Go to CNF homepage
The Canadian Nuclear FAQ  

by Dr. Jeremy Whitlock


Published in the Ottawa Citizen 2008 February 8:

Never a real nuclear danger

John G. Waddington
The Ottawa Citizen
Friday, February 8, 2008

This week the much-discussed safety upgrade to the National Research Universal Reactor at Chalk River was completed, significantly decreasing the risk of harm in the case of an accident. But just how unsafe was it?

On Jan. 29, Linda Keen, former president of the Canadian Nuclear Safety Commission, testified before a House of Commons Committee that the NRU reactor was 1,000 times more risky than international standards would allow, if the two coolant pumps that provide cooling in emergency situations were not connected to a new, seismically qualified power supply.

I believe Ms. Keen's assessment of the risks from operating the NRU reactor is seriously in error.

She had compared NRU to a new international standard that requires that the next generation of reactors should be designed such that the likelihood of a serious accident that could lead to serious harm to a member of the public should be less than one in a million. She then assessed that the likelihood of "fuel failures" in the NRU was one in 1,000. The difference between these two probabilities led to her statement of that the risk was "1,000 times greater than the international standard."

But at least two more independent "lines of defence" built into a plant must fail before radioactive material released from fuel that has overheated -- for whatever reason -- could be released to the environment and become a hazard to the public.

"Fuel failure" is not synonymous with public harm. The accident that occurred at the Three Mile Island reactor illustrates this point very well. The fuel in that reactor overheated and a large proportion of the fission products from the fuel was released into the containment building. No member of the public or any plant worker suffered any health effects as a result of that accident.

Ms. Keen's statement can only be based on an assumption that if an earthquake with a return frequency of one in 1,000 years occurred essentially directly under the plant (about a magnitude-six), all the plant's back-up power supplies would fail so that the cooling pumps would stop, the steel pipes that circulate the cooling water around the reactor would fail, the building in which the reactor sits would fail, the operations staff would do nothing in response and the radioactive material normally locked up in the reactor fuel would escape into the atmosphere.

Put simply, Ms. Keen seems to have assumed that, before the installation of the new seismically qualified power supply, the NRU plant had no defence at all against a serious earthquake, and therefore the probability of harming the public had the same probability as the occurrence of the serious earthquake itself.

In fact, the response of well engineered industrial facilities to severe earthquakes has been studied for 35 years. Based on these studies there would be a high likelihood that at least one of the original back-up power supplies would continue to provide power to the coolant pumps. There is also much evidence to show that the steel pipes of the cooling system itself are unlikely to fail following the earthquake, particularly those operating at low temperature and pressure such as those in NRU. From this, we can conclude that the likelihood of the fuel overheating and radioactive material leaking out into the environment is significantly less than the likelihood of the earthquake itself.

And what would happen if, despite all these studies, the cooling systems all failed, along with everything else? AECL is required to analyze this event as part of the safety case that it must submit to the CNSC before it can get a licence to operate. The analysis predicts that, in the case where every back-up cooling system fails, the steel pipes fail, the confinement building fails, and the NRU operators do nothing at all to remediate any failures, the maximum dose that might be received by the most exposed member of the public would be 2.4 millisieverts (mSv). The CNSC has said it must be less that 5 mSv to be licensable. It should be noted that, based on 60 years of study, no health effects from radiation, including cancer, have been observed in humans who have received doses of radiation of below about 100 mSv.

In plain language, no health effects to the public are expected should this serious, but very unlikely succession of events actually occur.

In summary, the likelihood of the plant systems failing to such an extent that radioactive material would be released to the environment are much lower than Ms. Keen has assumed, and consequences of such an event should it actually occur would be negligible. The combination of a low probability of occurrence and negligible consequences means that the NRU reactor did indeed meet existing international standards.

The seven upgrades to the NRU reactor that the CNSC sought from AECL and which were installed in 2005, together with the connection of the two pumps to the new back-up power system, are designed to ensure the risks of operating the reactor will be close to the new international standard for new reactors being built today.

John Waddington was a director-general of the CNSC for 10 years, retiring from that organization in 2002. He is a professional engineer with 40 years' experience in nuclear safety.