you have people left who have higher natural immunity and whatever other characteristics that helped them survive. Given that the linear no-threshold model is a worst case model. As an analogy lets say you tap one million people on their hand. No effect right? Now the linear no-threshold model would say that would have the same effect as taking that energy and applying it to a few people, which would shatter their hand. Exposing a large no. of people to low dose of radiation is the same as exposing a few people to a large dose. Not very credible either.
@ DocMartyn 25 Mar 10:31 PM
Thanks for this idea on H2 generation from the numerous discharged batteries. I think it depends on where the batteries are located.
If batteries were in R4 reactor building, batteries could have contributed H2 if vented to building top.
If batteries are in the separate turbine hall building, then no.
I’d guess that the batteries are in the turbine hall because that’s where the steam turbines, their surface condensers for condensing turbine steam and the main cooling water pumps (driven by batteries) are.
However, that’s just a guess. Somebody with actual experience would have to tell us where the batteries are located.
Now that I think of it, this discussion should be on the open comment thread.
Agreed William Please everyone switch your comments on Hormesis topic to Open Thread.
Thank you, Dr. Hannum, for a very complete overview of events at Fukushima. I have been involved with industrial chemistry for many years, and I take some issue with your acceptance of molten sodium as safe. In extreme events such as a 9.0 earthquake, it would not be out of the realm of possibility to experience a leak, possibly a large leak, of hot liquid Na. Even atmospheric water would be enough to generate hydrogen, and if there were a quantity of liquid water, it’s very likely there would be a large hydrogen explosion. All this can possibly designed-around, but I wouldn’t categorize it as “intrinsically” safe.
Since the safety systems kicked in and shit down the reactors, it sends much of the risk was posed by the spent fuel pools not being within containment. I have heard that spent fuel can remain as high as 90% radioactive. I am just an interested reader and not a nuclear scientist so I dont know if that is true. In any case, it seems we would want containment for the fuel pools. Is there a reason that was not part of the original design?
The analysis by Gerhard Wotawa written up in the New Scientist is interesting in one respect: the levels of iodine relative to caesium suggest that the reactor cores of units 1, 2 and/or 3 are the source (and pretty much the sole source) of the long-range release – leaving its size to one side for now. Iodine-131 in 5 & 6 and in all the spent fuel pools would be basically non-existent, but caesium-137 would still be present at only slightly-reduced levels compared to the reactor core.
With respect to the New Scientist article: it’s always a good idea to
check the primary sources. The actual conclusions from Wotawa’s website
My translation is (apologies, if there are errors, my german is quite
a bit rusty):
So: there are no data as yet beyond three days, even from these, in my
opinion, questionable estimation methods, and the New Scientist
article should therefore not integrate the source terms up to ten
These statements by Dr. Hannum need to be corrected. I raised the issue once before above, and nobody has followed up on it. But it is actually one of the main issues to be considered for this accident (venting and hydrogen management). There is no vent that leads directly into a major working area of the reactor building where fuel loading takes place and the spent fuel pond is located. This would be a major oversight in design, since this area is not kept free of oxygen and is no place to vent dangerous radioactive steam and gases.
E&E Publishing (appearing in the Times) looks at this issue today: “U.S. Experts Blame Fukushima 1 Explosions and Radiation on Failed Venting System.” They consider two alternatives. The first is the one I initially suggested (based on Mitsuhiko Tanaka’s account of the accident presented at an expert panel event in Japan): a failure of the primary containment vessel at the hatch for the primary containment vessel which is designed to withstand 4 atmospheres of pressure (when the primary containment vessel contained 8 atmospheres at the time, or twice the amount of design tolerances). The second is a failure of the venting system, which works by moving steam and radioactive gas from the primary containment vessel through the reactor building, and to the large 100 meter emissions stacks located outside the power plants (which can filter and diffuse the gas to minimize it’s environmental impact). But as it turns out, these stacks require fans to work properly, and thus electrical power. The speculation is that “most, if not all, of this dangerous mix of hydrogen gas seeped into the reactor building in Units 1 and 3. The hydrogen, being lighter than air, mixed with air in the upper large refueling floor area” (and subsequently exploded). Clearly, this is going to be a vital issue in the subsequent dissection of this accident, and should give rise to an “industry re-think” (as suggested by em1ss above).
