Seismic Stress Test
I was involved when we did this for Pickering A. Like the first Japanese BWR's, this was not designed in the slightest for earthquakes. They did some lateral loads according to the building code, but that was for wind and snow load. So, we were faced with throwing up some sort of seismic qualification, so that it could be rejuvenated. We chose a dying system called Seismic Margin Assessment. It was once a big thing down in the States, but all the money left it. The formal system cost a fortune, but we got around it by hiring all the consultants that had written it. :)
It really needed extensive modification for Canada and Candu stations. It is basically a process of assuming a test earthquake scenario, and then running through the whole plant looking for weaknesses. Looking at everything was impossible, so the list was narrowed down by choosing a couple of paths (operations) that would allow safe shutdown and cooling. All the systems on those paths would be looked at. This walkdown (examining the systems) would have picked out the Japanese generator problem in a flash!
So the first job was to pick a decent earthquake that would stress test the system. It would be no good to pick the 'Earthquake From Hell', that would leave everything in dust, mainly because you would be able to improve anything. Besides that earthquake is up there with a large meteor strike. We chose the 1 in 10,000 earthquake, as derived from pure seismological studies. Unfortunately, the US could only think in terms of peak ground acceleration and this was a 30% g earthquake, which usually isn't enough to dust off the shelves.
-more later, I have to bake cookies.