In my last post (about NRC Chairman Gregory Jaczko’s May 20th announcement about finding additional issues with the AP1000 ), I linked to Cassandra Sweet’s article in the WSJ, in which she wrote that “The NRC plans an inspection next week during which the agency will look at Westinghouse’s quality assurance and improvements.” I closed that post with the comment that: “Based on news stories citing NRC staffer reports, though, it appears that fresh attention is being given to the effect of temperature differentials on the concrete shield building, and to the application of the quality assurance procedures used in producing the design. I’ve noted these as concerns I had in my earlier post on the AP1000. If these reports are correct, that’s a very welcome sign.”
There has been a lot of commentary on (NRC Chairman) Jaczko’s announcement, and commentary on the commentary. Jaczko wrote an op-ed in the Huffington Post defending the NRC, and Westinghouse sent out press releases and statements expressing things from its own perspective.
One thing Jaczko said in his May 20th statement was that “NRC staff will examine Westinghouse’s quality assurance and corrective actions programs as part of an inspection next week.” (emphasis added) According to the reports available on the NRC’s website, it appears that there were some corrective actions that had to be taken as a result of the last inspection in 2008: http://www.nrc.gov/reactors/new-reactors/oversight/quality-assurance/qual-assure-license/inspect-reports/2008/
You might wonder: Inspection? Of what? Nothing’s built. The design has been submitted. Aren’t the procedures for verifying the supporting design calculations kind of ho-hum? Not necessarily. Sometimes there are twists to an engineering design problem, and sometimes there are twists to a licensing review process. The reason it could be a big deal — and an interesting story — here is that it’s a case of two such intersecting twists.
Here is one of the twists: The AP1000 was designed to keep the same “footprint” as the AP600, but to be able to produce almost twice as much power. That’s a very difficult design problem. You don’t just have to verify that the change you made is done correctly; you have to verify that all the things you wish to keep unchanged really can be left unchanged. You have to document that somehow. You have to document how they are affected by the change, too, even if you leave them as they are. That’s part of the engineering record of the new design.
Even small changes require that you think things through systematically. You usually can make a reasonable attempt at it, given enough resources. Still, even “simple” changes such as operating an existing nuclear plant at a power level increase of a few percent can require an extensive effort and cost millions to analyze. Even then, there can be unanalyzed consequences. For very large changes, it’s not that easy.
And the fact that this new plant design was concocted under the constraint that it be as much like the design for a smaller one as possible is not the only twist involved: the new “streamlined” licensing process for new nuclear power plants begat another. Many of the details of the design weren’t required to be determined until the stage at which someone applied for a construction license (or a Combined Operating License (COL), which is a combined construction license and operating license.) Now that the NRC is reviewing COL applications for the AP1000, the process has taken lots of people by surprise. Implementation gets very complicated, very quickly.
That’s why, I think, the meeting that took place this week about quality assurance procedures for verification of engineering design could turn out to be an important one — if it’s done right. The report won’t be out until next week. We’ll see.
As I’ll be completely tied up this weekend and part of next week, I don’t expect to post again soon. Due to these other commitments, I don’t have time to explain things further here as fully as I’d like. Insteaad, I decided to post pdfs of the letters I wrote to the NRC many years ago about the QA procedures governing engineering design as well as the one about the Heat of Solar Radiation and Ultimate Heat Sinks. Here they are, using the exact images from the NRC’s own files:
1. Letter “Ap1000 Fluid Systems Design & QA Procedures” July 30, 2003
2. Letter “Heat of Solar Radiation and AP1000 Ultimate Heat Sink” July 31, 2003
Hopefully, these two letters will explain the conceptual questions behind some of the issues now being looked at more closely.
A longer and more detailed account of the context in which these were presented is given in the earlier post on “AP1000 New Nuclear Power Plant Design – making the NRC accountable in doing its job . . . ” There are more documents linked to in that post, but I think the pdfs I’ve provided above are the most concise statements of the basic ideas involved.