Any additional thoughts on this important (and central issue) to this accident, are obviously welcome.
Standby gas treatment system requires fans to work when it’s taking air from secondary contaiment (reactor building outside of the primary containment vessel), since this air is normally at slightly under atmospheric pressure. If a steam/gas mixture of very high pressure (the kind of pressure that would threaten primary contaiment integrity) were to be discharged directly to SGTS, it would damage SGTS and probably leak to the auxiliary building instead of being vented out of the stack.
The direct hardened vent installed to Fukushima reactor buildings doesn’t require any fans, if the discharged steam/gas mixture is of high enough pressure. The pressure must be high enough to break the rupture disk valve anyway.
There are several scenarios of why this might not have been used, but we don’t know yet even if it was used or not. I can think of at least the following:
– damage to valves from the quake or tsunami, and/or structural damage preventing access to the valves for manual operation
– venting attempted at below the rupture disk pressure, or leaks before the rupture disk
– deliberate attempt to reduce activity release by venting to the secondary containment instead, perhaps hoping that SGTS could be restored before reaching explosive concentration
(the only “filter” for the direct vent line is that any material from the core would have gone through the suppression pool water first, with boil off from that water being of much lower activity)
Before we can assess whether a containment vessel leak is to blame or venting operations, or even both, more information needs to be secured. Radiolysis from the pools as one source should not be completely overlooked either, as in normal operation that is effectively removed from the secondary containment air when ventilation (at a rate of about one full exchange per day) is working.
In one of the threads there was mention that the Brown’s Ferry secondary containment vessel could not hold pressure at 70 psi (~4.5 atm) when they tried to pressure test it during construction. It leaked at the flange at the top of SCV. When they reduced pressure to 65 psi it held.
Then somewhere else in these threads there is a long video presentation by one of the design engineers. In it he postulates that steam and hydrogen being vented from the PRV resulted in the reported SCV pressure of 8 atm (~120 psi) which is about double the design pressure of 3.7 atm. The hydrogen being the lightest component collected, at high concentration, at the top of the SCV under the dome and leaked out at the flange when the SCV internal pressure doubled the design pressure.
This leakage would have vented the high concentration hydrogen gases directly into the operating floor penthouse at the top of the reactor building where it ignited with explosive force.
His theory seems to give a credible explanation as to how the hydrogen got to the top floor of the building.
I suck at searching in these forums so no links but I will post them when I find them.
I see that ELs post has the link to the presentation I mentioned. http://www.ustream.tv/recorded/13410573
For the life of me I can’t find the Browns Ferry link that mentions the 65 vs 70 psi pressure test, but it is here somewhere
Anyone seen this?
re post by: Red_Blue, on 25 March 2011 at 7:06 PM:
Red_Blue, thank you for your reply! You are right that I was apparently flipping the units, and that it was 1 & 2 that lost ECCS about an hour after the tsunami.
You may have in part answered my question – I was under the impression from a few different sources such as NRC that the RCIC ran solely by turbine – no battery or AC needed. From US NRC Reactor Concepts Manuel, Boiling Water Reactor Systems: http://www.nrc.gov/reading-rm/basic-ref/teachers/03.pdf
I also found a DBA scenario for loss of offsite power (LOOP) + guillotine main line break – so perhaps not totally applicable but I would think close, since it appears the initiation of the emergency systems we’ve been discussing is the SCRAM + main line isolation. http://www.enotes.com/topic/Boiling_water_reactor#The_safety_systems_in_action:_the_Design_Basis_Accident It states:
But you’re saying that RCIC requires batteries for the required valves? If so, then I’m still confused because I thought that all three units had batteries?
Let’s say that unit 1 & 2 had already shifted to the lower pressure RHR… both had battery power to run RHR for some time, didn’t they? So why would RHR have failed? Why would pressure have increased beyond what RHR could handle?
