TL: Update on the closure of 5 Swedish BWR reactors SO: Greenpeace Sweden (GP) DT: September 26, 1992 Keywords: nuclear power reactors shutdowns sweden scandinavia europe safety problems / Update on the closure of 5 Swedish BWR reactors =============================================== Based on a meeting with Mr Frigyes Reisch of the Swedish Nuclear Power Inspectorate (SKi) on 23 September 1992. Contents -------- 1. Background 2. The Barsebaeck-2 Incident 3. Sweden's 5 oldest BWRs not allowed to operate 4. SKi's criteria for allowing continued operation 5. Implications for the other Swedish reactors 6. Implications for other countries 1. Background ============= Due to the residual or decay heat of fission products in a reactor core, a reactor still generates a significant proportion of the normal operating heat output following a scram and this residual heat falls off only slowly. The amount of residual heat is sufficient to pose a danger of post-shut-down melting of the core and possibly a breach of the primary circuit unless cooling can be guaranteed under all possible conditions. The continuing provision of emergency core and containment cooling in the event of a LOCA is critical. In Sweden's 5 oldest BWRs the water supply for the emergency core and containment cooling systems is drawn solely from the condensation pool beneath the reactor pressure vessel. The water intakes to the emergency cooling systems could thus be blocked by any material falling into the pool. While the emergency cooling systems have been designed to cope with this, in an incident on July 28, they did not function as predicted. 2. The Barsebaeck-2 Incident ============================ On the 28th July an incident occurred in connection with preparations for the start up of the Barsebaeck-2 BWR following its annual outage. A safety valve inside the containment opened inadvertently because of a leaking pilot valve. This resulted in insulation material (approx. 100 kg) being carried by the emergency cooling water to the condensation pool beneath the core. The emergency core and containment spray systems take their cooling water from the condensation pool via strainers. After an hour, the strainers on the intakes to the emergency core and containment cooling systems were found to be blocked. The strainers were successfully cleared by reversing the water flow (back-flushing). Other safety systems functioned normally and there were no radioactive releases. In the Barsebaeck-2 incident, which occurred during heating when the pressure passed 30 bar (normal operating pressure = 70 bar) and at less than 2% of full reactor power, the strainers clogged at a point in time, which in this particular case, the containment spray was no longer needed. Initially the incident was provisionally rated at 1 on the IAEAs International Nuclear Event Scale (INES), but in its final rating it was uprated to 2 on the scale on the 17 September. 3. Sweden's 5 oldest BWRs not allowed to operate ================================================ It had been assumed that the strainers on the emergency cooling systems' intakes can become clogged by material in the condensation pool and the system had therefore being designed so that the strainers could be cleared by reversing the flow (i.e. back-flushing). However, in the safety case for Barsebaeck-2, and the other 4 BWRs which have the same/similar design features, it had been assumed that back-flushing would not be necessary within the first 10 hours of use of the emergency core and containment spray systems. For this particular design of BWR the back-flushing of the emergency core and containment spray systems is therefore a critical operation and it requires time to be conducted safely. Extrapolating from the Barsebaeck-2 incident, it was concluded that the time frame for back-flushing is a considerably shorter one than that envisioned in the safety analysis for pipe failure in the containment. It was determined that the premature clogging of the strainers would have a marked effect on safety in the event of a pipe failure in the containment while the reactor was at full power. Therefore SKi considered that safe operation as prescribed in the safety analysis on which operational permission was based is not fulfilled and that this is a serious deviation. Following this analysis, SKi requested during the week of September 7-11 that the operators of the 5 oldest Swedish BWR reactors should, as soon as possible state their opinions concerning the safety situation in the light of the events at Barsebaeck-2. By this time Barsebaeck-2 was in fact the only one of the 5 then in operation. Barsebaeck-1 and Oskarshamn-2 were down for annual outages, while Ringhals-1 had not been allowed to restart after its annual outage following the discovery of cracks in its feedwater pipes, and Oskarshamn-1 was ready for restart following a cold shutdown for turbine balancing and other work. Both SKi and the reactor operators considered that the function of the emergency core and containment cooling systems could not be guaranteed during operation of these reactors. Therefore Barsebaeck-2 was shut-down and on the 17 September SKi announced that none of these 5 reactors would be allowed to restart until approval is given by SKi. 4. SKi's criteria for allowing continued operation ================================================== A safety analysis on which permission to operate the reactors until their annual outage in 1993 shall be presented to SKi. A second separate presentation of proposals for the modifications or other measures to be undertaken during the annual 1993 outage shall also be made. 4.1. Safety analysis for operation until 1993 annual outage ----------------------------------------------------------- The reactor operators shall produce a analysis which shall present the safety case with respect to emergency core cooling and emergency containment cooling in the event of a pipe failure inside the containment and taking into account the modifications proposed, the effect on safety and other factors which the remedial measures could have. Related questions necessary for emergency procedures, their verification and operator training shall be considered in the analysis. The analysis is expected to be presented some time in the next two weeks, whereupon it will be reviewed by SKi. The following conditions are set by SKi for restart and operation until at least the 1993 outage: 1. Emergency core cooling shall be provided for every postulated top or bottom break in the primary circuit inside the containment, so that core damage can be shown to be avoided with a safety margin which can be shown to be not markedly less than described in previous safety analyses. 