Second quarter of 2001

Unprecedented leaks on fuel assemblies at Cattenom

WISE-Paris, 31 May 2001

[Posted 14/06/2001]

An "unprecedented" number of "leaking faults" of nuclear fuel was noted on 15 March 2001 at Cattenom, by the operating company, Électricité de France (EDF). During the shutdown for the ten-yearly visit of reactor 3, the user company noted leaks in not less than "28 assemblies, comprising 26 of the third cycle and 2 of the second cycle" (1), which included, "on average, 3 damaged rods", representing a total of 92 damaged cladings (2). Up to "18 rods" showing transverse cracks were found in one assembly, which is nearly as twice many as the "5 to 10 cracks of clads" usually counted "in the total number of reactors" every year.

Located in Lorraine, near Moselle, within a stone's throw of the German and the Luxembourg borders, the Cattenom site is composed of four 1300MW reactors. The core of each of the reactors is made up of 193 assemblies, each composed of 264 steel claddings, or rods, which contain the fuel in the form of uranium oxide pastilles. The leaking is due to cracks in the claddings, which constitute the first of the three containment barriers, fundamental for the reactor safety (the second and third barriers are respectively the reactor vessel and (biological) containment).

Last year EDF detected "an increase in the radioactivity of the water of the reactor's primary coolant circuit" (3) and, suspecting a leaking, prepared itself to "manage the problem" but only when it shut it down on January 27 2001- without anticipating it. But given the "unusual" number of "assemblies concerned", it suggested to classify the incident at level 1 on the INES scale (International Nuclear Event Scale, graduated from 0 to 7). The reactor, which was scheduled to restart at the latest in the beginning of May 2001, has not been restarted yet.

According to EDF, the claddings had never presented such -transverse and longitudinal- cracks, the length of which reached 1 cm and a half. One of the rods was found to be "broken", but the fuel pastilles were not released, assured the communication office of the Cattenom site.

The origin of this exceptional phenomenon has not been established with certainty. The first hypothesis is that of a quality failure of the claddings, based on a recent precedent: in August 2000, a fuel cladding was found broken at the level of reactor 2 of the Nogent-sur-Seine site.

The investigation revealed sometime later a problem of quality control concerning a certain number of rods at the Paimboeuf factory (4), only admitted a long time afterwards by Framatome, parent company of the claddings manufacturer, CEZUS (Compagnie Européenne de Zirconium).

However, the Safety Authority (Direction de la sûreté des installations nucléaires (DSIN)) affirms that "the assemblies in question are old, (...) in their third processing cycle" (5) and sees no relationship with the CEZUS case.

Asked to comment on this matter, Framatome, whose group provided the assemblies in question from the Pierrelatte, Dessel and Romans factories, says that "given the fact that the incident took place in the reactor and during the exploitation phase... once the assemblies are delivered" the supplier "doesn't have to comment" (6).

EDF's first investigations tended towards another track. The concerned assemblies are spread "homogeneously" and EDF has not established a cause and effect relationship between the core loading and the leaking. Yet, Mr. Dominique Minière, Director of the Cattenom site, put forward the theory according to which the cause of the leaking is linked to "fretting and vibrations in the lower grid", adding that they expected "a four to six-month inquiry" and that "several causes may be combined, for example damaged grids" (7).

The first elements of the inquiry have shown in fact "leaking problems... due to vibration phenomena", but the origin of the vibrations remains unknown. EDF suspects these vibrations of having induced "more sustained fretting" between the claddings and the springs that maintain them, resulting in term in their cracking.

According to the position expressed by the DSIN in April 2001 (8), the reactor will not be allowed to restart until the vibration phenomenon is clarified. EDF nevertheless sent "a technical proposal" to the Safety Authority, in which it suggests to replace the damaged assemblies and restart the reactor "under reinforced observation", given the fact that the "immediate cause" is now fully established.

Besides, the exploiting company will continue to study the origin of the vibration phenomenon and submit the results of its studies to the Nuclear Protection and Safety Institute (Institut de protection et de sûreté nucléaire (IPSN)) which is charged by the Safety Authority to check whether they are "coherent" (9). According to the Cattenom site, "part of these investigations" cannot be undertaken "unless the reactor is restarted".

As of June 6, EDF had "a one-month delay on the schedule" for maintenance works and replacing of the damaged assemblies, knowing that the average cost per day for the shutdown of a reactor, either programmed or not, is estimated by EDF at 1 million francs.

It is not the only problem the electrician is experiencing. In fact, on 12 March 2001, EDF informed the safety authority, that it detected a manufacturing anomaly on the emergency coolant circuit valves of 12 reactors (1,300 MW units), of which Cattenom's reactors (10), which may affect their operating in emergency situations.

In the case of important leakage from the primary circuit, and once the water reserves of the PTR reservoir are exhausted, the emergency coolant circuit permits to recover the water that flows into the sumps (11) within the reactor containment and to reinject it into the primary circuit.

EDF proposed to modify the valves on the five sites concerned by the anomaly. According to the DSIN, the new system is being put in place and the operation is almost completed (9).

The Safety authority reclassified this generic incident at level 2 on the INES scale, on 27 April 2001.



  1. This reactor is exploited within the framework of extended campaigns (GEMMES) of 18 months per cycle. The loading is programmed for one-third of the core at a time
  2. Unless specified otherwise, EDF's statements were made by the Press office of the Cattenom site, on 17, 28 and 31 May 2001, during telephone conversations
  3. See the Safety authority's 19 March 2001 communiqué on the official site (
  4. See "The Cezus affair: A flaw in the quality control of nuclear fuel", WISE-Paris, 20 December 2000
  5. Mr. Alain Delmestre, DSIN, telephone conversation, 2 May 2001
  6. Framatome, Press service, telephone conversation, 17 May 2001
  7. Cited by WOXX, published on 20 April 2001
  8. According to Nucleaonics Week, "EDF, Framatome still in the dark about cause of Cattenom leakers", 3 May2001
  9. Mr. Thomas Maurin, DSIN, personal conversation, 31 May 2001
  10. The other reactors being those of Belleville, Golfech, Nogent and Penly
  11. PTR reservoir, among other functions, stores the water necessary to the fuel handling pools. Sumps collect the water leaking from the primary circuit which is used to cool the core

Back to contents