Even if pressure did increase past what RHR could handle, couldn’t they have just switched back to high pressure RCIC?
So, we’ve got:
RCIC – high pressure, ??? driven – turbine driven (per NRC), battery (enotes), initiation on low RPV water level or manually
HPCI – high pressure, diesel driven, initiation on loss of coolant
RHR – low pressure, ?? driven, init ??
LPCS – low pressure, ?? driven, init ??
LPCI – low pressure RHR + CS, diesel driven, init ??
So I’m still confused…
Articles like this drive me crazy:
It sounds oh-so-bad, those rotten people should have been better prepared for tsunami, they were warned!!
Only a few problems if one reads with a bit of a well justified critical eye. First, the ‘well established’ historical precedent tsunami they note is specified for Sendai Bay and they don’t bother to tell us how high it was at the Fukushimi-1 & 2 sites where it could very very easily have been tremendously smaller. Next, we all tend to think “if it could be THAT large at x location, of course it could be at y location also” Only that isn’t true at all, because it all depends on where the fault lines are, how the ocean floor lays between those locations and the site of interest, what the elevation is at the site of interest, etc.
So, if the next one was overdue, what would it matter if historically it wasn’t higher than the 5.5m already planned for at Fukushimi? But they never address that aspect, do they?
But along those lines – I gather that the primary buildings & diesels were about 10m above sea level… where was the 5.5m tsunami/sea wall relative to those structures?
Finally – they note that the historical information and warning about tsunami risk was presented. That it was deemed to not be a significant risk apparently. In cases like this where hindsite picks out one or even a few people who warned about some catastrophe and others didn’t act on it, all too often it turns out that it wasn’t acted on NOT because of greed, or negligence, but because a number of experts evaluated the situation and didn’t feel that it was nearly the risk being portrayed. Only reports like this never bother to go into the OTHER side of the story, they only present the one that puts everything into the worst possible light.
Now I know that isn’t always the case, and it may not be in this situation – but all too often it is. When that is the case, you rarely (if ever) see a retraction of the original story, or even a new story explaining the other side. Its just not sexy enough.
It’s like all the flak about the Mark 1 containment – but if you read more about it, from the US NRC, and even GE for example, to get both sides, you find out that things aren’t anywhere near what they’re being portrayed to be in the general media.
The inaccuracy, the misrepresentation, the ambiguous nature, the FUD tactics, all drive me just about crazy.
You just beat me to it:) I was just about to ask you for the link which you have now provided above:)
On the “articles that drive me crazy” post I just did at 11:47 am, I had meant to add in the link to the article. It was from a post by: Shelby, on 25 March 2011 at 10:29 PM
Oh, and it also said:
So a reactor got swamed from a tsunami that originated far away – again, we would have to know if there was any credible risk to the fukushimi site for something of that nature to occur. Maybe the example they use was a reactor that got swamped by a 0.5m tsunami – if so, how is that relevant to the Fukushimi sites? But do they bother to tell us any of the key facts? NO. What ever happened to the idea that journalists were to answer the 5 W’s (who, what, where, when, why), and present facts in context?
> Rational Debate
> we would have to know if there was any credible
> risk to the fukushimi [sic] site … But do they
> bother to tell us any of the key facts? NO.
O RLY? read the article. It says in part:
“… evidence two years ago from geologists that the power station’s stretch of coast was overdue for a giant wave”
and goes on to give details including
“sediment samples showed the tsunami had a pattern of recurring every 800 to 1,000 years, according to a 2001 report by a research team funded by the government’s Science Ministry.”
“Yukinobu Okamura, who heads the government-funded Active Fault and Earthquake Research Center, asked Tokyo Electric why it hadn’t taken on board evidence of the tsunami risk.”
Please stop debating. Science bloggers — even many science blog readers — often do check sources. Debating points using claims so easy to refute doesn’t work well in science discussions.
re post by: Hank Roberts, on 26 March 2011 at 1:08 PM:
Except you haven’t refuted anything I said, Hank. The story never ties the historical evidence to the Fukushimi (or other power station) site, exactly as I already stated.
