Assessing the risk of nuclear technology

The nuclear industry is keen on coming up with probabilistic risk assessments that show that a serious accident at a nuclear power plant can only happen every 20,000 years or so (admittedly this number is from a somewhat outdated report from 1974). Apparently, this is supposed to keep people’s minds at ease. But it shouldn’t:

  • This should not have happened. At least not yet.

    This does not mean that once Chernobyl has happened, mankind will be safe for 20 millennia. It’s just statistical, so two accidents could happen in consecutive years (although according to these statistics, the next accident would then be – again statistically – 40,000 years away).

  • If you read carefully, you will notice that these statistics use “reactor years” instead of “years”. A reactor year is one year of a reactor’s activity, so if your country has 20 active nuclear reactors, the probability of a serious accident is already 1 in every 1,000 years.
  • It doesn’t take into account the slow deaths caused by constant radiation and increased rates of cancer in the areas surrounding a nuclear power plant.

But I have a much more general problem with these risk assessments: I just don’t believe that anyone can predict the safety for the next 20,000 years of a technology that has been around for 60 years. I think this is beyond the capability of any statistical model.

I have recently been thinking about our abilities to predict the risks of technologies because I had just come across a study by NASA’s Space Shuttle Safety and Mission Assurance Office that examined the risk of the first Space Shuttle missions and found that these were actually 10 times more dangerous than the current Space Shuttle flights.

Especially in light of the recent catastrophe in Fukushima where the possibility of a combined earthquake and tsunami had apparently not been considered (which in itself is a bit odd, considering that Japan is [a] an island that is [b] quite prone to earthquakes], this report by NASA provides valuable lessons for the risk assessment of nuclear power:

  1. The first Space Shuttle missions (in the 1970s) were ten times as dangerous as the current ones. We have many nuclear power plants currently active that are just as old. A scary thought.
  2. The main reason why later missions became less dangerous was that each mission learnt from the previous one. Nuclear power plants however are being used for decades, without any changes to the reactor’s design. Because a nuclear reactor is expensive to plan and build, but relatively easy to run, the nuclear industry has an incentive to squeeze as many years as possible out of each reactor.
  3. The biggest improvements in safety only followed what NASA calls “major events”, meaning an explosion with lethal consequences. This leads one to assume again that we are not capable of grasping all the risks that are technology-inherent, unless they are being pointed out to us by actual catastrophes.

And then there is of course still one fundamental difference: If a space shuttle explodes, the astronauts die. That is sad, but they had accepted that risk. – The radiation from a nuclear power plant however kills much more indiscriminately.

About Andreas Moser

You will most likely find me in the forest, next to the lake, reading a book. Just follow the cigar smoke!
This entry was posted in Economics, Environment, Politics, Statistics, Technology and tagged , , , , . Bookmark the permalink.

13 Responses to Assessing the risk of nuclear technology

  1. John Erickson says:

    I had read that is wasn’t the combination of earthquake and tsunami that did in the Fukushima Daishi plant, but the scale of each. 9.0 magnitude is a HUGE amount of energy, and the Richter scale is logarithmic – 8 is not twice as bad as a 4.0, it’s dozens or hundreds of times worse. (Sorry, plate tectonics class was a LONG time ago.) I had heard that the generators were protected by walls and raised intakes/outputs, but they only took into account a certain height (again, apologies for forgetting my sources), and the tsunami topped that by a couple of metres. And when you consider the amount of visible damage to the plant, the effect on the general population has been quite mild, especially compared to Chernobyl.
    The biggest problem is out-of-date safety regulations. Just like cars, airplanes, and public buildings. nuclear plants are built to a certain level of safety. As an example, the airbag and seatbelt on your car will save your life up to a certain speed (varies by country or political body), but if you drive into oncoming traffic at 150mph, you are dead, whether you drive a Mercedes SUV or a Fiat 500. And as always, regulations are “written in blood”, as you mentioned concerning shuttle flights.
    I’m not a huge proponent of current nuclear plants – even the newest ones here in the United States are well over 30 years old. New plants, with new safety regulations, could serve as a bridge to truly green technologies – but only temporarily. And I worry FAR more about the waste storage – look at the problems in Reactor No. 4 at Fukushima Daishi. Perhaps this incident will serve to push the development of far safer alternative sources such as wave and current power, as well as geothermal and others just coming online.

    • g says:

      The problem with the Fukushima Daishi plant that it wasnt a problem , The plant was put in that location to insure that if it did blow up that it wouldnt kill their country leaving a trail of death in its wake and billions in clean up cost. Instead it would be washed away and end up on other countries shores or lost at sea. Wonder if the whales can sue over it?

  2. I think you make a mistake in not differentiating between risk and uncertainty. You can only reduce risk based on what you know. It is quite easy to assess the risk of driving by relating the number of drivers, the number of accidents on the route you want to take and your driving experience to each other. So you end up with a probability of 1:100000 for you to have a fatal accident. This, however, is to assume that future equals past to a certain degree and it is to assume that remaining uncertainty will be distributed randomly, thereby reducing your personal risk. However, if, by any coincidence, an alien spaceship lands on the M4 exactly at that moment you happen to drive there, and as a result you got flattened, this materializing uncertainty is something no statistical model can account for, because it hasn’t happened before. However, past your flattening experience, models will include the possibility of spacecrafts landing on top of cars. This is what experience is all about and this is the problem with every risk assessment, you can’t include what you don’t know, you can only allow for a random error and Fukushima is well inside the confidence interval of statistical prediction given that a lot of unforeseens had to combine to bring about disaster, i.e., maintenance work, tsunamie, earth quake, total break down of power supply ..

  3. Mountainbear says:

    No offense, but if I want a risk assessment on nuclear power, I talk with an engineer, technician or physicist. You know, a professional, like the men at Fukushima and not someone commenting from thousands of miles away with no connection to the matter.

    “The radiation from a nuclear power plant however kills much more indiscriminately.”

    Statistical deaths don’t count. The actual deaths by nuclear power plants are pretty low.

    As John said, it was the scale of events. An 8.9 quake plus a tsunami, which, at some points, may have exceeded 20 meters. The system withstood the 8.9, while only be designed for an 8.2. An 8.9 is 7 times stronger than an 8.2 since it scales logarythmically. Then the tsunami knocked out any backups they had. A lot of bad luck. Such things happen.

    The death toll by quake and tsunami will be significantly higher than by Fukushima Daiichi. Right now, it’s still like this:

    Quake + tsunami: 10,000+ (20,000+ expected)
    Fukushima Daichi: 0

    “It doesn’t take into account the slow deaths caused by constant radiation and increased rates of cancer in the areas surrounding a nuclear power plant.”
    To that I won’t say anything. It’s the classic fear mongering by people who don’t understand radiation or nuclear power plants. To put it simply: there is zero evidence for this claim.

  4. Took me time to read all the comments, but I genuinely enjoyed the post. It proved to be Pretty helpful to me and I am sure to all the commenters here It is always great when you can not only be informed, but also entertained Im certain you had fun writing this write-up.

    • I usually delete spam like this, but I couldn’t help but notice the irony of an ad for penis enlargement after a blog posting that warns about our lack of ability to appropriately assess risks of new technologies.
      So: No thanks.

    • John Erickson says:

      It had to happen. Radiation released in a country that is crazy about robots – so this is the future. And I always thought mankind would be wiped out by Skynet after it nuked the world during Judgement Day. Instead, we get Spammed to death by netbots. I bet James Cameron never saw THAT one coming! :D

  5. Pingback: First New Year’s Resolution already broken in 2012 | The Happy Hermit – Andreas Moser's Blog

  6. Pingback: How people found my blog | The Happy Hermit – Andreas Moser's Blog

  7. Kyle Jameson says:

    “so two accidents could happen in consecutive years (although according to these statistics, the next accident would then be – again statistically – 40,000 years away).”

    No, it doesn’t. Statistically speaking.

  8. Pingback: Atomares Endlager? Das ist doch kein Problem! | Mosereien

  9. g says:

    I think we need to look at the risk of using oil,coal and gas which they tap from the earth! It fuels a fire that burns under surface of our earth. once the fire has no fuel it goes out and we become a dead planet Yes we get heat from the sun but the earth provides it own heat also and if that fire goes out we die with it! If we didn’t breed so much we wouldn’t need all this fuel so maybe a plant blowing up once in a while is good for us in some kinda way?

  10. Pingback: Don’t be stingy with the translator! | The Happy Hermit

What do you think?

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s