Severe Nuclear Reactor Accidents Likely Every 10 to 20 Years - New Max Planck Institute Study

[bananas]

"In light of our findings I believe an internationally coordinated phasing out of nuclear energy should also be considered," adds the atmospheric chemist.

http://www.sciencedaily.com/releases/2012/05/120522134942.htm

Severe Nuclear Reactor Accidents Likely Every 10 to 20 Years, European Study Suggests

ScienceDaily (May 22, 2012) — Western Europe has the worldwide highest risk of radioactive contamination caused by major reactor accidents.

Catastrophic nuclear accidents such as the core meltdowns in Chernobyl and Fukushima are more likely to happen than previously assumed. Based on the operating hours of all civil nuclear reactors and the number of nuclear meltdowns that have occurred, scientists at the Max Planck Institute for Chemistry in Mainz have calculated that such events may occur once every 10 to 20 years (based on the current number of reactors) -- some 200 times more often than estimated in the past. The researchers also determined that, in the event of such a major accident, half of the radioactive caesium-137 would be spread over an area of more than 1,000 kilometres away from the nuclear reactor. Their results show that Western Europe is likely to be contaminated about once in 50 years by more than 40 kilobecquerel of caesium-137 per square meter. According to the International Atomic Energy Agency, an area is defined as being contaminated with radiation from this amount onwards. In view of their findings, the researchers call for an in-depth analysis and reassessment of the risks associated with nuclear power plants.

<snip>

The computer simulations revealed that, on average, only eight percent of the 137Cs particles are expected to deposit within an area of 50 kilometres around the nuclear accident site. Around 50 percent of the particles would be deposited outside a radius of 1,000 kilometres, and around 25 percent would spread even further than 2,000 kilometres. These results underscore that reactor accidents are likely to cause radioactive contamination well beyond national borders.

<snip>

If a single nuclear meltdown were to occur in Western Europe, around 28 million people on average would be affected by contamination of more than 40 kilobecquerels per square meter. This figure is even higher in southern Asia, due to the dense populations. A major nuclear accident there would affect around 34 million people, while in the eastern USA and in East Asia this would be 14 to 21 million people.

"Germany's exit from the nuclear energy program will reduce the national risk of radioactive contamination. However, an even stronger reduction would result if Germany's neighbours were to switch off their reactors," says Jos Lelieveld. "Not only do we need an in-depth and public analysis of the actual risks of nuclear accidents. In light of our findings I believe an internationally coordinated phasing out of nuclear energy should also be considered ," adds the atmospheric chemist.

[FBaggins]

1. What incredbible nonsense

Oh yes... all reactors are identical and all are equally prone to the same dangers. Who cares about design differences... it says "reactor" on the label, right? If a new design comes out that's 1/10th as likely to suffer a meltdown... it doesn't matter. It's a reactor.

So all we have to do is take the total number of reactors in operation times the number of years they've operated... then divide by the number of accidents. Presto! We know how often accidents will happen in the future!



Think of all the money we could save by getting rid of the NHTSA? Why rate cars on crashworthyness or set insurance rates by driving record? We know how many cars have existed and how many fatal accidents there are per mile driven... the chances of a fatal accident are therefore already known.

[caseymoz]

4. I think the scientists considered what you said.

I haven't read their analysis, but if they're scientists I suspect they might be way ahead of you on this. They're professionals, and I would read their analysis before judging that they must have made amateur mistakes that even Bill O'Reilly wouldn't make. They don't want to cut out an absolutely rich source of energy, unless you believe the conspiracy the Warming deniers would point to.

Also, they were talking about current reactors, not future designs, as I understand.

Now, I'd like to see new reactor designs too, but this gives me pause. Fission reactors are all going to have the same drawback of needing enriched isotopes of Plutonium or Uranium-235. They are all going to have the same problem that they are going to create waste that is both toxic and radioactive. The handling of these materials, before and after fission is necessarily going to be dangerous. Another draw back is they are going to be expensive, at every step-- to prevent accidents. Those toxic, radioactive isotopes will then be with us, forever. I'm thinking they were looking at possible accidents with those, and not just meltdown.

Lastly, if they shut down the current plants, that doesn't mean that experimentation with new designs won't go ahead. In fact, having the currently designed plants in service probably stands in the way of that.

[FBaggins]

8. Think again.

You're right that it seems absurd that their "analysis" (sic) would be so simplistic... But it's just what they did.

To determine the likelihood of a nuclear meltdown, the researchers applied a simple calculation. They divided the operating hours of all civilian nuclear reactors in the world, from the commissioning of the first up to the present, by the number of reactor meltdowns that have actually occurred.

Also, they were talking about current reactors, not future designs, as I understand.

And there are large differences between current reactors. Chernobyl didn't "pop" because "it's a reactor and that's how often they fail"... It failed because of specific design weaknesses that are not shared by more than a handful of reactors.

[caseymoz]

17. I went through the inherent design flaws

In any fission reactor. Those are going to be true no matter what the design is. Those are always going to carry with it serious safety issues, not to mention huge expenditures those issues entail.

That mathematical model is taking all of what you said into account. It's absolutely fair to look at all the reactors in operation and not just the new ones. If you threw out the newer ones, wouldn't that be grading on a curve? It's easy enough to separate the newer reactors from the older ones, do the division to find out if there is improvement.

It might give you confidence, but that method is also invalid due to shorter operation times. You can't extrapolate improvement in the long term from the short term, especially when long term retrospective safety was exactly what this study was meant to find out. The scientists know this, and I'm informing you of it.

Moreover, just because the article doesn't mention that the scientists studied new reactors doesn't mean they didn't, however flawed that analysis might be. Reporter can't put everything into the story. Maybe you should read the original.

I seem to remember operator stupidity was actually what caused Chernobyl, and the reactor wasn't designed with withstand the utterly stupid things done. In the last decade, it's become fashionable to say it was a design flaw. No, it wasn't. True the reactor wasn't designed to be run in an idiocracy. No reactor can be. Not when every reactor has to be constantly cooled to prevent catastrophe.

A principle selling point of the old reactors everywhere by the industry in all its ideologies and nationalities was their safety. Now we have new, improved, reactors whose selling point is their safety. Given the industry track-record, worldwide, why should those assurances be considered true? Are they born again honest and humble this time? Don't you see a credibility problem?

The current reactors are considered improved by correcting yesterday's problems, not any new, unforeseen ones. New designs can introduce new, unforeseeable accidents.

The only way we can know if they're significantly safer is to run them long term. That hasn't happened yet. To sum up: you have no reason but industry assurances and blind faith that the new ones are safer. And no reason to think these scientists approach was simplistic or biased just by reading this article.

[FBaggins]

18. I'm sorry, but you've just admitted to be true that which I claimed.

You really do think that all reactors are essentially the same in terms of how often you could expect one to fail.

That's ridiculous in the extreme.

Moreover, just because the article doesn't mention that the scientists studied new reactors doesn't mean they didn't,

Actually, it does... particularly when it explicitly says that they didn't.

The Mainz researchers did not distinguish ages and types of reactors, or whether they are located in regions of enhanced risks

I seem to remember operator stupidity was actually what caused Chernobyl

That's correct. It's also relevant to this conversation... because the same stupidity at a western reactor could not have had anything like the same result. To use your wording, there were "inherent design flaws" that do not exist "in any fission reactor"

It might give you confidence, but that method is also invalid due to shorter operation times.

You may be misunderstanding what I'm saying. I'm not talking about the newest designs being better than the older ones (though they are). I'm just refering to differences between existing reactors. As has been pointed out here many times, the simple fact that US reactors (of the same period and basic design) have their backup generators in waterproof vaults very likely would have meant that Fukushima would never been more than a week-long news story. The hardened vents that most have could have avoided the hydrogen explosions. You could go on and on before you even come to the fact that very few reactors sit low on a tsunami-prone coastline or where record-sized earthquakes occur.

[caseymoz]

27. And your interpretation of what I wrote is false

And also offensive, insulting, oh, yes and ridiculous in the extreme. I did not read beyond that, and won't.