As I noted also, the story mentions a single person or group, Okamura etc., without bothering to present the OTHER side of the story – which may well shed an entirely different light on the subject.
As to checking sources, that is exactly what I expect – and I further expect that not only science types, but hopefully everyone will check not only the original story, but read critically as I’ve noted and then check the other side of the story before jumping to conclusions. All too often that doesn’t appear to be occurring. Journalists ought to be doing that for their readers – but they generally fail miserably on that account whether intentionally or not.
Joffan, on 26 March 2011 at 5:42 AM said:
I agree. It does seem that Wotawa’s study is more interesting at present as which isotopes are seen and not seen in the plume and their relative levels. If there had been large scale oxidation of zirconium cladding in the number 4 spent fuel pool, which I understand contains a full core, offloaded about 100 days ago, then I would expect to see some activated zirconium in the plume. Zr-95 can be produced from Zr-94 by neutron absorption, and Zr-94 is about 17% of natural zirconium. It has a half-life of 64 days and its decay scheme has several characteristic gammas in Nb and Mo.
[…] Posts Preliminary lessons from Fukushima for future nuclear power plantsFukushima Nuclear Accident – a simple and accurate explanation10+ days of crisis at the Fukushima […]
I don’t think there were any large pipe breaks or other leaks, since those would have required raising the alarm with NISA, which didn’t happen until tsunami struck.
There are many possible loads in the battery bus and how long the batteries last depends on which loads are useds, but typical estimates are 6 to 8 hours. When the plant is intact, it’s possible to operate almost all valves (outside of containment) manually, albeit with some difficulty. When conditions detoriorate, it might be impossible to reach such valves because of water leaks, high temperatures or dose rates etc. So when batteries to operate them motor driven or compressors to operate pressurised air valves run out, options become much more limited.
RHR (to be more precise, at least 1 RHR pump, and one service water pump) requires AC power and cannot be run with batteries. In general, there are no electric pumps that run on batteries, batteries are only good for limited amount of instrumentation and control.
You should really look up descriptions of these systems, their interrelations are more complicated than that. RCIC is used for the immediate phase of cooling when the turbine and reactor trips and the condenser and its (seawater in this case) cooling loops are no longer available. RCIC dumps the reactor heat as steam to the pressure suppression pool.
RHR/LPCI is then used to actually remove heat from containment to outside of the unit with the aid of the “ultimate heat sink”, seawater in this case. RCIC and HPCI cool just the core and dump the heat to the pressure suppression pool, from where it must be eventually removed. In any accident where feedwater pumps and RHR are completely unavailable, the third and last option to cool contaiment is to vent.
If at some point the RPV pressure had gone above what RCIC/HCPI turbo pumps (using steam turbines) require, that would not have helped to cool the pressure suppression pool, but to heat and drain it further instead. Whether that would have been a good idea at a certain point of time would have depended on the temperatures and pressures of the RPV and PSP as well as the RPV water level.
Without RHR you essentially have a closed system which is going to keep heating up regardless of how you circulate the water and steam back and forth inside. Which is why they had to vent to reduce the pressure and then introduce outside water source (starting with the unit fire pump) to make up for the boiled water.
Thanks again for your response and the information. I have been trying to look up information on these systems, that is where the links I posted came from – descriptions of these systems. I’ve had trouble finding much that is more detailed or useful – and am certainly open to any that you or anyone else can provide.
I understand that RCIC/HCPI only cool the core and eventually the close looped system of water becomes too hot for the cooling to be useful. I thought that I had said that – but I’ve been asking how long they should be expected to be functional BEFORE that occurs, after a scram? Is it really only an hour or two?
I’d also like to find out information about just what each system can be operated off of (e.g., decay heat steam turbine, diesel, battery), just what their operating parameters are, and that sort of information. If you know of any good information online that way, I’d be most grateful for links.
Thanks again – I have really appreciated your posts on a number of different subjects and across several threads.
