Bureaucrats follow the book. The book requires compliance of all systems. The Year 2000 Problem threatens compliance. Conclusion: shut them down before 2000.
This essay on Westergaard's site points to the problem. The author, Rick Cowles, has a Web site devoted to electrical utilities and 2000.
Nuclear power supplies 20% of U.S. power. Take this off-line in late 1999, and what happens if railroad coal shipments start breaking down in 2000? Oil delivery?
Baby, it's cold outside.
* * * * * * *
A lot of automatic safety and support systems are necessary to operate the average nuclear facility. What if all the nuclear plants in the U.S. were required to shut down as a precautionary measure prior to 01/01/2000 because no one knew if these systems were going to function properly?
The production of nuclear energy from the 108 domestic U.S. nuclear facilities accounts for nearly 20 percent of the total domestic (US) electrical generation capacity. In some regions of the country, this percentage is much greater. For example, the eastern seaboard of the U.S. depends on nearly 40% of its electricity being generated by splitting atoms. . . .
NRC regulations that were developed as a result of TMI [Three Mile Island, 1979] require all plants to do an exhaustive analysis of potential safety problems. These analyses are conducted to make sure that the plant operators training and understanding of how the plant operates doesn't change because of some type of failure mechanism that was never considered before (and Y2K is one heck of a potential failure mechanism that was never considered before). The detail, structure, and depth of these analyses would boggle the mind of the average person outside of the industry.
Here's an analogy: let's say the manufacturer of your car recommends that you use 89-octane fuel in your car. You, however, wish to switch to 91-octane fuel for better engine performance. What if Environmental Protection Agency regulations said that you couldn't legally switch fuel grades until you had a mechanic analyze the impact of such a change on every system of your car? And provide assurance that using a higher-octane fuel would not fundamentally alter the actual performance (or your understanding of the performance) of any aspect of your car? And that you couldn't drive your car until you had written certification of your mechanic's analysis?
. . . .Think about the magnitude of the Year 2000 computer problem - it touches virtually every area, administrative and operationally, of a nuclear power plant. Hundreds of automatic controls and embedded chips exist throughout a typical plant. Many important (and date dependent) databases and maintenance scheduling computer programs are administratively required for a plant to continue operating. And literally everything has to be reviewed and intensively documented for Y2K impact on the ability to safely operate the plant.
Unless a nuclear plant can prove conclusively that the Y2K bug will not impact the safe operation of the plant or the ability to safely shut it down, the plant doesn't run. Conservative, safe operation of a nuclear power facility dictates this thought process. The NRC has publicly stated that a certification of Year 2000 readiness will be required from every nuclear plant in the country. In the absence of that guarantee, the NRC has no recourse, by federal regulations, but to require the shutdown of the plants that aren't Y2K certified. I would expect this requirement to be enforced no later than the 3rd or 4th quarter of 1999. . . .
As noted before, a significant percentage of all domestic U.S. electrical power is produced by nuclear energy. At any given time, there's about a 15% reserve in makeup power available (provided all other generating plants are available and running). If the nukes aren't operating, where's the extra power going to come from? You can do the math on this one.