If you get what I wrote that wrong, read into it "admissions" and play gotcha, it also shows you definitely can't get nuclear power right. I'll listen to the experts and not dilettante fanboy critics.

And I'm not sorry for pointing that out. Goodbye.

[OKIsItJustMe]

21. While other reactor designs fail for other reasons

The Three Mile Island meltdown was caused by a different design failure than Chernobyl, while the Fukushima meltdowns were caused by yet another design failure.

Or were they? I would say that all 3 incidents share a similar cause, the same cause as the loss of 2 US “Space Shuttles,” hubris, pure and simple.

http://www.nytimes.com/2012/03/04/world/asia/japans-premier-says-government-shares-blame-for-fukushima-disaster.html

[font face=Serif][font size=5]Japanese Prime Minister Says Government Shares Blame for Nuclear Disaster[/font]

By HIROKO TABUCHI
Published: March 3, 2012

[font size=3]TOKYO — Prime Minister Yoshihiko Noda of Japan acknowledged on Saturday that the government shared the blame for the disaster at the Fukushima Daiichi nuclear power plant, saying that officials had been blinded by a false belief in the country’s technological infallibility, even as he vowed to push for the idled reactors to be restarted.

…

“The government, operator and the academic world were all too steeped in a safety myth,” Mr. Noda said in an interview with journalists from overseas news media organizations. “Everybody must share the pain of responsibility.”

…[/font][/font]

[FBaggins]

22. You've raised a good point.

I have no problem accepting a claim that accidents will be more common than a designer estimates (or that hubris is a big reason for that)... but that's very different from trying to quantify the impact of that hubris or assume that the correct conclusion is that there is no effective difference between designs.

If, for instance, airbags save far fewer lives than their original designers estimates, that doesn't mean that they aren't a safety improvement, and it doesn't keep us from estimating how much safer some enhancement to the technology is.

The Three Mile Island meltdown was caused by a different design failure than Chernobyl

Actually, many of the root causes were similar (inadequate training for one)... but we can't ignore that the results were very different, and we can't pretend that the difference in results had nothing to do with the difference in design.

It's very reasonable to assume that had TMI been of the RBMK design, America would have had its own Chernobyl.

[OKIsItJustMe]

24. The root cause of inadequate training is…hubris

One of my pet peeves is people driving at normal speeds on a snow-covered highway because they have a big honkin' SUV with "four wheel drive." They have a false sense of security (and a false sense of their own competence) and as a result, they endanger other drivers.

I'm not saying that the introduction of Gen IV (or Gen III+) reactors would be a net loss of safety. On the other hand, I think the magnitude of improvement, may not be as great as it should be.

[intaglio]

6. The people at the Max Planck Institute are incredible idiots?

Look, there's this bridge in London that is just sitting in my portfolio. It's a bit of a fixer-upper, sometimes the roadway just separates but I'm sure that can be fixed - I can let you have it cheap.

Seriously, sometimes even you must choke on the Kool-Aid fed to you by the nuclear industry.

[FBaggins]

9. I don't think they're idiots...

... I think they hope that you are.

Try reading the piece (or my quote above)... That's precisely what they did.

It's interesting that you appear to recognize that such a ridiculous assumption would be incredibly idiotic.

[kristopher]

10. Would you mind giving a proper citation for the sections you are referring to?

For a change.

[FBaggins]

11. Try reading what you respond to.

I pointed to the quote from the prior post... and it's from the same story as the link in the OP.

[kristopher]

14. I had already read the paper before the OP was posted.

I was giving you a chance to do the right thing and go get the rest of the discussion that shows you took the quoted portion out of context in order to change the significance of it.

Now why don't you go do the right thing and post what you left out.

[FBaggins]

15. The you are without excuse.

There's nothing wrong with putting the text from the article referenced in an OP without once again linking to what had already been linked.

And it's nonsense to talk about the quote as being "out of context" there is nothing in the remainder of the piece that in any way refutes what I said. They guestimated future accident occurence by assuming that all reactor/years are the same. Period.

Now why don't you go do the right thing and post what you left out.


You mean the part that further confirms what I said?

The Mainz researchers did not distinguish ages and types of reactors, or whether they are located in regions of enhanced risks, for example by earthquakes.