2. Operator actions which are required to maintain emergency core cooling can be taken into account in the safety analysis provided that they need to be initiated not sooner than 30 minutes after the activation of the systems. However: - Backflushing of the strainers as included in current instructions may not be considered as an option during this first 30 minutes; and - Operator interventions may be credited within the first 30 minutes only is a specific analysis is made which considers the nature of the measure, the time involved and the specific working situation. 3. The need for back-flushing the strainers to obtain emergency cooling water from the condensation pool shall be assumed to have a frequency of at least 3 times per hour during the first hours following initiation of LOCA conditions, timed from the start of the back-flushing. 4.2 Operation beyond the 1993 annual outage ------------------------------------------- For the five BWR reactors presently shut-down a general presentation of the measures suggested for introduction during the 1993 annual outage shall be presented to SKi not later than February 1993. A final safety analysis which shall form the basis for operation after the 1993 annual outage shall be sent to SKi no later than June 1993, or at least two weeks BEFORE the start of the annual outage if this is earlier. SKi anticipates that modifications or other measures will be necessary to guarantee the provision of emergency core and containment cooling based upon the conditions imposed by the current safety assessment. SKi require that the necessary safety systems have the redundancy required so that a fault in one individual system component will not prevent their effective functioning (single fault criterion). It appears that there are a restricted range of remedial measures which could be used separately or in isolation to address the problem in these 5 reactors. These include: 1. The provision of an alternative water source to the emergency core and containment cooling systems, which does not draw its water from the condensation pool; 2. Changing the configuration of the strainers over the intakes of the emergency core and containment cooling systems to increase their surface area and thus decrease the probability that they will become clogged over a given period of time for a given accident; 3. Change some or all the isolation inside the containment, to decrease the amount of insulation material likely to break off in the event of a pipe failure and thereby clog the strainers in the condensation pool. The current insulation is rock wool, which becomes increasingly brittle with age, alternatives include an all-metal insulation (known here as "mirror insulation"); and 4. Improve the efficiency of the back-flushing operations. On the basis of the criteria set by SKi, it seems possible that each of remedial measures 1-3 might be sufficient individually, while option 4 would need to combined with at least one of the other measures. 5. Implications for other Swedish reactors ------------------------------------------ All the remaining 7 Swedish reactors (4 BWRs and 3 PWRs) are required to send a report to SKi by October 31st covering the safety implications of the Barsebaeck-2 incident for their operations. Of the other 4 Swedish BWRs not presently effected by the SKi decision, problems are not expected - at least not on the same scale as in the 5 oldest BWRs. All four (Forsmark 1-3 and Oskarshamn-3) have annular (ring-shaped) rather than circular condensation pools - decreasing the possibilities for material to fall into the pools. In addition, Forsmark-1 and 2 both have a mixture of "mirror insulation" and rock wool insulation, while Forsmark-3 and Oskarshamn-3 have only "mirror insulation", reducing the volumes of material that would be likely to break-off in the event of a pipe failure inside the containment. I do not know what, if any, implications there are for the PWRs. 6. Implications for other countries with BWRs --------------------------------------------- The implications for other countries' BWRs are very much influenced by the particular designs that are used. I have heard that a number of countries' nuclear inspectorates have been in contact with the Swedish authorities, but it is probable that there are more. Those I know have been definitely in touch with SKi are: Switzerland, Spain and Finland. According to my information the Finnish authorities (STUK) do not expect any problems due to design differences. From the information I have, all I can recommend is that these are features to look for in your BWRs if the emergency core and containment cooling systems draw water from the condensation pool beneath the pressure vessel: 1. what type of insulation used inside of the containment; 2. the shape/surface area of the condensation pool; 3. the surface areas of strainers over the ECCS intakes; and 4. alternative water sources for the emergency cooling systems. ----------------------------------------------------------- Simon Carroll, 26 September 1992