LEAVING NO STONE UNTURNED

"You can't squeeze blood out of a stone," says an ancient proverb. True enough; but how about a little radon or extra electrical conductivity?

When rock comes under stress, its electrical conductivity changes, in part because the conductivity of any material is depends on its density (which increases under compression), in part for reasons that are not yet fully understood. And rocks do come under stress before an earthquake; the earthquake itself sets in when the stressed layers can take it no more. It has therefore recently been tried, with some success, to monitor the conductivity of rock to provide warning of impending earthquakes.

As reported earlier [Sept. 79], it has also been found that the radon concentration in water rises sharply prior to an earthquake. Uranium, the source of the natural decay chain that leads to radon gas, is a ubiquitous element, and virtually all rocks contain it, though some, such as granite, contain more than others. The radon then bubbles up through the overburden and into ground water. It may be that the pre-seismic stress will squeeze out additional amounts of radon, if not blood. In any case, the radon concentration of ground water is now being monitored as an earthquake prediction device.

An earthquake is a threat to any large-scale energy facility, but not to the same extent. The power plants facing the highest risk, in both probability and consequences, are hydroelectric (or any other) dams; the lowest are nuclear plants. The latters' safety lies in their siting, their earthquake-resistant structures, their elastic piping, the hollowness of the pressure vessel holding the reactor core, and the thorough tests that all components have to undergo to conform to outlandishly strict specifications.

By iron-clad logic, it is therefore up to the nuclear industry to defend itself, and it goes to extraordinary lengths to do so. The US utility-financed Electric Power Research Institute of Palo Alto, Calif., together with the Taiwan Power Co., has built two model nuclear plants including containment buildings within the dense net of an NSF seismic measuring array near Lotung, Taiwan. The two models are ½ and ¼ size replicas of PWRs, which do not produce any power, but are equivalent to the real thing in their earthquake-resistant properties, both in containment buildings and the actual reactors, including piping.

The area is subject to frequent earthquakes, and recently experienced an earthquake measuring 6.0 on the Richter scale. Evaluation of the data supplied by the heavily instrumented mockups will, among other things, enable computer model predictions to be compared with the experimental evidence, and the programs to be amended as necessary.

There is, of course, also a lot of unplanned evidence by now. Several Japanese plants went through earthquakes without missing a beat, and the biggest US earthquake since San Francisco in 1906, measuring 7.3 on the Richter scale, hit in October 1983 only 60 miles from the Idaho National Engineering Lab, where four reactors were in operation at the time. They shut down instantly, but as it turned out, unnecessarily, for none suffered any structural or internal damage [AtE Jun 84].

On January 31 of last year, a Richter 5 earthquake hit 10 miles south of Cleveland Electric's recently completed Perry-1 reactor, which was awaiting a low-power start-up licence. The earthquake exceeded the ground acceleration for which the reactor must be shut down (0.15g, where g=32.2 ft/sec/sec is the acceleration of terrestrial gravity), though it is designed to withstand accelerations many times higher. The acceleration-sensing switches tripped the reactor as required. There was no structural damage; the remaining was slight and only to nonessential items.

[More: "Breaking new ground in seismic research," EPRI Journal, Sept. 1986; reprint available from AIF, 7101 Wisconsin Ave, Bethesda, MD 20814.]