[kristopher]

19. Here is what the authors say

By integrating over a year we capture the annual range ofmeteorological conditions, thus providing a stochastic representation of the atmospheric transport and deposition pathways, accounting for the different seasons. During an actual accident the total deposition likely occurs over a much shorter time period, as was the case with the Chernobyl reactor. To assess the effect of individual accidents onewould need to simulate the actual emissions and meteorological conditions in a deterministic approach (http://flexrisk.boku.ac.at/). Such calculations can be performed when theemissions and meteorological conditions are known, which has been done for Chernobyl and needs to be done for Fukushima. For our risk calculations the total deposition overall meteorological conditions is relevant, rather than the actual time period of any individual accident. For this reason we assess the contamination risk per year in Sect. 4. To illustrate the intra-annual variability, we also present monthly contamination risk maps in Sect. 4 by assuming that the same emission occurs within one month rather than one year. Fur-thermore, in Sect. 6 we test our approach by comparing thecontinuous annual simulations with those in which we re-lease the radioactivity in weekly periods for two selected locations.
J. Lelieveld et al.: Global risk of radioactive fallout p.4247

[FBaggins]

20. You've confused unrelated portions of the piece.

In other words, you've invented a context that doesn't exist in report.

Intentionally?

They performed real science regarding the atmospheric chemistry involved after an accident is assumed. That is, after all, where they have actual expertise. And you'll find no post from me questioning the validity of those models.

But the estimates re: how often a reactor can be expected to "pop"... are not in any way "science".

You'll have to carry on alone from here... I'm off to a conference. Your apology for the error is assumed and accepted.

[intaglio]

12. So what you are saying is that MPI is being deliberately misleading?

That for some unidentified reason they wish the nuclear industry to fail; perhaps because they are wholly in thrall to the evil, anti-nuclear government of Germany.

Sorry that isn't how science works, but it is good fodder for conspiracy theorists.

BTW please link to your sources, not leave people to search amongst your verbiage

[FBaggins]

16. They could be... or it could be the reporting.

(Though it should be pointed out that this isn't the society putting their stamp on the claim, it's just some members of the society publishing a piece)... not at all the same thing as an official WHO study. You could, for instance, see other society members publishing contrary pieces.

They really didn't put any effort into identifying how often a reactor meltdown could be expected. Their real focus was on simulating the release/plume and the impact it would have on Europe. They were atmospheric chemists after all, that's their field.

I can't say from the summary whether they are the ones pushing the lie that they've somehow calculated how likely reactor accidents actually are... or whether that's just what the reporting picked up.

Either way it's faux science.

[intaglio]

23. As I said, conspiracy theory

Last edited Thu May 24, 2012, 03:23 PM - Edit history (1)

First; "Though it should be pointed out that this isn't the society putting their stamp on the claim, it's just some members of the society publishing a piece" The same could be said about that Nature article on a draft report produced by some members of WHO.

You are aware of failure analysis in engineering? It is pretty damn important after all. Take the simplest example I was told. you have a marine diesel that needs to be cooled. Essentially there are 4 choices open to engineer; air cooling, direct salt water, dual circuit salt/fresh and closed circuit fresh with a keel radiator. Air cooling is impractical for any but small engines because of the vast volumes of air needed for cooling. Keel cooling is complex to install and difficult to maintain. Single circuit direct salt water cooling is cheap, effective and very vulnerable to catastrophic failure by clogging and corrosion. Dual circuit with 2 pumps one for salt water and one for a closed freshwater system and heat transfered through an exchanger to the salt, it is very popular, cheap and standard maintenance is easy but the risk of pump failure is 4 times higher than for a single pump system. This is diagrammed as pump A fails, or pump B fails, or pump A non-critical fault causes B to fail or B non-crit fault causes A to fail (there is also a tiny chance of A and B failing independently). This is drastically over simplified but essentially the worst (dual pump) system is used on many pleasure craft with inboard engines because it is cheaper and failure will not often result in loss of life. The far more failure resistant Keel cooling system with it's problems of ballast and maintenance is rarely used because of it's expense.

