Fourth quarter of 2000
The Cezus affair: A flaw in the quality control of nuclear fuel
WISE-Paris, 20 December 2000
[Posted 22/12/2000]
Download the full version of the CEZUS affair (including annexes) in PDF, 14 pages (225 Ko)
1. Summary of events
In August 2000, the operator of reactor No. 2 at the Nogent-sur-Seine power plant, 120 km from the centre of Paris, France, detected a cladding failure in the nuclear core (abnormal radioactivity in the primary coolant system). The investigation — carried out by Framatome, the fuel manufacturer, on request from EDF (France's electricity company) — revealed what can truly be described as "The CEZUS Affair". What is especially worrying is that the problem was discovered in the manufacturing plant 18 months after its start, and then concealed — i.e. it was not revealed to those concerned, neither French or foreign clients nor safety authorities — at a time when a competitor, the British company BNFL, was embroiled in a scandal over the quality control of plutonium-containing MOX fuel, which brought BNFL to the brink of bankruptcy.
From August 1998, the Compagnie Européenne de Zirconium (CEZUS), then a subsidiary of COGEMA and Framatome, experienced, a quality control problem in its plant in Paimbœuf, France, which manufactures zircalloy tubes for nuclear fuel assemblies, both uranium oxide and plutonium-uranium mixed oxide (MOX) fuel. Although the problem was discovered by the plant operators in February 2000, CEZUS' management did not communicate this to anyone, neither to the clients nor to the safety authorities. Officially, it was Framatome — as client and parent company (by then holding 100 per cent of shares) — that was informed and which communicated the incident to the DSIN (on 6 November 2000) and to its own clients, the users of the fuel manufactured with the some 900,000 cladding tubes produced during the period in question. According to information transmitted by DSIN to WISE-Paris on 16 November 2000, clients included EDF (which also informed DSIN on 6 November 2000) and nuclear operators in countries in Europe (including Belgium, Germany Spain, Sweden, Switzerland); North America (USA); Africa (South Africa); and Asia (China, Japan, South Korea,). The DSIN also stated that it informed the safety authorities in the countries concerned before the 10 November 2000. However, later, DSIN withdrew Switzerland, Japan and South Korea from the list (see point 6.).
The DSIN declared this "generic incident" as at "level 1" on the International Nuclear Event Scale (INES), which has seven levels. This was attributed to EDF for failing, as operator, to monitor its supplier's quality control system. The French safety authority — one of whose assistant directors considered that the event "does not pose a problem for safety" — would have placed the event at level 0 except for the fact of the "long delay between discovery and declaration of the incident". This same assistant director also declared that there had been "no concealment" of the information. However, the other assistant director of this same organisation considered, for his part, that "such behaviour is entirely abnormal" and concluded that, "those who work in the nuclear sector do not have the right" to behave in this way.
The delay in communication has had an irreversible consequence: the tubes affected by the quality control failure, identifiable within batches so long as they are not used, have all left the Paimboeuf plant and have been delivered, assembled and, for the most part, loaded into reactors. Thus, according to information provided to WISE-Paris by DSIN, the tubes manufactured for EDF, were used in 1,263 assemblies of which 1,140 (more than 90 per cent) have already been loaded into 49 of EDF’s 58 pressurized water reactors. The situation is less clear for dozens of other CEZUS/Framatome foreign clients. In the view of the DSIN, it is now up to the safety authorities of the countries in question — which have been informed — to take the necessary steps.
Back
to contents