6-8 hours assuming the limiting factor is battery power. If DC power is secured (by aggressive load balancing and by bringing in replacement batteries/DC chargers connected to portable generators), then suppression pool pressure becomes the limiting factor. I haven’t seen any modelling of that exact scenario, but it should not be a difficult calculation. I think you could do a ballpark figure by simply taking the core thermal output curve from decay heat and applying that to the suppression pool volume at normal temperature. I don’t think that time is going to be higher by days though, maybe extending it by 10-15 hours.
It’s likely though that RCIC/HPCI pump cavitation will start before the suppression pool will rupture, but there are probably significant unknowns in these strengths, because unlike the reactor pressure vessel, I’m not aware of any desctructive “until failure” testing being done to those systems. They have estimated failure pressures that are usually quite conservative, meaning that the actual failure could occur at much higher than anticipated pressure, but will not occur before that. Then there are also large safety margings between that and the “design pressure”.
There is suspicion that the suppression pool of unit 2 suffered a leak, but there is no confirmation or information how it happened or how serious it is.
@Rational Debate 4:36PM
I hesitate to put out this calculation in the presence of the experts here. But perhaps they will correct my numbers and give the answer you’re looking for.
Please feel free. This is a semi-informed, back of the envelope calculation.
My understanding of the turbine driven emergency cooling is that it uses the high pressure steam from the reactor to drive a turbine driven pump to provide pressure for coolant circulation. The exhaust from that turbine is vented into the suppression pool. Such a closed system will heat the suppression pool whose pressure will rise until no more steam can be exhausted into it and the turbine will stop.
From this reference:
I got that the minimum volume of the suppression pool was 105600 ft^3 and the condensate tank 400000 gal for a total of 4504m^3. If it starts at 20C and heats to 100C with a specific heat of 1cal/gm, it will absorb 1.509*10^12joule.
From this reference:
The 1GW reactor is making somewhere between 200MW and 16MW in the period between the scram and 1 day after.
I gave it 40MW just to have a number in there.
40MW goes into 1.509*10^12joule 37720sec or 0.437 day.
Given my expertise in this field, this is pretty much just a numberized WAG but there you have it.
Corrections are welcomed.
Has anybody considered using a big thermoelectric generator to run cooling pumps? The efficency is pretty dismal, but in this circumstance I don’t see that as an issue. The advantage is having no fluid conducting openings to the heat source, just wires. Now your heat exhaust (cold junction) can be the ocean directly.
At the very beginning of this event i realised something that i saw as very good. Under the circumstances it seemed ill advised to comment this, but now it may be a better time. The good thing is that in all of 40 years the Japanese did not get sloppy on safety. I even recall an article on Tepco adressing its workforce on awareness for the dangers of getting sloppy.
To my understanding there is nothing ‘offcourse’ about this. During the fall of the Soviet Union the Bulgarians had quite some difficulty in maintaining safety standards, relying to some extend on volunteers to keep their plants running. I recall seeing a German tv documentary on this subject only. Never the less the event in Fukushima proves guaranteing safety level maintenance over decades is evidently realistic, which is (my opinion) a good thing.
Another thing the Fukushima event proved is the usefullness of extradiciplinary technology. It has been stated before: the presence of very little relatively cheap equipment would have made a big difference here. A difference in attitude or approach of the safety ussue may prove very productive. The box didn’t hold so lets think beyond it. If hammer-and-plyers-technology saves the day, it might be wise to invite it to stay.
A third point is the damage assesment. Nuclear science has the tendency to walk away from responsability for effects such as global panic or at least not see this as ‘their’ part of the damage. I consider this not only unrealistic, but also inproductive. A sound clear and conclusive public information policy greatly reduces this damage. I see myself sharing a problem with the pro-nuclear society there. Lack of competence in conclusive communication. To my knowledge adressing this problem reduces actual damage way faster than any on site safety measure. The answer to the question if this event is a victory or a defeat for nuclear energy relies almost entirely on the way it is presented to the public. Thank you.
Ernie Hamilton, on 26 March 2011 at 9:43 AM said:
I posted a seconday source for this a couple of times in the last week or so. You will find it at the following site with a lot of other information on for example the SFPs.
