Thank you, Mr. Chair.
Good afternoon, ladies and gentlemen. Thank you for asking me to reappear before you.
My name is John Waddington. I'm a professional engineer. I have spent over 40 years in the business of nuclear safety. I retired from the Canadian Nuclear Safety Commission in 2002 after 27 years with that organization. Since then, I have provided consulting services in matters of nuclear safety to various Canadian organizations, including AECL and the IAEA. I am appearing before you as a private individual.
Following a brief discussion with your chair and your staff, I thought it might be useful to give you my views on the question of whether the MAPLE reactors could be restarted to make radioisotopes. I have structured this presentation in the form of questions you might ask and my answers, as it seemed to be the easiest way to do this.
The first question is whether the MAPLE reactors could be restarted. The answer, in principle, is yes. But there would be many steps, and much human and financial effort would be needed and would need to be completed before such a direction could be successful.
Perhaps I could just refresh your memory a little as to why the MAPLEs were shut down. The first MAPLE, as was previously mentioned, was operating at power, though it was not producing radioisotopes at that time. It was in commissioning, when it was observed that its operating behaviour was not exactly as expected, when the reactor power was changed from one level to another. Although the difference between its observed behaviour and its expected behaviour was not large, it was significant, in that the reactor core was slightly more reactive at higher powers than at lower powers instead of being less reactive, as the original design intended it to be. The major difference, though, was not so much the change in the power coefficient of reactivity. It was the fact that there was a difference, a discrepancy, if you wish, between prediction and observation. In nuclear safety, the designer, the operator, and the regulator all expect a very high degree of correlation between analytical predictions of how a reactor operates and the measured observations of that operation. In other words, in the Nuclear Safety Commission, we expect the design and the operation to deliver what they said they were going to deliver.
After the expenditure of a very substantial effort by AECL, using very competent engineers and scientists, they could not find the reason for that discrepancy. They found some reasons for it, but they were not able to fully explain it, and the further steps needed to resolve the discrepancy were not readily apparent. Before the MAPLE reactors could be restarted, that discrepancy would still need to be explained to the satisfaction of AECL, as competent operators who are very conscious of the safety of the reactors, and of course to the satisfaction of the CNSC. And of course a licence would have to be applied for and granted by the CNSC. In my view, it would take likely a lot more time and effort to nail that discrepancy to the ground.
The second question you might ask is whether it is possible that the reason for the discrepancy could ever be found. My answer is yes. Given enough resources--there are some very smart engineers and scientists around--I would expect that AECL, together with support from the international community, would find that reason in time. However, the question remains: how much time and effort would be needed, and is it possible right now to estimate how much effort would be needed? The answer to that second question is no, it's not possible. A formal series of tests was put together using a very formal process of investigation, and it didn't come up with the answers. So we're starting from a head-scratching spot again, if I may put it that way.
Another question you might ask me is whether MAPLE could be operated safely with a small positive coefficient of reactivity. I would disagree with Michel here, and I would say that yes, it could.
Every reactor in the world has a positive coefficient for some accidents. In a PWR, it's a cold-water accident that puts a positive reactor in the core. In a CANDU reactor--or a power reactor, as Michel has pointed out--it's a loss of coolant accident; that is, the big pipe that would break. But every reactor in the world has some part of its design that, if put it into an accident condition, will result in a power increase. And the shutdown systems are designed to deal with that. In a CANDU reactor, there are two independent shutdown systems to deal with that.
Just as an aside, if Chernobyl had been fitted with a Canadian reactor shutdown system, there likely would not have been an accident at Chernobyl. That's just an aside.
So could it be operated safely with a small positive coefficient? Yes, but AECL would have to explain satisfactorily to itself and the CNSC what that discrepancy arises from before making the new safety case to the CNSC. Operation of a reactor with a small positive coefficient of reactivity would likely involve the operator moving the control rods that control the reactor more slowly than otherwise would have been the case. Obviously, if the power becomes more reactive as you increase power, you can control things a little more slowly to compensate. And I would expect that a successful prediction of observed reactor behaviour would be needed as an essential element in that safety analysis.
The question has arisen about an independent panel of experts. Could an independent panel of experts find the source of the problem? It is always possible that fresh eyes will see something that others have missed. I think that's evident to us all. Given the amount of effort that has been put in by AECL to date, and indeed by highly competent independent U.S. labs, a panel is also likely to take substantial time and effort, in my view. It would need additional experimental data to assist it in coming to a firm conclusion. So it's not going to be a simple thing. It's not a question of looking at the thing and saying “Eureka!” I can't see that happening.
The HANARO reactor in Korea generally works fine. Can't we learn from that? As I understand it, there was a significant amount of information transfer from HANARO to AECL while AECL was looking for the cause of the discrepancy. Well, the HANARO reactor works fine, so why doesn't MAPLE?
We don't know all the reasons for that. HANARO is a larger reactor than MAPLE, and of course it doesn't use the special high enrichment fuel that is used to make radioisotopes in MAPLE. I would expect, then, that since it's both larger and doesn't have this concentrated uranium in parts of the core, the neutron flux in the core would be a good deal less peaky. It's a good deal smoother in HANARO. In the MAPLE reactor, it's a very peaky process, which makes prediction of the reactor physics that much more challenging. And it does introduce different thermal stresses in the fuel of MAPLE that are likely not present in HANARO. However, whether these differences I've just outlined account for the difference in behaviour between MAPLE and HANARO is not clear. They are possible sources, but it's not clear.
Another question you might ask me is can HANARO be used to make medical radioisotopes? I can't see a fundamental reason why not. No doubt the Korean regulators would need a safety case to be made to them, and the Koreans presumably would need to develop and build a new processing plant to extract the moly-99. And I have no idea if Korea is looking at this possibility.
You may ask me if we have the right balance between nuclear safety and patient safety. You'd be asking me a question that I'm not really very competent to answer. I'm an engineer; I'm not in the medical business. And of course it is the most difficult question of all. The nuclear industry's designers, operators, and regulators seek a very high level of assurance of safety and a much lower level of risk to individuals and society from the operation of reactors than what is accepted in many other human activities. The medical industry is faced with a much more immediate life and death decision. So we have a clash here of quite different views of risk.
At the end of the day, Parliament may have to decide for itself what levels of risk are acceptable in a modern society, as you indeed did when you decided to instruct the NRU reactor to reopen 18 months or so ago.
The last question that I'd put into your mouths, if I may, is that some have said that the NRU will never start again. Will it? Personally, I think--in fact, I'm sure--it will. Corrosion is not an unknown phenomenon. The difficulties are finding the full extent of the corrosion, and that, of course, is a matter of inspection in a very difficult area, and then developing an appropriate method of repair, again in a physically very difficult area.
In addition to the need to produce radioisotopes, the long-term success of AECL's new power reactor design, the ACR-1000, is also dependent on timely experimental work planned to be done in the NRU. It seems to me that AECL's long-term future as a reactor designer of note—again, I would disagree with you, sir, on that, but we can discuss it later—the assurance of the long-term safety and economics of Canada's existing fleet of power reactors, and the continued development of a nuclear business in Canada depend on having a research reactor available for the supporting research that is needed to keep a complex machine running for 60 years.
At the moment, that means bringing the NRU back into service as soon as possible. In the long term, it means a replacement for the NRU will eventually be needed.
Thank you, Mr. Chairman.