Nuclear Power is Demonstrably Not Safe
By George Harvey
Proponents of nuclear power say it is the safest power source we have. Anti-nuclear people say it is the most dangerous. Proponents accuse those opposing nuclear power of being motivated by fear, rather than real science. Anti-nuclear people accuse proponents of being motivated by greed, rather than real science. For an impartial person listening to both sides,it is not easy to decide who is right.
Even when examining specific, narrowly-defined issues, it is difficult to find figures that can be accepted without controversy. Some proponents of nuclear power say that only 63 people died as a result of the Chernobyl Disaster. The Chernobyl Forum, composed of representatives of a set of UN agencies, came up with figure of 4000. Greenpeace seems to believe the number of people who will die early because of the disaster will exceed 100,000. Some governments put the number higher.
As it happens, there are numbers that are harder to dispute, and the truth may be discernible in these. Early in its history, the nuclear industry produced projections of the likelihood of core damage events (meltdowns), and today it continues to use both the same data and the same computational methods that created them. The core damage frequency (CDF), tells us the likelihood that a given reactor will melt down in any given year. Now, however, we have fifty years of experience. We can compare the old projections with empirical data representing what has actually happened in the real world, and see who is right and who is wrong.
A meltdown does not have to be a disaster. It can be a total meltdown, but it can also be relatively minor. Some of the great meltdowns are well known. A number of smaller meltdowns have happened, in which the core had a sufficiently small part melt that the reactor could be cleaned up and returned to service. All these are included for the calculations of the CDF.
When we review the list of meltdowns in civilian power reactors, this is what we find:
- In 1959, the Sodium Reactor Experiment, which despite its name was providing power to grid customers, melted down. The reactor was very near Los Angeles. After cleanup, the reactor was put back into service. At the time of the meltdown, the operators did not realize the high temperatures in the reactor constituted a serious problem.
- In 1966, Fermi 1, just outside Detroit, melted down. Though the reactor was cleaned up and restarted, it was never brought back to full power. The meltdown happened because operators did not notice a malfunction.
- In 1967, a minor meltdown occurred at Chapelcross, in Annan, Dumfreshire, Scotland. A single channel was blocked with graphite debris, and this caused some fuel to melt. The reactor was returned to service after a two-year cleanup.
- In 1969, Unit A1 of the Saint-Laurent in St. Laurent-Nouan, Loir-et-Cher, Centre, France had 50 kg of fuel melt.
- In 1977, the A1 reactor at Jaslovské Bohunice, in Czechoslovakia, had a fuel melt as a result of an accident during a refueling process. The accident happened because a worker had failed to remove packaging material before fuel rods were put into the reactor. The reactor suffered major damage and was permanently shut down.
- In 1979, Unit 2 of Three Mile Island experienced a partial meltdown in 1979. The meltdown was bad enough that the reactor was permanently disabled. The release of radiation is unknown, because it exceeded the limits of the monitoring equipment. The meltdown was the result of operator error.
- In 1980, the A2 reactor at Saint-Laurent in St. Laurent-Nouan, Loir-et-Cher, Centre, France had a partial meltdown. The reactor was cleaned up and returned to service.
- In 1986, the Chernobyl Disaster happened in the USSR. It is the only meltdown we know of from the Soviet Union. The reactor exploded causing many billions of dollars worth of damage, and killing many people. How many of either is a matter of controversy. The meltdown was the result of design flaws and operator error.
- In 1989, Greifswald Unit 5, in Lubmin, East Germany, had damage to ten fuel elements as a result of sticky electrical contacts causing multiple equipment failures. According to WISE-Stockholm, the underlying problem appears to have been sloppy construction.
- 2011 – Three reactors melted down at Fukushima Daiichi. Since fuel fragments were found over a mile from the plant, it is possible that one of the reactors exploded. The immediate cause of the meltdown was a tsunami that flowed over the plant’s seawall and entered unprotected areas where critical safety equipment had been installed. The meltdown was the result of bad engineering, however, including building a 5.7-meter seawall on the northeast coast of Japan, where the Tsunami of 1933 had a maximum wave height of 28.7 meters, and the Tsunami of 1896 had a maximum wave height of 38.2 meters.
These ten incidents involved twelve reactors. Other meltdowns may have happened in such places as the Soviet Union, but never made public for security reasons, so the number could be more than twelve.
Roughly 480 commercial nuclear reactors have been brought online, putting power onto power grids, for a total of about 16,000 reactor years (adding up the times in years all the reactors have actually run). Early reactors were claimed to have a CDF of 1-in-10,000. Later, as improved reactors were being designed, the CDF was claimed to have become 1-in 20,000. Even newer designs push the CDF to 1-in-50,000, but most that have been built are 1-in-20,000, with a few at 1-in-10,000.
Based on empirical data instead of calculated projections, however, twelve meltdowns in 16,000 reactor years means the actual CDF has been 1-in-1334. This shows the calculation of safety leading to values of 1-in-10,000 or 1-in-20,000 was very badly flawed.
We might note that twelve meltdowns out of 480 reactors means about 2.5% of all commercial reactors have melted down so far. We should also note that these reactors are, on average, about three-quarters of the way through their service lives, implying that an additional four meltdowns should be expected from them, if they retire as planned. And, since the older reactors are possibly more likely to develop problems, the number is very likely higher than four.
There is a reason why the nuclear industry’s projected safety calculations are so wrong. When probabilistic risk assessment is done, human error is very difficult to calculate. We should remember that human error is not just a problem of operations. It can result from design, manufacturing, construction, maintenance, or regulatory flaws. Here, all, or nearly all, of the failures resulted at least in large part from human error.
In the old days, the nuclear industry promised safety using projections of CDFs. Now, we are in a position now that the nuclear industry should scrap its outmoded projections of safety and start to live in the real world, using empirical data to represent real risks. Until they do that, we can only claim for certain that they are nowhere near as safe as they claim to be.
Two sources not linked in the text are: