Wampum

The World Information Service on Energy site hasn't been updated since mid-July (it is holiday time in Europe), so there is no news there. There are four back-up generators at Forsmark 1, a 968 MWe BWR that become operational in 1980. When power was lost to the reactor, in particular to its monitors and cooling pumps, two of the four generators could not be started.

Not confidence building. Since the MSM coverage has been wicked scanty, to the point of finding out which of three units experienced the failure, and what the failure was, is a multi-hour challenge, here's some of the RASK report down by SKi staff the following day.

From SKi:

On July 25 at the Forsmark unit 1 NPP there was a short circuit in a 400kV outdoor switching station. Due to this the plant scrammed in a way that included a number of subsequent events in a complex scenario.

After disconnecting the unit from the grid due to the short circuit there was a partial scram and both turbines for a short while transferred to house load operation. After the turbine trip the reactor scrammed. A number of conditions in the safety trains (in system 516, the reactor protection system) tripped: several scram conditions, I-isolation and N-chain. The reactor scram could be seen through WRNM even though the indication for control rod positions was unclear due to the unit partly having lost its power supply. Water was pumped in using two of the lines in system 327, the auxiliary feedwater system (2x22,5 kg/s). Four of the eight reactor recirculation pumps were in operation.

Pressure relief of steam from the reactor to the condensation pool was done through two pressure relief valves in system 314, the automatic depressurization system (about 2x50 kg/s) that had been opened via the N-chain.

Reactor pressure and water level in the reactor went down. The display of the reactor level was ambiguous since some actuators were not active due to loss of power. The water level was down 2 m and the pressure went down to 12 bar after about 20 minutes. The emergency cooling system which had already started on isolation signals pumped water into the reactor vessel for a short while when pressure had been reduced. Sprinkling was activated in the containment.

The shift team checked the level in the reactor vessel in order to be prepared to activate the automatic depressurization system if the level were to be reduced to 1.1 m, in accordance with the Emergency Operating Procedures.

After 23 minutes the shift team realized that there was a possibility to manually restart the two diesels that had stopped, and after this the situation was quite quickly stabilized. The 6 kV bus bars were then already operational.

The reactor was then at hot stand by.

After the shift handover to the ongoing shift the leaving head shift engineer had a debriefing with her team.

SKI concludes that the event badly affected important redundant components, namely the DC/AC inverters for feeding of the battery secured 500 kV-bus bar from a UPS (Uninterrupted Power Supply). This means that this is a common cause failure event. If the other two subs had been knocked out as well this would have led to a total loss of power, including the battery secured net. This is a more severe case than was anticipated in the Safety analysis report. During the visit there was no obvious direct cause for two subs being knocked out, whereas the other two were not.

SKI furthermore states that:

The work on assessing the course of the event seems to be well described in spite of the difficulties obtaining the information about it. The events in the reactor part, however, were not well described. In spite of a very unclear signal display, knocked out computer screens as well as the loud speakers being out of order, the control room personnel seems to have done their job according to their instructions. The control room also received valuable help from the control room personnel at units 2 and 3. The head shift engineer also summoned the next shift about an hour prior to the scheduled time. The motive for this was to make sure that they were informed of the event well in advance, and the head shift engineer also judged that it was uncertain whether her shift team could complete the whole shift.

Possible causes for the event and contributing conditions

The initiating event occurred in connection with maintenance work done by SVK (The company that administers and runs the national electrical grid in Sweden), and this was done while unit 2 was out of operation due to its refuelling outage. SVK had written a work order and had informed FKA about it. FKA would have had the right to react on the maintenance being done exactly at this moment (and has done so in other cases), however this time there was no need to react and ask for the maintenance work to be postponed. The reason for the short circuit in the switching station has not been ascertained and SVK has still to submit a report on the disturbance.

The 70 kV-net was probably instable. This is to be confirmed by SVK. The instable voltage in the 70 kV-net led to the 6 kV-net also being unstable. When automatic switch tried to connect the 500 V-net the 6 kV-net was too unstable, and automatic switch then tried to feed the 500 V-net from the diesels. It is essential that a complete picture of the steps in the event be put together and confirmed.

The reason for two of the battery secured bus bars being knocked out is, according to the primary analysis, that the voltage transient tripped the rectifier as well as the inverter, which according to the utility FKA is due to incorrect design. The inverter should have been in operation to make the batteries feed the 500 kV-net. The tuning of the protective devices should be done in such a way that these trip selectively, so that the DC/AC-converter for battery voltage to the 500 kV-net is protected.

The UPS (AEG delivered) were installed in about 1993-1994 as an improvement of the former rotating transformers. Information from AEG to the utility FKA, but not confirmed, claims that a similar event occurred in an NPP in Germany, and that AEG was aware of the problem and had taken measures to prevent this error reoccurring. This implies routines and practices connected to experience feedback need to be checked.

One problem was that the list of events was far from complete. Many events were registered, however with no time recorded, and probably some events were missing altogether. This has meant that detective work is needed to investigate the course of events.

That reads like the operators were prepared to breach into the containment vessel. What happens after that is a roll of many loaded dice.

Ironically, on April 27, 1986 workers at Forsmark were found to carry radioactive particles. The origin of the presumed leak was investigated and it eventually became clear that the contamination came from the atmosphere. This was the first place outside the Soviet Union that detected the Chernobyl plume of April 26, 1986.