Go down to “How Hydrogen May Have Gotten from Primary Containment into the Reactor Building”
Thanks for your reply. Do you have any knowledge of a venting pathway that would lead from the primary containment vessel and intentionally to the secondary containment structure? You are assuming this is possible to do (but I haven’t seen any indication that this is the case from reactor plans or second hand commentary). Why would you engineer a pathway to intentionally flood a primary working area of the building with dangerous radioactivity and gases from the primary core containment structure. You are suggesting this is done to diffuse the gas so it can be taken up by the standby gas treatment system, but this area is not kept free of oxygen with inert gas, and the hard vent was installed exactly for the function you describe (and without the consequence of damaging the reactor building and rendering one room of the site permanently uninhabitable)? If this is an option, how can this be done from the control room.
There is the possibility that the venting wasn’t intentional, as my answer to Ernie Hamilton shows that the primary containment could leak.
@ William Fairholm
That explains why I could not find the link. It was the Brunswick reactor… Not the Brown’s Ferry plant.
On another blog it has been suggested to change the cooling agent for the pool from water to air all together in emergency situations. Some drawings published indicate that above the pool there is free space up to the roof unless the crane is there. A fall-in-place chimney above the pool was suggested. If the water level reaches the cores, the pool is shock drained allowing large amounts of air to flow in at the bottom. It would not keep all radioactivity inside, but allow less to escape than in the current situation. According to some argument on this blog, the draft would be powerfull enough to keep the cladding from melting or interacting with steam and its passive, thus ruling out the need for outside power to keep it stable. Another advantage would be the escape from polluting the pool with seawater. Closing the bottom while shock filling the pool might put it right back in business. I have no means to do the math on this option, but it does not sound entirely impossible to me.
According to “rational”
” they’ll sue a driver for merely tossing a bit of trash out the window”
Is this typical or extreme?
Why use extreme instances to consider mainstream issues?
What is your evidence?
re post by: Mike, on 27 March 2011 at 10:16 AM said:
Mike, I’m sorry, but i have no idea what you are trying to get at. Also, note that my post said In general people aren’t unreasonably afraid of cars such that they’ll sue a driver for merely tossing a bit of trash out the window. So its not a good analogy. Basically the opposite of in meaning than if one only reads or considers the part of my sentence that you extracted….
@Red_Blue and DrD re the RCIC system parameters.
Thank you both for your posts. DrD, I’ll have to look at the links you provided.
If we assume battery isn’t the limiting factor because they’re available etc, then I thought that the next limiting factor to the system was temperature thoughout the water volume, not just in the suppression pool or torus, but all -e.g., suppression pool + torus + RPV…. and that it was only useful until the total volume of water in the loop was too hot to successfully cool the core?
It is, afterall a loop – the steam line runs thru turbine and dumps into the suppression pool, but that turbine drives the RCIC pump, which pulls water from both condensate tank and suppression pool, routing it right back into the RPV….
Red_Blue, I could also see how in that system if the pressure capability of the pool would/could come into play, but wouldn’t steam just be blown off/vented sufficient to keep that from being the limiting factor?
@Rational Debate: Remember, the whole plant was flooded with electrically conductive seawater when the tsunami hit. Who knows what systems were shorted out when that happened. The batteries may be designed to run for 6-8 hours, but if circuitry they’re hooked up to got shorted by seawater, they could easily have been drained MUCH quicker than that.
Some loss of battery power due to seawater short circuit must have taken place, assuming they were exposed at all. The extend varies enormously depending on the construction of the batteries themselves. My experience is that they are to quite some extend protected against it. At Tudor we covered the cell connectors with hard fat for that reason. It all depends on how much and for how long conducting surface was exposed and that varies from insignificant to conclusive. Again, it is custom to construct emergency backup systems to be able to survive the circumstances in which they may be needed.
The forementioned meaning a sensably designed emergency battery pack or cell array should be able to hold on to its capacity and remain operational, even after being fully immersed in sea water. Thus supporting Rational Debate’s assumption. Whether the rest of the circuitry does that too may be unsure, but that does not significantly influence the batteries capacity.