Now multiply the problem by a nuclear power station, where failure will result in loss of life and land.

People like the MPI make it their business to analyse failure modes of critical systems such as nuclear installations. They use complex computer models, factor in increased risk due to age, wear and poor maintenance, examine known sub-critical failures and failures of similar systems. This paper is item is the result of such an analysis.

Now you may want to slag off one of the great German Institutes and a recognised discipline such as failure analysis, just do not expect me to respect you for so doing.

[One_Life_To_Give]

25. Common factor is

the Human factor.

Chernobyl, TMI and Fukushima weren't the result of hardware failures. Not anymore than a traffic accident being the fault of insufficient brakes. The first two the automatic controls were disabled and the last a report of inadequacy was ignored.

From where I sit the only one to build a culture of safety equal to the risk was named Rickover.

[FBaggins]

26. Well you certainly won't catch me bad-mouthing Rickover

But saying it's the human factor just goes too far.

Both Mercedes and the DIY scooter will have human-factor errors... But one IS clearly safer than the other.

You'll also have a tough time convincing me that Fukushima was primarily a human-factor failing at all like the other two.

[bananas]

28. It's not nonsense, it was even used in MIT's 2003 report "The Future of Nuclear Power"

From page 48 of the full report at http://web.mit.edu/nuclearpower/

Also note they said more than 1 core damage before 2050 is unaccceptable,
there were 3 cores destroyed just last year - we've already passed what they considered acceptable.
And there will be more.

What is the expected frequency of accidents
today with the currently operating nuclear
plants? There are two ways to determine the frequency
of accidents: historical experience and
Probabilistic Risk Assessment.7 Since the beginning
of commercial nuclear power in 1957,
more than 100 LWR plants have been built and
operated in the U.S., with a total experience of
2679 reactor-years through 2002. During this
time, there has been one reactor core damage
accident at Three Mile Island Unit 2. The core
damage frequency of U.S. reactors is therefore 1
in 2679 reactor-years on average.

Probabilistic Risk Assessment (PRA) identifies
possible failures that can occur in the reactor,
e.g., pipe breaks or loss-of-reactor coolant flow,
then traces the sequences of events that follow,
and finally determines the likelihood of their
leading to core damage. PRA includes both
internal events and external events, i.e., natural
disasters. Expert opinion using PRA considers
the best estimate of core damage frequency to
be about 1 in 10,000 reactor-years for nuclear
plants in the United States. Although safety
technology has improved greatly with experience,
remaining uncertainties in PRA methods
and data bases make it prudent to keep actual
historical risk experience in mind when making
judgments about safety.

With regard to implementation of the global
growth scenario during the period 2005-2055,
both the historical and the PRA data show an
unacceptable accident frequency. The expected
number of core damage accidents during the
scenario with current technology8 would be 4.
We believe that the number of accidents expected
during this period should be 1 or less, which
would be comparable with the safety of the current
world LWR fleet. A larger number poses
potential significant public health risks and, as
already noted, would destroy public confidence.

[JayhawkSD]

2. If that's the case, why aren't they happening?

Some sixty years, two meltdowns, three if you count Three Mile Island which was only partial and about which the external release of radiation is questionable. That's a long way from "one every 10 to 20 years."

[kristopher]

3. Here is the full list

[bananas]

5. They are happening. In fact, there are three happening right now.

And that's not counting the spent fuel pools.

[villager]

7. Exactly. And none of the world's plants are getting younger. The apologetics for the nuclear gang,

though, is getting pretty long in the tooth.

[caraher]

13. I think this is the key to their analysis

The MPI analysis really doesn't seem concerned with all the INES 4-7 events listed in kristopher's post, but clearly does count Chernobyl and each individual Fukushima reactor meltdown.

It's not the last word on the risk of major releases, but it is based on the real-world track record versus models of failure, and the number of reactor-years we're logging by extending licenses suggests a dramatic decrease in the rate of severe accidents anytime soon is unlikely. One can quibble about whether to count Fukushima as 1 accident or 3 or 4, but the order of magnitude estimate holds whether you count Chernobyl + Fukushima as 2 or 5.