free web site hit counter
Search CNIC
Rokkasho Reprocessing Plant

Leak of 150 Liters HLW in Vitrification Facility

Active testing of the Rokkasho Reprocessing Plant continues to be plagued by problems and accidents. Since the beginning of 2008, Japan Nuclear Fuel Ltd. (JNFL) has successively extended the schedule for commencing commercial operations from February 2008 to May to July to November and then to February 2009. In January 2009, 150 liters of high-level liquid radioactive waste (HLW) leaked from pipes in the vitrification cell, so JNFL was forced to extend the start-up date by a further six-months to August this year. However, there is no scientific or technical basis for selecting August as the starting date. It is simply a target for JNFL.

Continuing problems with production of HLW glass canisters
The Rokkasho Reprocessing Plant produced its first MOX powder product in November 2006 and testing of the final facility, the Vitrification Facility, commenced in November 2007. However, vitrification testing has been a comedy of errors, with each attempt to fix one problem giving rise to further problems. As a result, in fourteen months only 107 glass canisters have been produced. (It is supposed to produce 1,000 in 12 months.)

Accumulation of platinum group elements
During the first two months, testing of the Vitrification Facility was initially suspended because of damage to the machine that welds the lids onto the vitrified waste canisters. In the latter half of this period, platinum group elements contained within the HLW liquid, including ruthenium, accumulated at the bottom of the vitrification furnace blocking the outlet. The vitrification tests require a flow rate out of the bottom of the furnace of 70 liters of molten glass per hour. This standard was not being met, so the tests were suspended again. Nevertheless, in February 2008 JNFL submitted a report on step 4 of the active tests, which claimed there were no safety problems with the production of glass canisters. The Nuclear and Industrial Safety Agency (NISA) accepted this report and approved resumption of the tests. However, pointing to the lack of specificity in the prediction that stirring the mixture in the bottom of the furnace would prevent the problem of accumulating platinum group elements, it required a follow-up report.

By allowing JNFL to continue testing the production of glass canisters, even though the tests had not been passed and the outcome of the follow-up report was completely unclear, NISA effectively exacerbated the technical problems of the vitrification furnace. JNFL proposed stirring the mixture to deal with the problem of accumulation of platinum group elements, even though this was not considered in the original design. As an afterthought, two stirring rods (one straight and one bent) were added to the top of the furnace.

Low viscosity fluids
New measures to deal with platinum group elements were introduced and testing was recommenced on July 2, 2008. However, the tests were suspended after just 32 minutes, without producing a single glass canister. The nozzle through which the glass was supposed to flow could not be heated, so the glass would not flow through.

It is conjectured that the cause of this problem was so-called "low viscosity fluids". Substances that are less viscous than molten glass appear to have flowed down more quickly than the glass and adhered to the nozzle. These substances included radioactive substances such as molybdenum. The substances that had adhered to the nozzle were removed and the tests were finally resumed on October 10.

Failure of the stirring rods
Testing was resumed and initially it seemed like third time would be lucky. On October 20 accumulation of platinum group elements was confirmed. JNFL shifted to so-called "cleaning operation" (only glass ingredients fed into the furnace) and reported that this was effective. Undissolved sludge was fed into the furnace for the first time on October 23.

On October 30, accumulation of platinum group elements appeared again and preparations were made to insert the straight stirring rod. However, on November 1, the window through which the rod was to be inserted would not open. JNFL spent until November 17 replacing the window. When the metal rod was finally inserted, it did not work properly, so a camera was inserted to see what the problem was. It was discovered that it was bent virtually at right angles. The rod was made of heat resistant Inconel Alloy, but it appears that it bent as a result of the heat of the furnace and the load.

In response to this problem, the rod and all the molten glass were removed from the furnace, but on December 24 it was discovered that one of the tiles (6 kg) had fallen from the ceiling of the furnace. It is very likely that it was damaged when the rod was removed.

What happened to 150 liters of leaked HLW?
On December 16, because of the problems of the bent stirring rod and the fallen tile, JNFL closed off flanges to stop the flow of HLW through two pipes connecting the HLW feed tank to the furnace. However, in January it was confirmed that 150 liters of HLW had leaked from this area. (See diagram below.)

Starting from January 9, an alarm in the vitrification cell indicated a rise in the level of liquid in the tray beneath the flanges and in the catchment tray on the floor. The alarm went off then stopped again repeatedly. The operators checked with cameras, but the level of the liquid in the trays was below the level at which the alarm should go off, so they disregarded the alarms. It wasn't until January 21, when the liquid in the trays was analyzed for the first time and the radioactivity reading was found to be very high, that JNFL recognized that there must be a leak of HLW liquid.

JNFL says that about 150 liters leaked from the HLW feed tank, but that only 16 liters was recovered. What happened to the rest of it? JNFL claimed that it evaporated within the vitrification cell. The reading for cesium 137 in the leaked liquid was 1.6 x 1010 Bq/ml, whereas cesium 137 Feed Tank A was only 3.6 x 109 Bq/ml. That means the concentration in the leaked liquid was 4.4 times that in the feed tank. This indicates that the liquid had become concentrated during the period while the leak was continuing (January 9~24). As the liquid gradually dripped from the flanges it accumulated in the tray below, then overflowed into the catchment tray at the bottom. The liquid then evaporated from the trays in the hot vitrification cell.

JNFL's explanation revealed that there was a series of errors. HLW Feed Tank A feeds waste liquid to the furnace by an "airlift" system. Although HLW liquid was not being fed to the furnace at the time of the leak, pressure was being applied to purge waste gas. JNFL suspects that for some reason the pressure rose and liquid was forced up the feed pipe. Compounding the problem, the flanges were not properly closed. (After the leak was confirmed, the flanges were tightened and the leak stopped.) Although alarms were going off and the level of the liquid in HLW Feed Tank A had fallen, JNFL failed to connect this with the possibility of a leak.

The incident is an indictment of JNFL's shoddy management of the Vitrification Facility. It is further evidence that JNFL is not capable of operating a facility as complex as the Rokkasho Reprocessing Plant.

Masako Sawai (CNIC)

Return to NIT 129 contents

Go to CNIC's page summarising the state of active tests at Rokkasho.


Citizens' Nuclear Information Center
Akebonobashi Co-op 2F-B, 8-5 Sumiyoshi-cho,
Shinjuku-ku, Tokyo, 162-0065, Japan