@bchtd1parrot: Fair point, but I was more talking about the seawater getting to things like pumps and valves the batteries were connected to, rather than to the batteries themselves. I was kind of assuming that if seawater got to the batteries themselves and they didn’t have any protection they’d be dead almost instantly.
Seawater getting to the batteries themselves is very unlikely capable to cause immidiate servere damage. You must understand the short circuit in that case is due to electrolytic process, not conduction like with metal. Pure water does not or hardly conduct at all. Its the salt that does the trick. Thats why the connecting surface matters and because it does, battery array manufacturers tend to close that up by sealing as much of it as possible. We used warm fat brushed or sprayed on the connectors after welding and then briefly melt it with a torch.
Again, my main point was that any shorts in the rest of the circuit would decrease battery life, possibly quite significantly. Of course the batteries, if they’re closed units, would survive so long as no seawater got into their casings.
If this is another feature of the ipad format it should be over some time monday.
Following this post is doable, the open post takes 4000+ swipes per hour.
(ipad has no scroll besides swipe)
It’s a WordPress thing, totally out of our hands.
The low voltage power part of the system is hardly influenced by sea water. That’s why its possible to weld under water. It would stricktly speaking be possible to imagine a construct that would drain the batteries, but nothing sudden and its still a bit of a 6er in the Lotto. Assuming the battery capacity was not significantly influenced by the sea water, either direct or indirect, is the statisticly more likely variant.
I’d like to suggest we need to update and correct this statement as well. I think we have increasing evidence that this was a design based accident (and not a natural one): poor planning over an extended period of time for tsunami and earthquake site characteristics, poor primary containment design (we appear to have several breaches, spilled water, and pipe breaks throughout the site), inadequate containment for spent fuels (with water as the only primary containment barrier), inadequate evacuation and accident response plans (WSJ suggests they delayed cooling efforts during initial 24 hours with seawater at crucial initial phases of response), and more. We can definitely do better, and we no longer have to repeat the mistakes of 40 years ago to make nuclear power continue to be safe and reliable in the present.
Appearing in the Times yesterday, the verdict appears to be in on the question of tsunami planning and preparedness. Government guidelines overlooked the issue in 2006 (despite nonbonding recommendations in 2002, and additional warnings two years ago). Challenged with public and regulatory scrutiny after 1995 Kobe and 2007 Kashiwazaki earthquakes, still few changes (and nothing on tsunamis). A nuclear engineer working for Hitachi for 40 years described the safety and risk assessment culture this way: “Japanese safety rules are deterministic because probabilistic methods are too difficult … the US has a lot more risk assessment methods.” Tsunami and earthquake experts in hindsight described the many oversights this way: “a cascade of stupid errors that led to the disaster” (and more of the same from these experts over here). And this account from a Toshiba engineer on the early planning for the Fukushima plants (and safety culture in Japan): “We didn’t take a tsunami into account.” And a 10 year plant extension (received one month before accident) failed to take seriously these and other many concerns. I think I risk stating the obvious, but the more we learn about this design based accident, the more we learn about the many ways it could have been prevented.
[…] Guest Post on Bravenewclimate.com , an Australian website, attempts to put the nuclear event into some […]
I think the close proximity of reactors does hamper operations too… perhaps a wider distribution of reactor buildings must be enforced in the future. The radiation of neighboring buildings *should* be of lesser concern to workers when trying to cool some unit manually (or working on it otherwise).
Units 5 and 6 are not right beside Units 1-4. That has removed radiation from Units 1-4 as a serious concern. There has been speculation that the hydrogen explosion at Unit 4 came from hydrogen generated at Unit 3, through the common vent stack or some other route. Not likely in my estimation, but if it proves true that would be another indication that reactors should not be built too close together.
There is also the consideration ot siting of multiple units on one ste, with the possibility of a common source of multiply failure, as in this case. Now it has been found that multiply reactors on one site helps to quickly disperse information amoung operators of leasons learned at the other reactors. There is a study of CANDU reactors in Ontario that shows that. I read about it many years ago so do not have a link. So that is a plus of having many reactors on one site.
