|
On the Computer Troubles at Japanese Nuclear Power Station on 1 January 2000
6 January 2000
Agency of Natural Resources and Energy, MITI
(1) Four glitches occurred on January 1, 2000 at nuclear power stations in Japan. Three of them were attributed to the year 2000 problems in computers. (In addition, three glitches in administrative computers and one in a nuclide sample analyzer took place and those were also triggered by the Y2K problem.)
These glitches took place in non-power related systems, such as safety parameter display system, and did not disturb safe/stable operation of nuclear power plants. There was in fact no shutdown of nuclear power plants.
Investigations of the causes started immediately after the occurrence of the glitches and all of the troubled computers were brought back to normal operation in a few minutes or within the day, at the latest.
It is to be noted that except for the (4) below, the computer glitches experienced at nuclear power stations on January 1 were not of the kind requiring immediate trouble information release under the regulation. The trouble (4) was evaluated as 0 - an event having nothing do with safety - in INES (International Nuclear Event Scale) scaling. In spite of the said nature of the glitches, due to the fact that Japan was one of the earliest year 2000 rollover countries which operates nuclear power plants, information on any minor event could be useful to the rest of the world for a better rollover management. Based on this conviction, we think all nuclear facility operating countries may need to take the similar policy when the world face a challenge as Y2K in the future.
(2) The following are the non-power related computer glitches experienced on January 1 at nuclear power stations in Japan. Among them, (1), (2), and (4) were identified as caused by the Y2K problem.
(1) Tohoku Electric Power's Onagawa Nuclear Power Plant Units 1 and 2
At 0:02 on January 1, an indicator was activated to point an anomaly in Process Computer. In fact, one of the every ten minutes reporting system on seawater radiation monitors, etc. skipped but automatically restarted at the next reporting time, i.e. 0:12. The glitch was caused by the erroneous conversion of time from two-digit to four-digit, i.e., the time "24:00 on December 31, 1999" was misinterpreted as "24:00 on December 31, 2099".
(2) Hokuriku Electric Power's Shika Nuclear Power Plant Unit 1
At about 0:00 on January 1, malfunction of the Safety Parameter Display System occurred. Investigation revealed that mishandling of four-digit time data in the data-receiving program led to the glitch.
SPDS is a contingency data transfer system from the plant to the Ministry of International Trade and Industry (MITI) and not usually used.
(3) Kansai Electric Power's Takahama Power Station (non Y2K)
At about 2:00 on January 1, it was found that the data from two monitoring posts installed in Maizuru City for Takahama Power Station had not been sent to Kyoto Prefecture.
Investigation revealed that the glitch was caused by a failure in a computer program controlling the data storage areas. This was not triggered by the Y2K problem.
(4) Tokyo Electric Power's Fukushima-Daini Unit 1
During rated power operation, an alarm of "Control Rod Position Indicating System Multiple Failure" was actuated at about 8:58 on January 1, which resulted in loss of "control rod position indication". The process computer monitored the control rod positions and the plant operation was safely continued.
The cause was that in some of the parts of circuit boards for control rod coordinate display and position display using GMT mis-recognized year 2000 as year 1900 at the millennium rollover (i.e. 09:00 on January 1, local time). This resulted in mishandling of time and consequently the time January 1, 2000 was erroneously set at February 6, 2036, which led to the failure in the error indicator of the control rod position indicating system panel.
The system came back to normal by setting correct time in the mounted real time clock.
Rollover Reports | More Information
|
