(July 21, 1995) Investigations following the "loss of coolant accident" (LOCA) at Pickering reactor 2 (Canada) on December 10, 1994 have revealed that the written instructions followed by reactor operators actually contributed to the accident. The same design flaws that caused the Pickering accident exist at other CANDU reactors.
WISE Amsterdam (436.4306) - A similar but less severe accident happened at Bruce reactor 5 only a few months later, despite a detailed analysis by Ontario Hydro following the Pickering accident. In both accidents, super-heated, radioactive cooling water leaked from the reactor core because of faulty valves. LOCAs can result in overheating and melting of reactor fuel and large releases of radioactivity from the station. In both of these accidents and a third similar accident at Wolsung CANDU reactor 1 in South Korea, damage to the fuel bundles was apparently avoided.
LOCA - Pickering, December 10, 1994
This accident resulted in the use of the emergency coolant injection system, a special safety system, for the first time ever at a CANDU reactor. Initially, a liquid relief valve in the reactor core cooling system stuck open, allowing cooling water to enter the bleed condenser (an overflow tank for the cooling system). Following automatic shutdown of the reactor, pressure increased in the cooling system, causing bleed condenser pressure relief valves to open. The piping to one of these valves cracked because of severe pipe vibration, spilling cooling water into the reactor building. A total of 185 tonnes of water were spilled.
The emergency coolant injection system pumped about 140 tonnes of water into the core over several hours to prevent overheating and melting of the fuel. Workers stopped the leak after 1 1/2 hours by manually closing all liquid relief valves on the core cooling system. About 200 workers were involved in the clean-up effort.
Radioactivity was released to the environment following the accident at levels comparable to a normally operating station, even though all four reactors were shut down pending the investigation and design changes. For example, during January, Pickering "A" radioactive emissions included 600 billion becquerels of tritium to air and 37 trillion becquerels of tritium to Lake Ontario. It is estimated by Ontario Hydro that Pickering reactor 2 will be out-of-service until October 1995, since biennial maintenance work planned for this fall was rescheduled to overlap with the investigation of the accident. The cost for repairs and retrofits at the four Pickering "A" reactors is currently estimated to be $10.5 million.
The Atomic Energy Control Board established an Investigation Team to review the accident circumstances. The AECB Investigation Team determined that some actions taken by reactor operators (as instructed through the Power Reduction Action Guide) actually contributed to the accident. In addition, stopping the leak by closing all the pressure relief valves meant there was no overpressure protection for the entire primary cooling system, and this was done without regulatory approval. According to the Team, Ontario Hydro's Root Cause Analysis is incomplete. The Investigation Team made a number of recommendations to Ontario Hydro including: revision of reactor operating procedures; audit of the human performance evaluation system; and review of nuclear emergency procedures to make the Ontario Hydro plans consistent with the province's nuclear emergency plan. The Team also recommended that the AECB develop and implement an emergency response plan and establish a system for reviewing safety reports, accident reports and operating procedures in a timely way. The Team concluded that in-depth inspections are needed at nuclear stations, in addition to the usual compliance inspections done by the AECB staff, and that the process for reviewing operating procedures must be examined, since the instructions followed by the operators contributed to the accident in this case.
Significance for other CANDU stations
AECB staff reviewed the implications of the loss of coolant accident at Pickering for other CANDU stations because all CANDU reactors have similar core cooling system and bleed condenser overpressure protection designs. The root cause of these recent accidents has been traced to failure of a diaphragm in a liquid relief valve of the reactor core cooling systems. The secondary accident stage involved damage to the bleed condenser relief valves, and in the case of Pickering, the piping as well. The review concluded that design of the bleed condenser pressure relief system is inherently flawed in CANDU reactors. AECB staff have recommended design changes for each nuclear station in Canada with the exception of Pickering. Pickering "A" had design changes made since the December accident. Pickering "B" reactors apparently have different bleed condenser piping arrangements that won't lead to this type of accident. AECB staff noted that the analysis carried out by Ontario Hydro Nuclear at Bruce "B" prior to the May accident failed to identify the possibility of the very situation that did arise.
Heavy Water Coolant Spill - Bruce reactor 5, May 14, 1995
This accident was the result of a liquid relief valve leading from the reactor cooling system to the bleed condenser sticking open while the reactor was running at 88% of full power. The reactor shut down automatically, despite efforts by the operators to reduce the reactor power levels slowly. Unexpectedly, the cooling system pressure increased, causing two bleed condenser relief valves to open, dumping heavy water coolant into the reactor building sump. 87 tonnes of heavy water were spilled and 68 tonnes were automatically recycled back into the cooling system during the accident. Most of the remaining 19 tonnes were recovered later by clean-up crews. About 20 staff were involved in the clean-up. Ontario Hydro downplayed the seriousness of the accident by noting the amount of heavy water that had to be manually cleaned up as if it were the total amount spilled, and by stating publicly that there had been no radiation releases from the station. However, 180 billion becquerels of tritium were released to the air during May 14, 15 and 16.
Source: Nuclear Watchdog Bulletin #1, June 1995 (in igc:energy.news)
Contact: Nuclear Awareness Project; Box 2331 Oshawa, ON L1H 7V6 Canada.