Factual errors in the article:
“After the normal feed water supplies were exhausted, they activated the system to supply sea water to the core, knowing this would render the plant unfit to return to operation.”
Wrong. There is/was no system in place to supply seawater. This was improvised using mobile pumps and fire hoses.
“Since this is a Boiling Water Reactor (BWR), it is good at boiling water. Operating with the water level 1.7 to 2 meters below the top of the core, they mimicked power operation; the core normally operates at power with the water level well below the top of the core, the top part being cooled by steam.”
Wrong. Even in a BWR the entire core must be below the water surface at all times.
I kind of lose interest in the article when it gets simple facts like these wrong.
You are breaking the BNC moderating rules here(please check these on the About page) by not supplying references to support your contentions. Further comments without these links will be deleted and you will be asked to re-post.
[…] Guest Post on Bravenewclimate.com , an Australian website, attempts to put the nuclear event into some […]
Very belatedly, I easily found a source for the last correction for which the Moderator requested evidence (“in a BWR the entire core must be below the water surface”):
Click to access 03.pdf
“The worst case loss of coolant accident, with respect to core cooling, is a recirculation line break …. In this event, reactor water level decreases rapidly, uncovering the core. However, several emergency core cooling systems automatically provide makeup water to the nuclear core within the shroud, providing core cooling.”
Click to access BWR-basics_Fukushima.pdf
” water inventory lost needs to be made up again in order to keep the fuel covered with water. Without electrical power to provide a source of water to the reactor vessel, the level in the reactor vessel decreased and eventually the fuel became uncovered. Once the fuel is uncovered, fuel damage can occur as a result of high temperatures and reactions between the cladding (zirconium based metal tubes surrounding the fuel pellets) and the steam being generated by the decay heat. This cladding-steam reaction also generates large amounts of hydrogen.”
Note, I only check in here every year or so, and only to check sources and factual claims if I find them left hanging. Many are.
I wish there were a trusted librarian/reference desk on call here to address questions like this one. The Moderator (at least back in 2011) could have found this information, as could the original poster — neither one did.
We need to try harder to educate one another about this stuff, not try to make it the other person’s job to check the facts.
I’m convinced we need fission plants. I’m also convinced they can’t be operated as they have been without doing more damage than we’d like to accept. I am inclined to think this is a cynical calculation that more damage will occur because that’s how the world works — nothing would ever get built if people knew in advance what it will actually cost.
Not cynical enough yet, but working on it.
There are no paid staff on this blog. The Mod is a volunteer. It is incumbent on the poster to do the research and supply the links/refs not the Mod or any commenter reading the post.
By the way — was Dr. Hannum able to provide a source for what he wrote about core cooling by steam with water level below the top of the fuel? It seems only fair for the Moderator to require sources from all those stating factual claims. I have been looking as hard for support for what Dr. Hannum wrote as for the other claims — and I haven’t found a source supporting what he wrote. I assume one exists somewhere and would appreciate seeing it.
Here’s a timeline for the course of events (more recent than this thread):
“… the water level dropped to the top of the fuel about three hours after the scram (6 pm) and the bottom of the fuel 1.5 hours later (7.30 pm). The temperature of the exposed fuel rose to some 2800°C so that the central part started to melt after a few hours and by 16 hours after the scram (7 am Saturday) most of it had fallen into the water at the bottom of the RPV…. (Oxidation of the zirconium cladding at high temperatures in the presence of steam produces hydrogen exothermically, with this exacerbating the fuel decay heat problem.)
And from the same world-nuclear page
Event sequence following earthquake (timing from it: 14:46, 11 March)
Unit 1 Unit 2 Unit 3
Loss of AC power + 51 min + 54 min + 52 min
Loss of cooling + 1 hour + 70 hours + 36 hours
Water level down
to top of fuel* + 3 hours + 74 hours + 42 hours
Core damage starts* + 4 hours + 77 hours + 44 hours
So — 1, or 3, or 2 hours elapsed between the time the top of the fuel was exposed, and the time core damage started.