Computer Problems at Nuclear Plants When the Millennium Changes

When the next millennium arrives, will the nuclear power plants shut down?... or maybe "blow up" ? What will happen: explosion, sudden shutdown, or will everything be normal?

For the past several months, even years, the computer industry has been discussing this theme, sometimes heatedly, at other times apparently unconcerned: how will the computer programs, operating systems, microprocessors and clocks react, if the jump is into the next millennium? What will happen if, from one second to the next on 31.12.1999, the date 1.1.2000 follows? Will the computers and their software swallow that without complaint, and continue operating in accordance with their programs? Or will they simply "crash" and quit?...and if so, what will be the consequences?

Many programs developed in the 60's. 70's, and 80's didn't take into consideration that the century would come to an end. Their main concern back then was storage....each bit and byte had to be turned three times, because even the large "mainframe" computers were equipped with only one, two, or sometimes eight megabytes of memory. So the programmers simply abstained from mentioning the century, in order to save storage space. But with this logic, somebody who was born 31.10.99 must have first seen the light of the world in the year 1899.

On 31.12.1999 about 23:59:59, a single second will ring in the new year 1.1.00 with champagne and fireworks. But which computer will actually know what hour has come? Will the clock fall back to 1.1.1900? Or maybe to 1.1.1970 (the artificial birthdate of WIntel and Co.)? Or will the calculator just count on, unimpressed by the new millennium?

Thousands of programmers are devoting time to this "Y2P", (the Year 2000 problem). In the digital world, millions, even billions of dollars are being considered for the production of software "updates", "patches " or by totally replacing whole EDP-systems. Sometimes the danger is fully overrated. Horror scenarios are depicted at sales-generating seminars where the main motivation is the potential profits of the "experts".

It’s true that YP2 could produce strange effects for banks, insurance companies, stores, or government administrative offices. For example, such a "time jump" might mean that the cash registers at a supermarket couldn’t be opened and that the safes in banks would be permanently locked...because the computers thought that, since the date was 1.1.1900, money would be needed again only in a hundred years. Or perhaps your tax returns (in Euro?) could lie around for 30 years? That will have surely unpleasant material consequences. Things would be put out of joint. That might happen in our complex world of rules and systems, if these Y2P "bugs" (programming errors) were suddenly to come to the surface. Free reign should be granted to the imagination, because even the craftiest experts probably won’t think of everything.

The opponents of nuclear power already have a reputation as "enemies of technology" and "defeatist", but this topic has an especially bitter taste.

Because who can be definitely certain that when the new millenium arrives, there won’t be any problems in the 420 computer-controlled atomic plants that are now in operation throughout the world? The possibility of a big bang – an uncontrolled reaction – at midnight, just because someone overlooked a single switch somewhere in all that complex hard- and software, is not that far-fetched.

There remains only the possibility of an immediate shutdown. That’s been our demand for some time: "shut down NOW!" But wouldn’t the computers also have to be programmed to handle such an emergency?

Siemens – and the other atomic geniuses throughout the world – say that we atomic opponents have a "bee in our bonnets". They accuse us of "scaremongering

devoid of real groundwork", calling us "flipped-out green rioters" if we bring out these questions. Hopefully, they’re right: in that case our fears are without foundation. Surely, those herds of highly qualified computer scientists must have searched the software at least a hundred times over. German security technology leads the world in this respect. That’s what they tell us.

But what about the many reactors that can’t be updated, because they use neither German nor American software and the technology isn’t interchangeable?

Or aren’t being fixed for economic reasons or a lack of consciousness of the problem? How much do we really know about the reactors in the former Soviet Union, China, or elsewhere that were built with non-Western technology?

Therefore during the coming months we’ll devote this page to discussion as to whether there are actually any atomic installations, including reactors, research institutions and reprocessing plants that still haven’t solved their Y2P problem, where the possibility of an uncontrolled reaction can’t be completely eliminated. We ask all computer scientists, physicists, atomic technicians and control engineers to turn your concrete thoughts to this topic and then to spread your ideas world-wide through this site.

Chernobyl has demonstrated to all of us that the world is round and that radioactivity knows no borders. Therefore we’re calling on anybody who is familiar with this problem to enter the discussion, gather information, check facts, and share with us, in order that we all may discover any possible security gaps. The end of the 20th century mustn’t mean a repetition of Chernobyl in the new millenium.

We at CASTOR-NIX-DA! Are only concerned laymen. If our worries should be unfounded you may laugh at us, and we’ll be glad to laugh with you.

But first we have many questions. For example,

- Which hardware is used in for the control of an atomic plant? Discreetly based calculators? Microprocessors? INTEL-Chips? Or other types? And since when? Which time control chips?

- Which software steers the controls? Is it independent from the operating systems and close to the machines? Or is it a modern type capable of operating in developing environments? Within the regulatory or control system, can an hour/date problem have any effect at all, or can this be excluded?

For example, we know that in America during the 80’s, the "Altos" computer was installed to control atomic plants. These computers employed the latest 286 chips, ran on a Unix operating system, and were considered absolutely modern.

But was the Y2P problem taken into account then? If not, have the computers been switched, or are they still in use in accordance with the well-known EDP principle, "never touch a running system"?

At the end of the 80’s in Germany at least, programs for atomic plants were developed for use with "Primos" computers. These systems are definitely not Y2P-proof. In what language were the programs written -– Fortran 77, Cobol, or Pascal? In any case the "Primos" operating system recognizes only two-digit years.

For the widely used Unix computers, the well known Sun company has prepared an update package for their Solaris control system. This is said to include a Y2P-patch....but who can definitely say that all problems have thereby been solved?

Programmers have defined "time", in the absence of definitive standards, according to their own moods and desires. Some start their systems at 1.1.00, by which they mean 1.1.1900. Others use the "hour of birth of the p.c.", fixed on 1.1.1970 in order that the computer can grasp the transcience of time. Still others, the programmers of Excel, for example, (refer to c’t August and September issues) even require complex curves to calculate how many hours have passed from 19:00 until 3:00 in the morning; if a negative value is given the program answers straightaway with "error".

What happens, if in the year 2000 the value "00" must suddenly be calculated?

"Stack overflow", "segmentation violation", a crash because of an attempt to divide by zero?

"Murphy's laws" are valid...not only for EDP generally, but specifically at atomic plants. "What can happen, will happen." Like at the Esensham/Unterweser

plant a few weeks ago when a fuel rod got stuck during a change. It wouldn’t go forwards or backwards....and that wasn’t anticipated. As of now we haven’t heard that they’ve come up with a solution. "Murphy" lives again...and will also in the year 2000!

We ask everybody to set their minds to this exchange with friends and fellow workers...worldwide, because the problem is international. Send your thoughts or ideas by e-mail to CASTOR-NIX-DA! If you are aware of other sources, facts or data on the Internet, please let us know, and we’ll publish them on our link lists.


Dieter Metk, 6.9.(19)98

Bearbeitet am: 14.09.1998/ad

zurück zur Homepage